✏️Prompts
Inside this playbook

Prompts • Skills • Connectors • Agents • Campus workflows • Tools

AI Playbook
for Higher Education Administrators

For business office, finance, advancement, enrollment, and operations staff at colleges and universities. A practical guide to running the institution with AI: advancement and alumni relations, grants, financial aid, budgeting, and campus operations such as bookstores and housing.

100Ready-to-Use Prompts
12Claude-Ready Skills
12Ways to Connect Your Work
12AI Agents for Higher Ed Administration
How to use this playbook
Start with your role, copy one prompt, and work up to skills and agents. All interactive elements save to your browser.
→
Choose your next step
Need a quick first draft? Start with prompts. Do the same reporting, review, or communication work every cycle? Build a skill. Need AI to see your CRM, ERP, or reports? Connect your work. Ready to have routine monitoring done on a schedule? Explore agents.
→

Who this playbook is for

This playbook is for the people who run the institution, not for teaching or research. It covers the business, advancement, enrollment, and operations side of colleges and universities. It doesn't cover AI in the classroom, course design, research methods, or student use of AI.

This playbook is for
  • CFOs, budget directors, and controllers
  • Business office and student accounts staff
  • Advancement, alumni relations, and development staff
  • Sponsored programs and grants administrators
  • Financial aid, enrollment, auxiliary services, and facilities staff
Not covered here
  • Faculty teaching or course design
  • Research methods or scholarly writing
  • Student use of AI
  • Students or parents

Start With Your Role

Choose the role closest to your day-to-day work. Start with one useful prompt, make it repeatable, then add the information it needs.

CFO / Budget Director
  • Start with: variance narrative or enrollment revenue scenarios
  • Save as a skill: reporting format, thresholds, assumptions
  • Connect: ERP and budget system reports
  • First agent: quarterly budget variance drafter
  • Measure: reporting time, forecast accuracy
BudgetPlanningTrustees
Controller / Business Office
  • Start with: P-card review or close status summary
  • Save as a skill: policy rules and exception formats
  • Connect: ERP exception reports
  • First agent: P-card & AP exception queue
  • Measure: exceptions resolved, findings
CloseAuditAP
Advancement & Alumni Relations
  • Start with: annual giving review or prospect profile
  • Save as a skill: segments, appeal formats, research ethics
  • Connect: Raiser's Edge NXT connector or CRM exports
  • First agent: annual giving pace monitor
  • Measure: retention, participation, dollars
GivingProspectsAlumni
Sponsored Programs / Grants
  • Start with: funding opportunity summary or burn rate review
  • Save as a skill: checklist formats and rules
  • Connect: research administration and grant ledger exports
  • First agent: grant burn rate & closeout watch
  • Measure: late reports, disallowed costs
Pre-awardPost-awardCompliance
Financial Aid Director
  • Start with: discount rate model or award letter clarity review
  • Save as a skill: aggregate bands, communication standards
  • Connect: aggregate aid reports only
  • First agent: aid communications & deadline assistant
  • Measure: processing time, call volume
AidComplianceCommunication
Enrollment, Auxiliary & Campus Operations
  • Start with: funnel narrative, housing forecast, or work order backlog
  • Save as a skill: metrics and report formats
  • Connect: CRM, housing, and work order exports
  • First agent: enrollment funnel watch or housing occupancy forecaster
  • Measure: yield, occupancy, backlog age
EnrollmentHousingFacilities
Start with a task you already do
Pick something your unit does every month or cycle with the same information and one person who signs off. The variance narrative is a great first choice.
→

100 AI Prompts for Higher Education Administrators

Ready-to-use prompts for advancement, alumni relations, grants, financial aid, budgeting, the business office, enrollment operations, auxiliaries, facilities, and compliance. Copy one, add your institution's context, and review the result.

Start with a prompt

Follow your data classification policy, use aggregates and IDs, verify every figure and rule, and review the result before it leaves your unit.

Prompts for annual giving, major and planned gifts, campaigns, and stewardship. Use donor IDs and sourced information; gift officers own relationships.

Annual Giving Performance Review
You are an annual giving analyst at a college. Review this fiscal year's performance. Data by constituency (alumni by decade, parents, friends, faculty/staff) and channel (mail, email, phone, giving day, peer-to-peer): [donors, dollars, retention, new donors, reactivated, cost] Produce: performance vs. last year and goal, cost per dollar raised by channel, retention by segment, the 3 biggest opportunities, and questions for the annual giving team.

Shows which constituencies and channels drive results and which drain budget.

Risks: Gift counting rules affect totals. Control: Follow your institution's counting policy (e.g., CASE standards) when comparing.

Donor Segmentation for Appeals
You are an advancement data analyst. Segment donors for the fall appeal. Data (constituent ID): [constituency, class year, last gift date and amount band, giving years, event attendance, volunteer status] Produce: segments (current, LYBUNT, SYBUNT, never-given alumni by decade, leadership annual giving prospects), counts, recommended ask string or approach, and channel per segment.

Tailored segments raise participation and average gift.

Risks: Data quality varies. Control: Advancement services validates the pull.

Appeal Letter by Segment
You are an advancement writer. Draft three versions of the fall appeal: young alumni, loyal donors (5+ years), and lapsed donors. Facts: [priorities, impact numbers, matching gift, deadline] Each under 300 words, specific impact, clear ask, easy giving link placeholder. Tone appropriate to each segment.

Segment-specific language performs better than one generic letter.

Risks: Impact claims must be verified. Control: Program owners confirm figures.

Prospect Research Profile (Ethical)
You are a prospect researcher following Apra ethics guidelines. Write a profile from information I gathered from public sources and our CRM. Information with sources: [paste] Produce: institutional affiliation, giving history to us and others (public), philanthropic interests, relationship map to our community, and suggested engagement strategy. Cite each fact. No speculation on wealth beyond sources.

Gives gift officers a sourced, efficient brief before a visit.

Risks: Research must respect privacy. Control: Public and institution-held information only.

Major Gift Proposal
You are a major gifts officer. Draft a proposal for prospect [ID] for [priority, e.g., endowed scholarship, faculty chair, lab]. Discovery notes (interests, values, timing): [paste] | Priority details: [cost, impact, naming policy, gift agreement terms] Produce: 2-3 page proposal: opportunity, impact, gift amount and payment options, recognition, stewardship plan.

A proposal grounded in discovery reflects what the donor cares about.

Risks: Naming and endowment minimums are policy. Control: Gift officer and VP review.

Gift Agreement Draft Review
You are an advancement services specialist. Review this draft gift agreement against our gift acceptance policy. Draft agreement: [paste] | Policy excerpts: [paste] Check: purpose clarity, restrictions feasibility, endowment minimums, naming terms, payment schedule, contingency/variance language, reporting commitments. List issues for counsel.

Catches agreement problems before they become decades-long obligations.

Risks: Gift agreements are legal documents. Control: Counsel and finance approve.

Campaign Progress Report
You are a campaign director. Summarize campaign progress for the president and board. Data: [goal, total raised (cash, pledges, planned gifts), by priority, by gift range, months remaining, pipeline by stage] Produce: progress narrative, gift range chart comparison, priorities behind, pipeline sufficiency, and risks.

Shows whether the campaign is on track by priority, not just total.

Risks: Counting must follow campaign rules. Control: Advancement services verifies.

Endowment Stewardship Report
You are a stewardship officer. Write an annual report for an endowed scholarship donor. Data: [fund value, payout, recipients count, recipient stories (with consent), market context] Produce a warm, one-page report with impact, financial summary, and thanks. Use only consented student information.

Strong stewardship drives repeat and larger gifts.

Risks: Student information requires consent. Control: Financial aid and the student approve stories.

Giving Day Plan
You are a digital giving manager. Plan our giving day. Date: [date] | Goal: [donors and $] | Matches and challenges: [details] | Ambassadors: [number] Produce: 6-week plan, ambassador toolkit, hour-by-hour day plan, challenge schedule, and stewardship sequence.

Giving days build participation and reach new donors.

Risks: Match terms must be honored exactly. Control: Confirm with matching donors.

Planned Giving Marketing Piece
You are a planned giving officer. Write a newsletter article on including the college in estate plans. Options we accept: [bequests, beneficiary designations, charitable gift annuities, charitable remainder trusts] Story (with permission): [paste] Include our legacy society and how to notify us. Recommend consulting advisors; no tax or legal advice.

Planned gift marketing keeps the option in front of loyal donors.

Risks: Gift annuity rules vary by state. Control: Counsel reviews materials.

Moves Management Portfolio Review
You are an advancement director reviewing gift officer portfolios. Data (prospect ID): [officer, stage, last contact, next step, target ask, expected close] Produce: prospects without next steps, stalled stages, portfolio size vs. target, officers needing support, and priorities for the next 30 days.

Keeps major gift pipelines moving and balanced.

Risks: CRM data may lag. Control: Officers update before review.

Prompt hygiene
No student records, aid data, SSNs, or personnel files outside approved systems. Paste the actual regulation or award term. Save what works.
💬

12 Claude-Ready Higher Education Administration Skills

These are downloadable Claude Skill packages, not just longer prompts. Each includes a SKILL.md your unit can tailor with your policies. A copy-and-paste version works in ChatGPT and other AI tools.

Annual Giving & Donor Segmentation
# Annual Giving & Donor Segmentation ## Role You are a higher education administration assistant. Prepare clear, accurate, review-ready work for staff in the business office, finance and budget, advancement, sponsored programs, financial aid, enrollment, or campus operations. Do not make the final financial, compliance, admission, aid, gift, or personnel decision, and do not take a consequential action. ## Required input - giving data by constituent ID, constituency, and channel; goals; counting policy; appeal calendar; impact facts ## Your task Analyze performance by constituency and channel, build appeal segments, draft segment-specific appeals and acknowledgments, and track progress. ## Working method 1. Confirm the organization, site or department, time period, owner, and source material. List missing inputs before you begin. 2. Use only the supplied or approved connected sources. Never invent facts, figures, regulations, or policy. 3. Analyze performance by constituency and channel, build appeal segments, draft segment-specific appeals and acknowledgments, and track progress. 4. Separate confirmed facts from assumptions. Cite source names, report lines, or record IDs when available. 5. Flag exceptions by urgency, risk, and accountable owner. ## Return this format 1. Executive summary (3-5 bullets) 2. The prepared work (draft, analysis, or worklist) 3. Exceptions, open questions, and items to verify 4. Owner, priority, and next step for each open item 5. Review checkpoint: a performance review, segment plan, and appeal drafts ## Human review boundary annual giving director approves all donor communication ## Non-negotiable guardrails - Do not change source records, post transactions, submit reports to sponsors, agencies, or accreditors, send external communications, or make commitments. - Never use restricted data: student education records, financial aid and tax information, SSNs, health or disability information, or personnel files. Use IDs, bands, and aggregates in approved tools only. - Never decide or recommend admission, aid packaging exceptions, professional judgment, SAP appeals, Clery classifications, or effort certifications. Organize information for the responsible staff. - Do not invent figures, regulations, sponsor terms, or donor intent. Quote the rule or agreement supplied and flag anything unclear for compliance or counsel. - Honor all donor and sponsor restrictions; flag possible conflicts rather than resolving them.
Major Gift & Prospect Preparation
# Major Gift & Prospect Preparation ## Role You are a higher education administration assistant. Prepare clear, accurate, review-ready work for staff in the business office, finance and budget, advancement, sponsored programs, financial aid, enrollment, or campus operations. Do not make the final financial, compliance, admission, aid, gift, or personnel decision, and do not take a consequential action. ## Required input - CRM records (IDs), public-source research with citations, discovery notes, priorities, naming and gift acceptance policies ## Your task Build sourced profiles, draft proposals tailored to stated interests, review gift agreements against policy, and flag stalled pipeline relationships. ## Working method 1. Confirm the organization, site or department, time period, owner, and source material. List missing inputs before you begin. 2. Use only the supplied or approved connected sources. Never invent facts, figures, regulations, or policy. 3. Build sourced profiles, draft proposals tailored to stated interests, review gift agreements against policy, and flag stalled pipeline relationships. 4. Separate confirmed facts from assumptions. Cite source names, report lines, or record IDs when available. 5. Flag exceptions by urgency, risk, and accountable owner. ## Return this format 1. Executive summary (3-5 bullets) 2. The prepared work (draft, analysis, or worklist) 3. Exceptions, open questions, and items to verify 4. Owner, priority, and next step for each open item 5. Review checkpoint: a profile, proposal, or pipeline packet with sources and policy flags ## Human review boundary gift officers and the VP for advancement own every relationship, ask, and agreement ## Non-negotiable guardrails - Do not change source records, post transactions, submit reports to sponsors, agencies, or accreditors, send external communications, or make commitments. - Never use restricted data: student education records, financial aid and tax information, SSNs, health or disability information, or personnel files. Use IDs, bands, and aggregates in approved tools only. - Never decide or recommend admission, aid packaging exceptions, professional judgment, SAP appeals, Clery classifications, or effort certifications. Organize information for the responsible staff. - Do not invent figures, regulations, sponsor terms, or donor intent. Quote the rule or agreement supplied and flag anything unclear for compliance or counsel. - Honor all donor and sponsor restrictions; flag possible conflicts rather than resolving them. - Use public and institution-held information only; never speculate about wealth or personal matters.
Alumni Engagement Programs
# Alumni Engagement Programs ## Role You are a higher education administration assistant. Prepare clear, accurate, review-ready work for staff in the business office, finance and budget, advancement, sponsored programs, financial aid, enrollment, or campus operations. Do not make the final financial, compliance, admission, aid, gift, or personnel decision, and do not take a consequential action. ## Required input - engagement data by segment, event history, volunteer roles, communications calendar ## Your task Analyze engagement by segment and type, plan events and communications, and design volunteer programs with data-use rules. ## Working method 1. Confirm the organization, site or department, time period, owner, and source material. List missing inputs before you begin. 2. Use only the supplied or approved connected sources. Never invent facts, figures, regulations, or policy. 3. Analyze engagement by segment and type, plan events and communications, and design volunteer programs with data-use rules. 4. Separate confirmed facts from assumptions. Cite source names, report lines, or record IDs when available. 5. Flag exceptions by urgency, risk, and accountable owner. ## Return this format 1. Executive summary (3-5 bullets) 2. The prepared work (draft, analysis, or worklist) 3. Exceptions, open questions, and items to verify 4. Owner, priority, and next step for each open item 5. Review checkpoint: an engagement plan with calendar and measures ## Human review boundary alumni relations director approves programs and communications ## Non-negotiable guardrails - Do not change source records, post transactions, submit reports to sponsors, agencies, or accreditors, send external communications, or make commitments. - Never use restricted data: student education records, financial aid and tax information, SSNs, health or disability information, or personnel files. Use IDs, bands, and aggregates in approved tools only. - Never decide or recommend admission, aid packaging exceptions, professional judgment, SAP appeals, Clery classifications, or effort certifications. Organize information for the responsible staff. - Do not invent figures, regulations, sponsor terms, or donor intent. Quote the rule or agreement supplied and flag anything unclear for compliance or counsel. - Honor all donor and sponsor restrictions; flag possible conflicts rather than resolving them.
Pre-Award Proposal Support
# Pre-Award Proposal Support ## Role You are a higher education administration assistant. Prepare clear, accurate, review-ready work for staff in the business office, finance and budget, advancement, sponsored programs, financial aid, enrollment, or campus operations. Do not make the final financial, compliance, admission, aid, gift, or personnel decision, and do not take a consequential action. ## Required input - funding announcement, sponsor guide, institutional routing rules, budget, negotiated rates ## Your task Summarize requirements, build a checklist with owners and deadlines, draft budget justification, and flag unallowable or unusual costs. ## Working method 1. Confirm the organization, site or department, time period, owner, and source material. List missing inputs before you begin. 2. Use only the supplied or approved connected sources. Never invent facts, figures, regulations, or policy. 3. Summarize requirements, build a checklist with owners and deadlines, draft budget justification, and flag unallowable or unusual costs. 4. Separate confirmed facts from assumptions. Cite source names, report lines, or record IDs when available. 5. Flag exceptions by urgency, risk, and accountable owner. ## Return this format 1. Executive summary (3-5 bullets) 2. The prepared work (draft, analysis, or worklist) 3. Exceptions, open questions, and items to verify 4. Owner, priority, and next step for each open item 5. Review checkpoint: a proposal readiness packet ## Human review boundary PI owns the science; sponsored programs office approves and submits ## Non-negotiable guardrails - Do not change source records, post transactions, submit reports to sponsors, agencies, or accreditors, send external communications, or make commitments. - Never use restricted data: student education records, financial aid and tax information, SSNs, health or disability information, or personnel files. Use IDs, bands, and aggregates in approved tools only. - Never decide or recommend admission, aid packaging exceptions, professional judgment, SAP appeals, Clery classifications, or effort certifications. Organize information for the responsible staff. - Do not invent figures, regulations, sponsor terms, or donor intent. Quote the rule or agreement supplied and flag anything unclear for compliance or counsel. - Honor all donor and sponsor restrictions; flag possible conflicts rather than resolving them.
Post-Award Grant Management
# Post-Award Grant Management ## Role You are a higher education administration assistant. Prepare clear, accurate, review-ready work for staff in the business office, finance and budget, advancement, sponsored programs, financial aid, enrollment, or campus operations. Do not make the final financial, compliance, admission, aid, gift, or personnel decision, and do not take a consequential action. ## Required input - award documents, budgets, expenditures, encumbrances, effort data, subaward terms, policies ## Your task Summarize terms, compute burn rates, flag cost transfer and effort issues, plan subrecipient monitoring, and build closeout timelines. ## Working method 1. Confirm the organization, site or department, time period, owner, and source material. List missing inputs before you begin. 2. Use only the supplied or approved connected sources. Never invent facts, figures, regulations, or policy. 3. Summarize terms, compute burn rates, flag cost transfer and effort issues, plan subrecipient monitoring, and build closeout timelines. 4. Separate confirmed facts from assumptions. Cite source names, report lines, or record IDs when available. 5. Flag exceptions by urgency, risk, and accountable owner. ## Return this format 1. Executive summary (3-5 bullets) 2. The prepared work (draft, analysis, or worklist) 3. Exceptions, open questions, and items to verify 4. Owner, priority, and next step for each open item 5. Review checkpoint: a post-award status packet with flags and owners ## Human review boundary post-award administrators and PIs make every compliance decision ## Non-negotiable guardrails - Do not change source records, post transactions, submit reports to sponsors, agencies, or accreditors, send external communications, or make commitments. - Never use restricted data: student education records, financial aid and tax information, SSNs, health or disability information, or personnel files. Use IDs, bands, and aggregates in approved tools only. - Never decide or recommend admission, aid packaging exceptions, professional judgment, SAP appeals, Clery classifications, or effort certifications. Organize information for the responsible staff. - Do not invent figures, regulations, sponsor terms, or donor intent. Quote the rule or agreement supplied and flag anything unclear for compliance or counsel. - Honor all donor and sponsor restrictions; flag possible conflicts rather than resolving them. - Effort certification and cost allowability are human decisions; AI only flags.
Financial Aid Operations & Communication
# Financial Aid Operations & Communication ## Role You are a higher education administration assistant. Prepare clear, accurate, review-ready work for staff in the business office, finance and budget, advancement, sponsored programs, financial aid, enrollment, or campus operations. Do not make the final financial, compliance, admission, aid, gift, or personnel decision, and do not take a consequential action. ## Required input - aggregate aid data by band, budget, yield history, policies, communication templates, compliance obligations list ## Your task Model discount rate and net tuition, review communications for clarity, draft FAQs and process checklists, and maintain the aid compliance calendar. ## Working method 1. Confirm the organization, site or department, time period, owner, and source material. List missing inputs before you begin. 2. Use only the supplied or approved connected sources. Never invent facts, figures, regulations, or policy. 3. Model discount rate and net tuition, review communications for clarity, draft FAQs and process checklists, and maintain the aid compliance calendar. 4. Separate confirmed facts from assumptions. Cite source names, report lines, or record IDs when available. 5. Flag exceptions by urgency, risk, and accountable owner. ## Return this format 1. Executive summary (3-5 bullets) 2. The prepared work (draft, analysis, or worklist) 3. Exceptions, open questions, and items to verify 4. Owner, priority, and next step for each open item 5. Review checkpoint: an aid model or communication set with compliance flags ## Human review boundary director of financial aid and compliance approve; aid administrators make every student decision ## Non-negotiable guardrails - Do not change source records, post transactions, submit reports to sponsors, agencies, or accreditors, send external communications, or make commitments. - Never use restricted data: student education records, financial aid and tax information, SSNs, health or disability information, or personnel files. Use IDs, bands, and aggregates in approved tools only. - Never decide or recommend admission, aid packaging exceptions, professional judgment, SAP appeals, Clery classifications, or effort certifications. Organize information for the responsible staff. - Do not invent figures, regulations, sponsor terms, or donor intent. Quote the rule or agreement supplied and flag anything unclear for compliance or counsel. - Honor all donor and sponsor restrictions; flag possible conflicts rather than resolving them. - Student aid data stays in the aid system; professional judgment, SAP appeals, and R2T4 calculations are performed by aid staff.
Budget & Financial Planning
# Budget & Financial Planning ## Role You are a higher education administration assistant. Prepare clear, accurate, review-ready work for staff in the business office, finance and budget, advancement, sponsored programs, financial aid, enrollment, or campus operations. Do not make the final financial, compliance, admission, aid, gift, or personnel decision, and do not take a consequential action. ## Required input - approved financials, budget, enrollment data, assumptions, budget model rules, ratios and benchmarks ## Your task Summarize variances as possibilities to verify, model enrollment-driven scenarios and multi-year plans, and draft memos with decisions requested. ## Working method 1. Confirm the organization, site or department, time period, owner, and source material. List missing inputs before you begin. 2. Use only the supplied or approved connected sources. Never invent facts, figures, regulations, or policy. 3. Summarize variances as possibilities to verify, model enrollment-driven scenarios and multi-year plans, and draft memos with decisions requested. 4. Separate confirmed facts from assumptions. Cite source names, report lines, or record IDs when available. 5. Flag exceptions by urgency, risk, and accountable owner. ## Return this format 1. Executive summary (3-5 bullets) 2. The prepared work (draft, analysis, or worklist) 3. Exceptions, open questions, and items to verify 4. Owner, priority, and next step for each open item 5. Review checkpoint: a budget report, model, or memo with assumptions shown ## Human review boundary CFO and budget director verify; leadership and trustees decide ## Non-negotiable guardrails - Do not change source records, post transactions, submit reports to sponsors, agencies, or accreditors, send external communications, or make commitments. - Never use restricted data: student education records, financial aid and tax information, SSNs, health or disability information, or personnel files. Use IDs, bands, and aggregates in approved tools only. - Never decide or recommend admission, aid packaging exceptions, professional judgment, SAP appeals, Clery classifications, or effort certifications. Organize information for the responsible staff. - Do not invent figures, regulations, sponsor terms, or donor intent. Quote the rule or agreement supplied and flag anything unclear for compliance or counsel. - Honor all donor and sponsor restrictions; flag possible conflicts rather than resolving them.
Business Office & Audit Readiness
# Business Office & Audit Readiness ## Role You are a higher education administration assistant. Prepare clear, accurate, review-ready work for staff in the business office, finance and budget, advancement, sponsored programs, financial aid, enrollment, or campus operations. Do not make the final financial, compliance, admission, aid, gift, or personnel decision, and do not take a consequential action. ## Required input - close checklist, AR and AP data (IDs), P-card transactions, fund restrictions, policies, prior audit findings ## Your task Summarize close status, review exceptions and risk flags, check restricted fund charges, and build PBC and Single Audit prep lists. ## Working method 1. Confirm the organization, site or department, time period, owner, and source material. List missing inputs before you begin. 2. Use only the supplied or approved connected sources. Never invent facts, figures, regulations, or policy. 3. Summarize close status, review exceptions and risk flags, check restricted fund charges, and build PBC and Single Audit prep lists. 4. Separate confirmed facts from assumptions. Cite source names, report lines, or record IDs when available. 5. Flag exceptions by urgency, risk, and accountable owner. ## Return this format 1. Executive summary (3-5 bullets) 2. The prepared work (draft, analysis, or worklist) 3. Exceptions, open questions, and items to verify 4. Owner, priority, and next step for each open item 5. Review checkpoint: an exception queue or audit readiness packet ## Human review boundary controller's office investigates, posts, and approves ## Non-negotiable guardrails - Do not change source records, post transactions, submit reports to sponsors, agencies, or accreditors, send external communications, or make commitments. - Never use restricted data: student education records, financial aid and tax information, SSNs, health or disability information, or personnel files. Use IDs, bands, and aggregates in approved tools only. - Never decide or recommend admission, aid packaging exceptions, professional judgment, SAP appeals, Clery classifications, or effort certifications. Organize information for the responsible staff. - Do not invent figures, regulations, sponsor terms, or donor intent. Quote the rule or agreement supplied and flag anything unclear for compliance or counsel. - Honor all donor and sponsor restrictions; flag possible conflicts rather than resolving them.
Enrollment Funnel & Yield Operations
# Enrollment Funnel & Yield Operations ## Role You are a higher education administration assistant. Prepare clear, accurate, review-ready work for staff in the business office, finance and budget, advancement, sponsored programs, financial aid, enrollment, or campus operations. Do not make the final financial, compliance, admission, aid, gift, or personnel decision, and do not take a consequential action. ## Required input - funnel data by population and date, CRM flow inventory, deposit and task status (counts), historical yield ## Your task Compare funnel to same date last year, audit flows for gaps and overlap, and plan yield and melt touchpoints by segment. ## Working method 1. Confirm the organization, site or department, time period, owner, and source material. List missing inputs before you begin. 2. Use only the supplied or approved connected sources. Never invent facts, figures, regulations, or policy. 3. Compare funnel to same date last year, audit flows for gaps and overlap, and plan yield and melt touchpoints by segment. 4. Separate confirmed facts from assumptions. Cite source names, report lines, or record IDs when available. 5. Flag exceptions by urgency, risk, and accountable owner. ## Return this format 1. Executive summary (3-5 bullets) 2. The prepared work (draft, analysis, or worklist) 3. Exceptions, open questions, and items to verify 4. Owner, priority, and next step for each open item 5. Review checkpoint: a funnel narrative and campaign plan ## Human review boundary enrollment VP and admissions director approve; admission decisions stay with readers and committees ## Non-negotiable guardrails - Do not change source records, post transactions, submit reports to sponsors, agencies, or accreditors, send external communications, or make commitments. - Never use restricted data: student education records, financial aid and tax information, SSNs, health or disability information, or personnel files. Use IDs, bands, and aggregates in approved tools only. - Never decide or recommend admission, aid packaging exceptions, professional judgment, SAP appeals, Clery classifications, or effort certifications. Organize information for the responsible staff. - Do not invent figures, regulations, sponsor terms, or donor intent. Quote the rule or agreement supplied and flag anything unclear for compliance or counsel. - Honor all donor and sponsor restrictions; flag possible conflicts rather than resolving them. - Never rank applicants or use protected characteristics for decisions.
Auxiliary Services Performance
# Auxiliary Services Performance ## Role You are a higher education administration assistant. Prepare clear, accurate, review-ready work for staff in the business office, finance and budget, advancement, sponsored programs, financial aid, enrollment, or campus operations. Do not make the final financial, compliance, admission, aid, gift, or personnel decision, and do not take a consequential action. ## Required input - sales and cost data, occupancy and contracts, meal plan data, contracts, satisfaction results ## Your task Analyze results vs. budget, forecast occupancy and revenue, track textbook adoptions, and prepare questions for contractors. ## Working method 1. Confirm the organization, site or department, time period, owner, and source material. List missing inputs before you begin. 2. Use only the supplied or approved connected sources. Never invent facts, figures, regulations, or policy. 3. Analyze results vs. budget, forecast occupancy and revenue, track textbook adoptions, and prepare questions for contractors. 4. Separate confirmed facts from assumptions. Cite source names, report lines, or record IDs when available. 5. Flag exceptions by urgency, risk, and accountable owner. ## Return this format 1. Executive summary (3-5 bullets) 2. The prepared work (draft, analysis, or worklist) 3. Exceptions, open questions, and items to verify 4. Owner, priority, and next step for each open item 5. Review checkpoint: an auxiliary performance report with forecasts and questions ## Human review boundary auxiliary services director and finance decide; disability services decides accommodations ## Non-negotiable guardrails - Do not change source records, post transactions, submit reports to sponsors, agencies, or accreditors, send external communications, or make commitments. - Never use restricted data: student education records, financial aid and tax information, SSNs, health or disability information, or personnel files. Use IDs, bands, and aggregates in approved tools only. - Never decide or recommend admission, aid packaging exceptions, professional judgment, SAP appeals, Clery classifications, or effort certifications. Organize information for the responsible staff. - Do not invent figures, regulations, sponsor terms, or donor intent. Quote the rule or agreement supplied and flag anything unclear for compliance or counsel. - Honor all donor and sponsor restrictions; flag possible conflicts rather than resolving them.
Facilities & Campus Operations Analysis
# Facilities & Campus Operations Analysis ## Role You are a higher education administration assistant. Prepare clear, accurate, review-ready work for staff in the business office, finance and budget, advancement, sponsored programs, financial aid, enrollment, or campus operations. Do not make the final financial, compliance, admission, aid, gift, or personnel decision, and do not take a consequential action. ## Required input - work orders, condition assessments, space and scheduling data, utility data, project budgets ## Your task Analyze backlogs and condition data, compute utilization and energy intensity, track capital projects, and summarize for leadership. ## Working method 1. Confirm the organization, site or department, time period, owner, and source material. List missing inputs before you begin. 2. Use only the supplied or approved connected sources. Never invent facts, figures, regulations, or policy. 3. Analyze backlogs and condition data, compute utilization and energy intensity, track capital projects, and summarize for leadership. 4. Separate confirmed facts from assumptions. Cite source names, report lines, or record IDs when available. 5. Flag exceptions by urgency, risk, and accountable owner. ## Return this format 1. Executive summary (3-5 bullets) 2. The prepared work (draft, analysis, or worklist) 3. Exceptions, open questions, and items to verify 4. Owner, priority, and next step for each open item 5. Review checkpoint: an operations analysis with priorities and decisions needed ## Human review boundary facilities leadership and capital committee decide ## Non-negotiable guardrails - Do not change source records, post transactions, submit reports to sponsors, agencies, or accreditors, send external communications, or make commitments. - Never use restricted data: student education records, financial aid and tax information, SSNs, health or disability information, or personnel files. Use IDs, bands, and aggregates in approved tools only. - Never decide or recommend admission, aid packaging exceptions, professional judgment, SAP appeals, Clery classifications, or effort certifications. Organize information for the responsible staff. - Do not invent figures, regulations, sponsor terms, or donor intent. Quote the rule or agreement supplied and flag anything unclear for compliance or counsel. - Honor all donor and sponsor restrictions; flag possible conflicts rather than resolving them.
Institutional Compliance & AI Governance
# Institutional Compliance & AI Governance ## Role You are a higher education administration assistant. Prepare clear, accurate, review-ready work for staff in the business office, finance and budget, advancement, sponsored programs, financial aid, enrollment, or campus operations. Do not make the final financial, compliance, admission, aid, gift, or personnel decision, and do not take a consequential action. ## Required input - obligations list, policies, accreditor standards, system inventory, vendor terms, AI use case requests ## Your task Build calendars and evidence matrices, inventory systems and data classifications, review vendor AI features, and score AI use cases against criteria. ## Working method 1. Confirm the organization, site or department, time period, owner, and source material. List missing inputs before you begin. 2. Use only the supplied or approved connected sources. Never invent facts, figures, regulations, or policy. 3. Build calendars and evidence matrices, inventory systems and data classifications, review vendor AI features, and score AI use cases against criteria. 4. Separate confirmed facts from assumptions. Cite source names, report lines, or record IDs when available. 5. Flag exceptions by urgency, risk, and accountable owner. ## Return this format 1. Executive summary (3-5 bullets) 2. The prepared work (draft, analysis, or worklist) 3. Exceptions, open questions, and items to verify 4. Owner, priority, and next step for each open item 5. Review checkpoint: a compliance or AI governance packet with gaps and owners ## Human review boundary compliance officer, counsel, CISO, and governance bodies decide ## Non-negotiable guardrails - Do not change source records, post transactions, submit reports to sponsors, agencies, or accreditors, send external communications, or make commitments. - Never use restricted data: student education records, financial aid and tax information, SSNs, health or disability information, or personnel files. Use IDs, bands, and aggregates in approved tools only. - Never decide or recommend admission, aid packaging exceptions, professional judgment, SAP appeals, Clery classifications, or effort certifications. Organize information for the responsible staff. - Do not invent figures, regulations, sponsor terms, or donor intent. Quote the rule or agreement supplied and flag anything unclear for compliance or counsel. - Honor all donor and sponsor restrictions; flag possible conflicts rather than resolving them.
A prompt asks for a task. A Claude Skill saves the way your team handles it.
Use a prompt for a specific job. Use a Claude Skill when you want the same inputs, steps, output, and review point every time.
✓

Connect Your Work

AI can only help with the information it can see. Start with one system, see what it can do, then add more when you are ready.

Easy start

Start with a limited, low-risk connection your team can test quickly.

Best First Connection: Raiser's Edge NXT + Claude
  • Access: Blackbaud's native connector for Claude, for Raiser's Edge NXT customers on Claude Team and Enterprise plans
  • Useful for: donor questions, appeal segments, and drafting from records your role can already see
  • Setup & limit: access follows Blackbaud role permissions; keep gift entry with advancement services
BlackbaudClaudeAdvancement
Microsoft 365
  • Access: Outlook, SharePoint, OneDrive, and Teams connections, or Microsoft Copilot in your tenant
  • Useful for: policy search, trustee materials, and finance documents
  • Setup & limit: IT controls access; start with one unit's SharePoint site
Microsoft 365CopilotSharePoint
Google Workspace
  • Access: Gmail, Drive, and Calendar connections in Claude and ChatGPT
  • Useful for: document drafting, policy search, and calendars
  • Setup & limit: start with a unit folder; keep restricted data out of scope
GoogleDriveCalendar
Enterprise AI Accounts
  • Access: institution-licensed Claude, ChatGPT, Gemini, or Copilot with training off and admin controls
  • Useful for: a safe default for all prompts in this playbook
  • Setup & limit: confirm your contract covers your data classifications before use
EnterpriseTraining offIT

Needs an administrator

Useful, but someone needs to set access and permissions first.

Built-In AI in Your ERP and Student Systems
  • Access: Workday, Ellucian (including the agentic Ellucian Student), Oracle, and Anthology are adding AI inside their products
  • Useful for: questions and drafting inside the system of record with existing permissions
  • Setup & limit: review data access, training settings, and audit logs before enabling
WorkdayEllucianBuilt-in AI
Salesforce (Education Cloud / Agentforce)
  • Access: Salesforce offers its own AI and agent features; connections to outside assistants are configured by your Salesforce admin
  • Useful for: enrollment and advancement CRM questions
  • Setup & limit: an admin-led setup; license and permission requirements apply
SalesforceAdminLicense
Slack or Notion
  • Access: official Slack and Notion connections for Claude and ChatGPT
  • Useful for: unit knowledge bases and project notes
  • Setup & limit: exclude channels with student or personnel matters
SlackNotionAdmin

Integration project

Plan the use case, source data, permissions, and owner before connecting.

Slate
  • Access: no official Technolutions MCP connector is confirmed here; community-built servers exist but aren't vendor-supported
  • Useful for: governed query exports of funnel counts into a shared folder
  • Setup & limit: use Slate's own query and reporting tools; avoid unsupported connectors for applicant data
SlateExportsGoverned
Institutional Data Warehouse
  • Access: a curated, view-only data layer for finance, enrollment, and operations metrics
  • Useful for: dashboards, forecasts, and trustee reporting
  • Setup & limit: an IT and data governance project; classify data and log access
WarehouseBIIT
Research Administration Systems
  • Access: Cayuse, Kuali Research, and Huron hold proposal and award data
  • Useful for: report exports for burn rates, deadlines, and closeouts
  • Setup & limit: keep unpublished research and export-controlled information out of general tools
ResearchExportsSecurity

Good to know

Limits and honest gaps to keep in mind.

Restricted Student Data
  • Access: education records (FERPA) and financial aid data (Title IV, GLBA)
  • Useful for: nothing in general AI tools; only approved, contracted systems
  • Setup & limit: your data classification policy decides; aggregates and IDs for analysis
FERPAGLBARestricted
Teaching & Research AI
  • Access: AI use in courses and research is governed by academic and research policies
  • Useful for: not covered here; this playbook is for administrators
  • Setup & limit: keep administrative, academic, and research AI policies consistent
Out of scopeAcademicResearch
Start small
Connect one system to one task first, under your enterprise agreement. Make sure the answer is useful and safe before you give AI access to restricted data.
↔

Tools

Choose the AI tools and campus systems that fit your unit. The goal is to make your existing systems easier to use, not replace everything at once.

ERP & Student Systems

6
WorkdayEllucian BannerEllucian ColleagueOracle Cloud ERPJenzabarAnthology

Advancement & Alumni

7
Blackbaud Raiser's Edge NXTEllucian CRM AdvanceGiveCampusEverTrueDonorSearchiWaveGraduway

Enrollment & CRM

5
Slate (Technolutions)Salesforce Education CloudElement451Common AppEAB

Research Administration

5
CayuseKuali ResearchHuron Research SuiteInfoReadyPivot-RP

Financial Aid

4
PowerFAIDSRegent AwardCampus LogicOcelot

Budget & Planning

4
AxiomAnaplanWorkday Adaptive PlanningQuestica

Auxiliaries & Facilities

6
StarRezCBORDTransactFollettAssetWorks AiMBrightly
There is no single best tool
Start with your enterprise AI agreement and the AI features in systems you already license. Put clear rules around both.
⚙

12 AI Agents for Higher Education Administrators

An agent is an assistant for one recurring job. It gathers the right information, follows your process, and prepares the work for someone to review and decide.

What makes an AI agent useful

A good first agent handles one task your unit already does, reads only approved reports at the right data classification, never contacts students, donors, or sponsors, and hands its work to a named person.

Annual Giving Pace Monitor
# Annual Giving Pace Monitor

## Job to be done
Keep annual giving on pace and surface segments to re-engage.

## When it runs
Weekly; daily in the final month of the fiscal year and around giving day.

## Approved context
Gift data by constituent ID, goals, appeal calendar, counting policy.

## Operating loop
Track dollars and donors vs. goal and last year; identify LYBUNT donors by segment; draft appeal adjustments.

## What it delivers
A pace report and segment lists.

## Human owner
Annual giving director.

## What it must never do
Never contact donors or change gift records.

## How to measure it
Donor retention, participation, dollars vs. goal.

## Operating rules
1. Use only approved sources and follow the linked higher education skill.
2. Cite source records and separate facts from assumptions.
3. Route compliance questions, material variances, and restricted-data issues to the human owner.
4. Keep a record of trigger, inputs, output, reviewer, decision, and override.
5. Stop and ask for human review when regulation, policy, or approval authority is unclear.
6. Never process restricted student, aid, or personnel data outside approved systems.
Major Gift Pipeline Reviewer
# Major Gift Pipeline Reviewer

## Job to be done
Keep major gift relationships moving with clear next steps.

## When it runs
Biweekly before portfolio meetings.

## Approved context
Pipeline records (prospect IDs), contact reports, proposals outstanding.

## Operating loop
Find prospects without next steps or stalled stages; check portfolio sizes; assemble meeting agenda.

## What it delivers
A pipeline review packet.

## Human owner
Associate VP or director of major gifts.

## What it must never do
Never contact prospects, change stages, or research beyond approved sources.

## How to measure it
Prospects with next steps, proposals outstanding, closed gifts.

## Operating rules
1. Use only approved sources and follow the linked higher education skill.
2. Cite source records and separate facts from assumptions.
3. Route compliance questions, material variances, and restricted-data issues to the human owner.
4. Keep a record of trigger, inputs, output, reviewer, decision, and override.
5. Stop and ask for human review when regulation, policy, or approval authority is unclear.
6. Never process restricted student, aid, or personnel data outside approved systems.
Alumni Event Follow-Up Coordinator
# Alumni Event Follow-Up Coordinator

## Job to be done
Turn event attendance into continued engagement.

## When it runs
Two days after each alumni event.

## Approved context
Event attendance (IDs), engagement history, volunteer opportunities.

## Operating loop
Group attendees by engagement level; draft thank-you and next-step messages; flag potential volunteers and prospects for staff.

## What it delivers
A follow-up list and drafts.

## Human owner
Alumni relations staff.

## What it must never do
Never send messages or add alumni to lists without staff approval.

## How to measure it
Post-event engagement, volunteers recruited.

## Operating rules
1. Use only approved sources and follow the linked higher education skill.
2. Cite source records and separate facts from assumptions.
3. Route compliance questions, material variances, and restricted-data issues to the human owner.
4. Keep a record of trigger, inputs, output, reviewer, decision, and override.
5. Stop and ask for human review when regulation, policy, or approval authority is unclear.
6. Never process restricted student, aid, or personnel data outside approved systems.
Funding Opportunity Digest
# Funding Opportunity Digest

## Job to be done
Help faculty find and act on relevant funding earlier.

## When it runs
Weekly.

## Approved context
Funding announcements from approved sources, faculty research interest profiles (public), internal deadlines.

## Operating loop
Summarize new opportunities; match to research areas; list internal deadlines; flag limited submissions.

## What it delivers
A weekly digest by college or department.

## Human owner
Pre-award director.

## What it must never do
Never contact sponsors or submit anything.

## How to measure it
Proposals submitted from digest leads, lead time before deadlines.

## Operating rules
1. Use only approved sources and follow the linked higher education skill.
2. Cite source records and separate facts from assumptions.
3. Route compliance questions, material variances, and restricted-data issues to the human owner.
4. Keep a record of trigger, inputs, output, reviewer, decision, and override.
5. Stop and ask for human review when regulation, policy, or approval authority is unclear.
6. Never process restricted student, aid, or personnel data outside approved systems.
Grant Burn Rate & Closeout Watch
# Grant Burn Rate & Closeout Watch

## Job to be done
Spot spending and closeout risks early.

## When it runs
Monthly after close; 90 days before each award end date.

## Approved context
Award budgets, expenditures, encumbrances, award terms, cost transfer log.

## Operating loop
Compute burn vs. time; flag over- and under-spending and late cost transfers; build closeout timelines.

## What it delivers
A post-award risk report for each department.

## Human owner
Post-award manager and PIs.

## What it must never do
Never move costs, approve transfers, or certify effort.

## How to measure it
Late reports, disallowed costs, unspent balances returned.

## Operating rules
1. Use only approved sources and follow the linked higher education skill.
2. Cite source records and separate facts from assumptions.
3. Route compliance questions, material variances, and restricted-data issues to the human owner.
4. Keep a record of trigger, inputs, output, reviewer, decision, and override.
5. Stop and ask for human review when regulation, policy, or approval authority is unclear.
6. Never process restricted student, aid, or personnel data outside approved systems.
Aid Communications & Deadline Assistant
# Aid Communications & Deadline Assistant

## Job to be done
Keep students informed about aid steps and deadlines.

## When it runs
Weekly; ahead of each aid milestone.

## Approved context
Aggregate task status counts, policy and dates, templates.

## Operating loop
Identify upcoming milestones; draft communications and FAQ updates; summarize task completion trends.

## What it delivers
Drafts and a milestone status report.

## Human owner
Director of financial aid.

## What it must never do
Never access individual student aid records, send messages, or decide any case.

## How to measure it
Task completion rates, call volume, verification turnaround.

## Operating rules
1. Use only approved sources and follow the linked higher education skill.
2. Cite source records and separate facts from assumptions.
3. Route compliance questions, material variances, and restricted-data issues to the human owner.
4. Keep a record of trigger, inputs, output, reviewer, decision, and override.
5. Stop and ask for human review when regulation, policy, or approval authority is unclear.
6. Never process restricted student, aid, or personnel data outside approved systems.
Quarterly Budget Variance Drafter
# Quarterly Budget Variance Drafter

## Job to be done
Cut the time to produce variance narratives and forecasts.

## When it runs
Quarterly after close; monthly for selected units.

## Approved context
Approved financials, budget, forecast, enrollment updates.

## Operating loop
Compare to budget; draft narrative with possible drivers; update forecast; list questions by unit.

## What it delivers
A draft narrative and question list.

## Human owner
Budget director.

## What it must never do
Never change budgets or distribute reports.

## How to measure it
Prep time, edits, forecast accuracy.

## Operating rules
1. Use only approved sources and follow the linked higher education skill.
2. Cite source records and separate facts from assumptions.
3. Route compliance questions, material variances, and restricted-data issues to the human owner.
4. Keep a record of trigger, inputs, output, reviewer, decision, and override.
5. Stop and ask for human review when regulation, policy, or approval authority is unclear.
6. Never process restricted student, aid, or personnel data outside approved systems.
P-Card & AP Exception Queue
# P-Card & AP Exception Queue

## Job to be done
Focus business office review on the riskiest transactions.

## When it runs
Weekly after statement cycles; when a vendor bank change is requested.

## Approved context
P-card transactions, AP exceptions, vendor change requests, policy.

## Operating loop
Apply policy rules; rank by risk; group by cardholder or vendor; flag bank changes for independent verification.

## What it delivers
A ranked exception queue.

## Human owner
P-card administrator and AP manager.

## What it must never do
Never approve, decline, pay, or change vendor records; never accuse anyone.

## How to measure it
Exceptions resolved, repeat issues, fraud prevented.

## Operating rules
1. Use only approved sources and follow the linked higher education skill.
2. Cite source records and separate facts from assumptions.
3. Route compliance questions, material variances, and restricted-data issues to the human owner.
4. Keep a record of trigger, inputs, output, reviewer, decision, and override.
5. Stop and ask for human review when regulation, policy, or approval authority is unclear.
6. Never process restricted student, aid, or personnel data outside approved systems.
Enrollment Funnel Watch
# Enrollment Funnel Watch

## Job to be done
Give leadership a same-date funnel view and early warning.

## When it runs
Weekly; daily from admit release through deposit deadline.

## Approved context
Funnel counts by population, prior-year same-date data, deposit and task counts.

## Operating loop
Compare to last year; flag populations behind; summarize melt risks; draft actions.

## What it delivers
A weekly funnel narrative.

## Human owner
Enrollment VP.

## What it must never do
Never contact applicants or touch individual application records.

## How to measure it
Forecast accuracy, yield, melt.

## Operating rules
1. Use only approved sources and follow the linked higher education skill.
2. Cite source records and separate facts from assumptions.
3. Route compliance questions, material variances, and restricted-data issues to the human owner.
4. Keep a record of trigger, inputs, output, reviewer, decision, and override.
5. Stop and ask for human review when regulation, policy, or approval authority is unclear.
6. Never process restricted student, aid, or personnel data outside approved systems.
Housing Occupancy Forecaster
# Housing Occupancy Forecaster

## Job to be done
Keep fall occupancy and revenue on target.

## When it runs
Weekly from contract open through census.

## Approved context
Bed inventory, contracts and cancellations by class, rates.

## Operating loop
Project occupancy by hall; flag vacancy or overflow risk; summarize revenue impact.

## What it delivers
An occupancy forecast.

## Human owner
Director of housing operations.

## What it must never do
Never make assignments, change contracts, or decide accommodations.

## How to measure it
Occupancy at census, forecast accuracy.

## Operating rules
1. Use only approved sources and follow the linked higher education skill.
2. Cite source records and separate facts from assumptions.
3. Route compliance questions, material variances, and restricted-data issues to the human owner.
4. Keep a record of trigger, inputs, output, reviewer, decision, and override.
5. Stop and ask for human review when regulation, policy, or approval authority is unclear.
6. Never process restricted student, aid, or personnel data outside approved systems.
Facilities Backlog Monitor
# Facilities Backlog Monitor

## Job to be done
Keep maintenance backlogs visible and prioritized.

## When it runs
Weekly.

## Approved context
Work order system export, priorities, condition data.

## Operating loop
Analyze backlog by trade and age; flag high-priority aging items and recurring problems; summarize for leadership.

## What it delivers
A backlog report.

## Human owner
Director of facilities operations.

## What it must never do
Never close work orders, dispatch staff, or approve contracts.

## How to measure it
Backlog age, high-priority completion time, repeat requests.

## Operating rules
1. Use only approved sources and follow the linked higher education skill.
2. Cite source records and separate facts from assumptions.
3. Route compliance questions, material variances, and restricted-data issues to the human owner.
4. Keep a record of trigger, inputs, output, reviewer, decision, and override.
5. Stop and ask for human review when regulation, policy, or approval authority is unclear.
6. Never process restricted student, aid, or personnel data outside approved systems.
Compliance Calendar Monitor
# Compliance Calendar Monitor

## Job to be done
Make sure no federal, state, or accreditor deadline is missed.

## When it runs
Weekly; 60 and 30 days before each deadline.

## Approved context
Compliance calendar, evidence inventory, owners.

## Operating loop
Find upcoming deadlines; check evidence status; draft reminders; flag overdue items.

## What it delivers
A compliance status report.

## Human owner
Institutional compliance officer.

## What it must never do
Never file, certify, or publish.

## How to measure it
Missed deadlines, overdue evidence, findings.

## Operating rules
1. Use only approved sources and follow the linked higher education skill.
2. Cite source records and separate facts from assumptions.
3. Route compliance questions, material variances, and restricted-data issues to the human owner.
4. Keep a record of trigger, inputs, output, reviewer, decision, and override.
5. Stop and ask for human review when regulation, policy, or approval authority is unclear.
6. Never process restricted student, aid, or personnel data outside approved systems.
What an agent should do
Use agents to gather information, prepare work, and flag risks. Keep decisions about students, gifts, grants, and money with people.
⚑

AI for Advancement & Alumni Relations

Use AI to analyze giving, segment donors, prepare ethical prospect research, and plan alumni engagement. Gift officers own relationships.

Start With This Stack
  • Prompts: annual giving review, prospect profile, alumni engagement score review
  • Skills: annual giving & donor segmentation; major gift & prospect preparation; alumni engagement programs
  • Connect: Raiser's Edge NXT via Blackbaud's Claude connector, or your CRM (Ellucian CRM Advance, Salesforce) through governed exports
  • Agent: annual giving pace monitor; major gift pipeline reviewer
Giving Analysis
  • What AI does: Performance by constituency and channel
  • Helps: Invest where giving grows
  • Control: Follow counting policy
Appeal Segments
  • What AI does: Builds segments and tailored appeals
  • Helps: Higher participation
  • Control: Director approves
Prospect Profiles
  • What AI does: Sourced briefs from public and CRM data
  • Helps: Better-prepared visits
  • Control: Ethical research only
Proposals & Agreements
  • What AI does: Drafts proposals; checks agreements against policy
  • Helps: Faster, safer gifts
  • Control: Counsel approves agreements
Stewardship
  • What AI does: Impact reports for endowed funds
  • Helps: Repeat and larger gifts
  • Control: Student consent for stories
Alumni Engagement
  • What AI does: Plans events, volunteers, and communications
  • Helps: Lifelong connection
  • Control: Data-use agreements

Advancement Implementation Checklist

Workflow

Before You Start

  • Clean constituency codes and gift data
  • Document counting and gift acceptance policies
  • Write research ethics guidelines
  • Confirm connector role permissions
  • Baseline: retention, participation, prep time

After You Start

  • Review all donor-facing drafts
  • Audit profiles for sources
  • Track retention monthly
  • Steward every endowed fund annually
  • Report campaign progress quarterly
Advancement Guardrails
  • Ethics: Prospect research follows professional ethics (e.g., Apra) and uses public and institutional data only.
  • Privacy: Donor data stays in the CRM; minimum necessary in any export.
  • Restrictions: Honor gift restrictions and naming policy.
  • Honesty: Never overstate impact or invent stories.
  • Relationships: Gift officers own every ask and agreement.
Top Advancement tools
Blackbaud Raiser's Edge NXTEllucian CRM AdvanceSalesforce Education CloudGiveCampusEverTrueDonorSearchiWaveGraduway

AI for Grants & Sponsored Programs

Use AI to summarize opportunities and terms, draft justifications, and monitor spending. Sponsored programs staff and PIs own compliance.

Start With This Stack
  • Prompts: funding opportunity summary, award terms summary, burn rate review
  • Skills: pre-award proposal support; post-award grant management
  • Connect: research administration system reports (Cayuse, Kuali Research, Huron), ERP grant ledgers
  • Agent: funding opportunity digest; grant burn rate & closeout watch
Opportunity Summaries
  • What AI does: One-page summaries of announcements
  • Helps: Faster go/no-go
  • Control: PI reads full announcement
Compliance Checklists
  • What AI does: Sponsor and internal requirements with owners
  • Helps: Fewer rejected proposals
  • Control: Check current guides
Budget Justifications
  • What AI does: Drafts justifications and flags unallowables
  • Helps: Fewer sponsor questions
  • Control: Rates verified
Award Terms
  • What AI does: Summarizes special terms and deadlines
  • Helps: Nothing buried
  • Control: Post-award reviews
Spending Monitoring
  • What AI does: Burn rates and cost transfer flags
  • Helps: Avoid overruns and findings
  • Control: PIs confirm plans
Closeouts
  • What AI does: Timelines for final reports and subawards
  • Helps: On-time closeouts
  • Control: Confirm award terms

Sponsored Programs Control Checklist

Controls

Before You Start

  • Collect negotiated rates and policies
  • Export award and expenditure data
  • List prior audit findings
  • Define cost transfer and effort rules
  • Baseline: late reports, findings

After You Start

  • Verify all summaries against documents
  • Track burn-rate flags to resolution
  • Review effort and cost transfer patterns
  • Close out on time
  • Brief auditors on AI-assisted processes
Grants Guardrails
  • Uniform Guidance: Federal awards follow 2 CFR 200 and sponsor terms; humans decide allowability.
  • Certifications: Effort and financial reports are certified by people.
  • Integrity: Never invent research results or progress.
  • Confidentiality: Unpublished research and proposals stay in approved tools.
  • Export controls and security: Follow research security and export control rules for sensitive projects.
Top Research Administration tools
CayuseKuali ResearchHuron Research SuiteInfoReadyPivot-RPGrants.govResearch.govWorkday Grants

AI for Financial Aid Administration

Use AI for modeling with aggregates, clearer communication, and organized compliance. Student aid data and every student decision stay with aid professionals.

Start With This Stack
  • Prompts: discount rate model, award letter clarity review, aid FAQ
  • Skills: financial aid operations & communication
  • Connect: aggregate reports from your aid system (PowerFAIDS, Banner, Colleague, Workday Student)
  • Agent: aid communications & deadline assistant
Aid Modeling
  • What AI does: Discount rate and net tuition scenarios from aggregates
  • Helps: Align aid and revenue
  • Control: VP and CFO review
Award Letters
  • What AI does: Clarity reviews against best practice
  • Helps: Families compare offers
  • Control: Compliance approves changes
Student Communications
  • What AI does: Verification, SAP, and deadline messages
  • Helps: Faster completion
  • Control: Content verified yearly
Self-Service Content
  • What AI does: FAQ and chatbot answers
  • Helps: Lower call volume
  • Control: Individual cases go to staff
Process Checklists
  • What AI does: R2T4, verification, and PJ documentation templates
  • Helps: Consistent, auditable processes
  • Control: Verify against FSA Handbook
Compliance Calendar
  • What AI does: Recurring federal and state obligations
  • Helps: Nothing missed
  • Control: Compliance confirms

Financial Aid Control Checklist

Controls

Before You Start

  • Confirm GLBA safeguards and data classification
  • Define aggregate bands for modeling
  • Collect policies and annual dates
  • Review vendor AI features in the aid system
  • Baseline: call volume, processing time

After You Start

  • Keep student-level data in the aid system
  • Review communications each award year
  • Document every professional judgment by staff
  • Track processing times
  • Include AI processes in audit prep
What AI Should NOT Do in Financial Aid
  • Do NOT: Paste FAFSA/ISIR data, tax information, or student aid records into general AI tools.
  • Do NOT: Make professional judgment, SAP appeal, dependency override, or verification decisions.
  • Do NOT: Perform R2T4 calculations outside the aid system.
  • Do NOT: Send messages about individual students' aid.
  • Do: Treat aid data as protected under GLBA, FERPA, and Title IV requirements.
Top Financial Aid tools
PowerFAIDSEllucian Banner Financial AidEllucian ColleagueWorkday StudentRegent AwardCampus LogicOcelotNASFAA resources

AI for Budgeting & Finance

Use AI for variance narratives, scenarios, audit prep, and exception reviews. The CFO's and controller's offices own the numbers.

Start With This Stack
  • Prompts: variance narrative, enrollment revenue scenarios, P-card review
  • Skills: budget & financial planning; business office & audit readiness
  • Connect: ERP reports (Workday, Oracle, Banner Finance, Jenzabar), budget system (Axiom, Anaplan, Adaptive)
  • Agent: quarterly budget variance drafter; P-card & AP exception queue
Variance Narratives
  • What AI does: Drafts quarterly narratives from approved figures
  • Helps: Faster reporting
  • Control: Unit officers confirm drivers
Scenario Models
  • What AI does: Enrollment-driven revenue and 5-year plans
  • Helps: See gaps early
  • Control: Leadership agrees on assumptions
Trustee Memos
  • What AI does: Drafts finance committee memos
  • Helps: Clear decisions
  • Control: Controller verifies
Exception Reviews
  • What AI does: P-card, AP, and restricted fund flags
  • Helps: Focus review time
  • Control: Flags aren't findings
Audit Prep
  • What AI does: PBC and Single Audit preparation
  • Helps: Shorter fieldwork
  • Control: Confirm scope with auditors
Budget Model Explainers
  • What AI does: Plain-language guides for deans and chairs
  • Helps: Better unit decisions
  • Control: Budget office reviews

Finance Implementation Checklist

Controls

Before You Start

  • Define approved data sources
  • Document thresholds for variances and exceptions
  • Collect policies and prior findings
  • Set up read-only reporting access
  • Baseline: reporting hours

After You Start

  • Reconcile every figure to the ERP
  • Track flagged exceptions to resolution
  • Review models quarterly
  • Mark drafts until reviewed
  • Brief auditors
Finance Guardrails
  • Accuracy: Every number ties to the system of record.
  • Postings: AI never posts, approves, or pays.
  • Restrictions: Restricted funds follow donor and sponsor terms.
  • Public institutions: Records and open meetings laws may apply to trustee materials.
  • Decisions: Budgets and tuition are leadership and trustee decisions.
Top Finance tools
Workday FinancialsOracle Cloud ERPEllucian Banner FinanceJenzabarAxiomAnaplanWorkday Adaptive PlanningUnit4

AI for Campus Operations & Auxiliaries

Use AI to forecast occupancy, track adoptions, analyze dining and store results, and prioritize maintenance. Operations leaders decide.

Start With This Stack
  • Prompts: housing occupancy forecast, textbook adoption tracker, work order backlog
  • Skills: auxiliary services performance; facilities & campus operations analysis
  • Connect: housing (StarRez), dining (CBORD), store, and work order system exports
  • Agent: housing occupancy forecaster; facilities backlog monitor
Housing Forecasts
  • What AI does: Occupancy and revenue by hall
  • Helps: Protect debt service
  • Control: Update weekly
Course Materials
  • What AI does: Adoption tracking and affordability analysis
  • Helps: Lower student costs
  • Control: Faculty choose materials
Dining Analysis
  • What AI does: Plan mix, participation, and peak usage
  • Helps: Better service
  • Control: Contract terms govern
Maintenance Backlog
  • What AI does: Backlog by trade, age, and priority
  • Helps: Fix what matters first
  • Control: Validate categories
Space & Energy
  • What AI does: Utilization and energy intensity
  • Helps: Avoid unnecessary building
  • Control: Registrar decides scheduling
Capital Projects
  • What AI does: Budget and schedule status
  • Helps: See risk early
  • Control: Reconcile monthly

Campus Operations Checklist

Workflow

Before You Start

  • Export occupancy, sales, and work order data
  • Collect contracts for outsourced services
  • Define priorities and standards
  • Confirm minors-on-campus and accommodation policies
  • Baseline: occupancy, backlog age

After You Start

  • Compare forecasts to actuals
  • Track backlog trends
  • Review contractor performance quarterly
  • Report auxiliary results annually
  • Survey students on services
Campus Operations Guardrails
  • Accommodations: Disability services decides housing accommodations; AI never does.
  • Assignments: Room assignments follow policy and are made by housing staff.
  • Safety: Clery and emergency decisions stay with designated officials.
  • Academic freedom: Faculty choose course materials.
  • Contracts: Outsourced services operate under contract terms.
Top Campus Operations tools
StarRezAdirondack SolutionsCBORDTransactFollettBarnes & Noble CollegeAssetWorks AiMBrightly (SchoolDude)

AI for Enrollment Operations

Use AI for funnel analysis, communication audits, and yield and melt planning. Admission decisions and holistic review stay with people.

Start With This Stack
  • Prompts: funnel dashboard narrative, communication flow audit, melt prevention plan
  • Skills: enrollment funnel & yield operations
  • Connect: CRM reports (Slate, Salesforce Education Cloud, Element451, Ellucian CRM Recruit)
  • Agent: enrollment funnel watch
Funnel Narratives
  • What AI does: Same-date comparisons by population
  • Helps: Early warning
  • Control: Note calendar changes
Flow Audits
  • What AI does: Finds gaps and over-messaging in CRM flows
  • Helps: Clearer communication
  • Control: Test before launch
Completion Campaigns
  • What AI does: Nudges by missing item
  • Helps: More reviewable files
  • Control: Cap frequency
Yield Planning
  • What AI does: Touchpoints by segment
  • Helps: Higher yield
  • Control: Follow law and policy
Melt Prevention
  • What AI does: Task-based summer reminders
  • Helps: Keep deposited students
  • Control: Owner per task
Forecast Memos
  • What AI does: Headcount and credit hour projections
  • Helps: Align finance and academics
  • Control: Update weekly

Enrollment Operations Checklist

Workflow

Before You Start

  • Export funnel counts with same-date history
  • Inventory CRM flows
  • Define segments within legal and policy limits
  • Confirm reader and committee processes remain human
  • Baseline: yield, melt

After You Start

  • Review narratives weekly
  • Test flow changes
  • Track melt by task
  • Compare forecast to census
  • Review segmentation with counsel annually
Enrollment Guardrails
  • Decisions: AI never scores, ranks, or reads applications for admission decisions.
  • Law and policy: Segmenting and targeting follow current law and institutional policy; counsel reviews.
  • Privacy: Application materials stay in the CRM; analysis uses counts and IDs.
  • Accuracy: Program facts and deadlines verified each cycle.
  • Accessibility: Communications meet accessibility standards.
Top Enrollment tools
Slate (Technolutions)Salesforce Education CloudElement451Ellucian CRM RecruitCommon AppEABLiaisonMongoose

Set Up & Get Running

If you already use prompts, here is how to turn one into a repeatable skill and then a useful first agent. Start with a task that already works.

From prompt to first agent

1. Pick One Working Prompt
  • Choose a recurring unit task
  • Use real data at the right classification
  • Good first choices: variance narrative, funding digest, or giving review
RepeatableLow riskReal work
2. Turn It Into a Skill
  • Write down inputs, output, and reviewer
  • Test on three real cycles
  • Improve when results are unclear
InputsInstructionsReview
3. Add Only What It Needs
  • Start with reports your reviewer trusts
  • Use enterprise accounts and vendor connectors with role permissions
  • Keep restricted data in systems of record
ReportsRead-onlyOwner
4. Build Your First Agent
  • Start with a monitoring task
  • Set when it runs; a person checks every output
  • Track time saved and edits
MonitorReviewMeasure
A good first agent: Grant Burn Rate & Closeout Watch

It's useful, low-risk, and easy to check. It compares spending to time elapsed and flags closeout deadlines. It doesn't move costs, approve transfers, or certify effort.

Copy these builders

Use the first after a prompt works. Use the second only after the skill has been tested on real work.

Prompt → Skill Builder
You are a higher education operations designer. Turn the successful prompt below into a reusable skill for an administrative unit. Successful prompt: [PASTE THE PROMPT] What good output looks like: [PASTE A STRONG EXAMPLE OR DESCRIBE IT] Context the skill may use (with data classification): [LIST APPROVED FILES, REPORTS, OR SYSTEMS] Create a skill with: 1. A clear name and the job it performs 2. When to use it and when NOT to use it 3. Required inputs and missing-information questions 4. Step-by-step instructions 5. Exact output format 6. Examples of acceptable vs. unacceptable output 7. Human reviewer, approval boundary, and stop rules 8. A short test plan using three real examples Rules: - Never invent figures, regulations, sponsor terms, or donor intent - Never use restricted data (student records, aid data, SSNs, personnel) outside approved systems - Keep decisions about students, gifts, grants, and money with people
Skill → Agent Builder
You are a higher education operations designer. Turn the tested skill below into a safe, human-reviewed agent blueprint. Tested skill: [PASTE THE FULL SKILL] Trigger or schedule: [EXAMPLE: MONTHLY AFTER CLOSE / 90 DAYS BEFORE AWARD END DATES] Approved sources (read-only, with data classification): [LIST SYSTEMS OR REPORTS] Named human owner: [NAME OR ROLE] Create an agent blueprint with: 1. One clear job to be done 2. Trigger or schedule 3. Approved context and permission scope 4. Step-by-step operating loop 5. Review-ready deliverable 6. Escalation rules 7. Actions the agent must never take 8. Human approval point 9. Audit-log fields 10. Success measures and a two-week pilot plan Rules: - Start read-only and deliver a review packet - Never post, submit, certify, or contact students, donors, sponsors, or agencies - Stop and escalate when regulation, policy, or authority is unclear
Implementation Timeline

Days 1-30: Foundation

  • Name an executive sponsor and unit AI leads
  • Confirm enterprise AI accounts and data classification rules
  • Review AI features in ERP, CRM, and student systems
  • Pick 1 pilot per unit (max 3 units)
  • Record baselines
  • Train pilot users

Days 31-60: Expand

  • Save pilots as skills; test 3 cycles
  • Add a 2nd workflow per unit
  • Turn on one vendor connector with role permissions
  • Draft the administrative AI policy
  • Start an AI use case inventory
  • Measure time saved and edits

Days 61-90: Standardize

  • Policy through governance approval
  • Pilot one review-only agent
  • Write SOPs
  • Report results to cabinet
  • Expand to additional units
  • Plan the next wave

Implementation Success Measures

Measurement

30-Day Targets

  • Data rules followed
  • Pilots running with review
  • Baselines recorded
  • No restricted-data incidents

60-Day Targets

  • Pilot time down noticeably
  • Second workflows live
  • One connector in use safely
  • Draft policy reviewed

90-Day Targets

  • Policy approved
  • First agent piloted
  • SOPs written
  • Use case inventory live
  • Decision on next wave
A realistic pace
Pilot in a few units, then expand. Avoid launches during fiscal year-end, census, or aid packaging peaks.
📅

How the Pieces Fit Together

Start with a prompt. Save the ones that work. Add the systems and files they need. Then use an agent for recurring work, with someone checking the result.

1. Prompts
  • Ask AI to do one clear task
  • Use it for a quick draft, analysis, or explanation
  • Check the result before you use it
TaskContextReview
2. Skills
  • Save the step-by-step instructions for a task your team repeats
  • Add your policies, examples, and review steps
  • Update it when your process or regulations change
InstructionsExamplesReview
3. Connectors
  • Bring the right systems, files, and reports into the task
  • Stop copying and pasting information by hand
  • Start read-only, with only the sources the task needs
SystemsFilesRead-only
4. AI Agents
  • Run a skill on a schedule or when something happens
  • Gather the information, prepare the work, and flag issues
  • Leave student, gift, grant, and financial decisions with people
ScheduleFlagsApprovals
Take it one step at a time
Use a prompt to solve one task. Save it as a skill when it works. Connect the files and systems it needs. Then set it up to run regularly, with someone checking the work before anything important happens.
⚙

Where AI Can Help Higher Education Administrators

Use AI to get routine administrative work into a clear first draft. Keep decisions about students, gifts, grants, and money with your teams.

Advancement & Grants
  • Giving analysis and segmented appeals
  • Ethical prospect profiles and proposals
  • Funding opportunity summaries
  • Burn rates and closeout plans
Finance & Aid
  • Variance narratives and scenarios
  • Exception queues and audit prep
  • Discount rate models from aggregates
  • Clearer aid communications
Enrollment & Operations
  • Funnel narratives and flow audits
  • Housing and dining forecasts
  • Textbook adoption tracking
  • Maintenance backlog analysis
What Stays With People
  • Admission and aid decisions
  • Gift agreements and asks
  • Grant certifications
  • Trustee and regulatory submissions
Start with preparation, not decisions
The first win is a better work packet: facts, exceptions, and drafts ready for the person who decides.
✓

AI Capabilities Explained

No jargon. What each kind of AI does, and what it looks like in campus administration.

Generative AI (LLMs)

Claude, ChatGPT, Gemini, and Copilot write, summarize, and follow instructions.

On campus: appeals, trustee memos, award term summaries, FAQ content.

Connected Assistants

AI that reads approved systems through connectors with role permissions.

On campus: asking Claude about donor segments through the Raiser's Edge NXT connector.

Embedded Agents

AI built into ERP, student, and CRM systems that can take steps within the product.

On campus: vendor agentic features in Ellucian, Workday, or Salesforce, after review.

Forecasting

Projects numbers from patterns and assumptions.

On campus: enrollment revenue, housing occupancy, grant burn rates.

Rules Checking

Compares transactions to policy and flags exceptions.

On campus: P-card reviews, cost transfers, restricted fund charges.

Document Q&A

Answers questions from documents you supply.

On campus: 'What prior approvals does this award require?'

The common thread
AI predicts likely answers and can be confidently wrong about regulations, sponsor terms, and figures. Always verify.
🧠

Guardrails: Data Classification, Compliance & Stewardship

Simple rules for administrative AI use in higher education. Students, donors, sponsors, and regulators all trust you with their data and money.

Never Put Into Unapproved Tools
  • Student education records (FERPA)
  • Financial aid and tax data (Title IV, GLBA)
  • SSNs, health, or disability information
  • Personnel files and confidential trustee matters
Human Decisions
  • Admission and holistic review
  • Professional judgment, SAP appeals, R2T4
  • Effort certification and cost allowability
  • Clery classifications and timely warnings
Stewardship
  • Figures tie to the system of record
  • Donor and sponsor restrictions honored
  • Ethical prospect research
  • Trustees and leadership approve budgets
Enterprise Controls
  • Institution-licensed AI with training off
  • Data classification decides tool use
  • Vendor AI features reviewed by IT and counsel
  • Access removed when staff change roles
Public & Private Differences
  • Public: records and open meetings laws
  • Private: Form 990 and donor confidentiality
  • All: accreditor standards
  • All: state reporting obligations
What Stays With People
  • Decisions about students
  • Gift agreements and asks
  • Sponsor and regulatory submissions
  • Personnel and contract decisions

Administrative AI Readiness Checklist

Controls

Leadership & Policy

  • Executive sponsor for administrative AI
  • Administrative AI policy approved through governance
  • Enterprise accounts with training off
  • Vendor AI review process in place
  • Role-based training delivered

Practice & Review

  • Data classification applied to every use case
  • Human review on external and trustee materials
  • No AI in student decisions
  • AI use case intake and inventory
  • Annual review of AI use and incidents
Sample Administrative AI Use Policy

Scope: Administrative units (finance, business office, advancement, sponsored programs, financial aid, enrollment, auxiliaries, facilities). Teaching and research AI use are governed by separate policies.

Approved tools: [institution-licensed AI accounts; reviewed AI features in ERP, CRM, and student systems; the Blackbaud connector for Raiser's Edge NXT if licensed]

Data rules: Public and internal data may be used in approved tools. Confidential data only in tools approved for that classification. Restricted data (student education records, financial aid and tax data, SSNs, health, personnel) only in contracted systems of record.

Human decisions: AI does not decide admission, aid, appeals, effort certification, cost allowability, Clery classifications, gifts, or personnel matters.

Accountability: A named staff member reviews and owns every output used externally or with trustees.

Accessibility & records: AI-assisted content meets accessibility standards and is retained per records schedules.

Approval: [Cabinet / governance body], [date]

Golden rule
Your data classification policy decides which tool can see what. When in doubt, use aggregates.
🛡

AI Maturity Model for Higher Education Administration

See where your administrative units are today. Choose a realistic next step.

1

Curious + Ad Hoc

  • Individuals use consumer AI
  • No administrative AI policy
  • Vendor AI features unreviewed
  • No use case inventory
2

Enterprise Accounts & Rules

  • Enterprise AI with training off
  • Data classification guidance
  • Prompt libraries by unit
  • Human review on outputs
3

Embedded Unit Workflows

  • Saved skills in several units
  • Policy approved through governance
  • Vendor connectors with role permissions
  • Use case inventory
4

Connected & Governed

  • Review-only agents with owners and audit logs
  • Governed data layer
  • Measured outcomes across units
  • Annual policy and agent review

Maturity Self-Assessment

Assessment

Leadership

  • Executive sponsor
  • Policy approved
  • Unit AI leads
  • Vendor review process

Practice

  • Prompt libraries
  • Saved skills
  • SOPs
  • Governed exports

Protection

  • Classification applied
  • No AI student decisions
  • Human review
  • Access reviews

Measurement

  • Baselines
  • Time saved
  • Errors caught
  • Annual review
Your target state
Most institutions can bring several units to Level 3 within a year. Move at the pace of your governance processes.
🎯