Prospect research that turns one sentence into a cited, one-page donor briefing, with every weak or protected detail left out and shown
A cited brief before every visit in minutes, not days, each wealth figure confirmed twice.
See one case, screen by screen ↓Prospect research is the work an advancement office does before a visit to learn who a donor is to the institution, what they could give, what they care about and what not to raise. Prospect Research gives gift officers that as a cited, one-page brief built by agents from the CRM, giving history and public filings, with every weak or protected detail left out and shown, and an analyst approving any rating change.
AI in education: a buyer's guide to agentic AI for colleges, universities and school districts →
Updated
Prospect research is how an advancement office learns, before a visit, who a donor is to the university, what they could give, what they care about and what not to say. At a university where a few people give most of the gift dollars, that page decides how a major-gift visit goes. Today a prospect researcher builds it by hand from the advancement CRM, a wealth screening file, county property records, SEC filings, foundation tax returns and the news, and a full profile takes hours.
So research reaches the top of the pipeline, and many visits are booked before an analyst can get to them. A single-source net worth from a blog slips into a profile and an officer overasks. Health and religion details turn up in contact reports, and an alumnus's student records, or a child's, are education records under FERPA. Capacity ratings set by a screening years ago stay in the CRM. Prospect Research puts eight agents in front of the gathering and checking, so the officer gets one cited page and an analyst decides what changes.
Estimated split of the hours a researcher spends on one full profile by hand, and with the agents. The 'after' column is a person's reading and checking; the agents' own run takes under a minute. Our arithmetic, kept modest.
Eight agents gather, verify and exclude, and build one cited page; an analyst approves any rating change.
Claire Ashby, a major gifts officer at Halden State University, has lunch on Thursday with Renata and Paul Okonkwo-Hale, constituent HSU-0418-2276. Here is the prospect research behind that lunch, screen by screen, in the working solution.
Claire's home is a question box and her week: five visits, four briefed, and an Advisory Board dinner with 12 of 14 briefs ready. She types one sentence instead of filling in a research request form.
“Brief me on Renata and Paul Okonkwo-Hale before Thursday's lunch.”
A run line names each agent at work: the Brief coordinator, the Records agent, the Public records agent, the Verification agent, the Privacy agent and the Rating agent. The page opens with the relationship: Renata, B.S. Civil Engineering, class of 1996, on the Engineering Advisory Board since 2022; Paul, an independent director of Brennick Materials. Eleven consecutive fiscal years of giving, $186,400 lifetime. Twelve sources sit in the column on the right, each with its date.
Built in 39 s · 12 sources · 5 removed
The five-year giving capacity is $1M–$2.5M, from three figures, each confirmed twice: $2.34M of real estate in county records, matched by the wealth screening file; $2.82M of Brennick Materials stock, 41,200 shares at $68.40, in an SEC Form 4 and the proxy statement; and about $8.2M from selling 17% of Renata's firm, in a business journal and the state business registry. The method is on the page: 7.5% of the verified assets at the low end; 10% at the high end, once the rest of her Ridgeline stake is counted at 60% of its reported value.
Nothing weak or protected reaches the officer, and nothing disappears silently. Five items are struck through, each with the agent and the reason: a blog's net worth of $40 million with one source, a 2019 property sale for a different Paul D. Hale, a contact-report line about surgery, a congregation's donor list, and a family member's student records under FERPA.
Removed · 5
Claire clicks S11 and reads the business journal's own words, highlighted, with the source date of March 3, 2026, when the agents read it, what it was used for, and that the research policy allows it. Opening it is itself written to the audit trail.
“The deal values the 180-person firm at about $48 million.”
The verified range moves the prospect from the C band to the B band, $1M–$4.99M, so the Rating agent proposes the change with its evidence. Seen as Claire, Approve, Return to research and Reject are locked: ratings in an officer's own portfolio are approved by an analyst.
Tomás picks RC-01, New verified liquidity event, and approves. The approval is recorded, rating B is written to the CRM, and the CRM confirms the write-back at 2:32 PM. A jump of two levels or more would also need the Director of Prospect Research.
The header now reads Rating B, approved by Tomás Ferreira, Oct 21, 2:31 PM. Claire can print the page, which is watermarked and logged, or ask to share it with a dean, which waits for the Director of Prospect Management.
Rating B · approved by Tomás Ferreira, Oct 21, 2:31 PM
The same loop closes after every visit. From Claire's notes on Monday's coffee with Gerald Ostrowski ’82, the agent drafts the purpose, what she learned and the next step, citing each line of her notes, while the privacy agent strikes a health detail and a grandson's college plans. His planned retirement goes to research as a fact to verify. Nothing reaches the CRM until she approves.
The trail lists people and agents alike: the request, the sources read, the three protected details removed, the brief released with 12 sources and 5 removed, the proposal, Tomás's approval and the CRM write-back. An export pack answers an internal audit or a donor's data request, and the export itself is logged.
The same brief, seen by the four people who carry it: what their day looked like, and what it looks like now.
Eight agents gather, verify and exclude, and build one cited page; an analyst approves any rating change.
Reads the request, matches the constituent, plans the lookups and assembles the page in order, tied to the visit on the calendar.
Reads the CRM record, gift history, contact reports and the alumni degree record.
Reads county property records, SEC filings, foundation tax returns, the state business registry and the news archive.
Keeps a wealth figure only when two independent sources agree, checks identity by name, age and address, and strikes the rest into Removed.
Removes protected details and FERPA education records, and honors GDPR objections and do-not-research flags.
Works out the five-year capacity range with the institution's method and proposes a rating change.
Watches every rated prospect each night for new filings, sales, foundation returns and news.
Turns the officer's notes into a contact report draft and lists new facts for research to verify.
Approves rating changes with a reason. The agents propose; a named person decides.
The officer and the research team can ask about any brief in plain words. Every answer names its sources, and anything that changes the page or leaves advancement still waits for a person.
Why is the $40 million net worth not in the brief?
Because it had one source: a personal-finance blog post from June 2025 that gives no method. None of the 12 sources agrees, and the research policy keeps a wealth figure only when two independent sources agree. The three figures that did agree add up to $13.3M of verified assets, which is where $1M–$2.5M comes from.
Who in my portfolio had a liquidity event this year?
Two of your 142 prospects, each confirmed by two sources, and one still being checked: Renata Okonkwo-Hale, the sale of 17% of Ridgeline Geotechnical, about $8.2M; Dr. Felix Amadi ’90, whose consultancy a listed firm bought in May, about $3.6M; and Gerald Ostrowski ’82, the business succession you reported on Monday, which research is verifying.
Draft three questions for Thursday that test the lab naming idea
Three, each from what she has told you: what would the lab need for your firm to hire from it; would you want transfer students to have a place in it; and would the family foundation want a part in this, since it gave $96,500 in 2024, three grants to STEM education. I added them to the brief under Open questions, marked as yours.
Share this brief with the Dean of Engineering
A brief that leaves advancement needs Grant Holloway's approval. I prepared the request: the dean receives the page without the Removed list or the source passages, and the copy is logged. It is waiting for Grant now.
13 screens from the working solution, on its sample data. Pick one to see it large.
The rules every brief is built against are settings, changed only by the administrator and kept on the audit trail.
| Setting | Default | Choose from |
|---|---|---|
| Wealth figures | Two independent sources required | A single-source figure goes to Removed, with the reason |
| Identity confidence | 0.95 | Below it, the coordinator stops instead of guessing |
| Capacity method | 7.5% (low) and 10% (high) of verified assets | Private holdings counted at 60% |
| Rating scale | A $5M+ · B $1M–$4.99M · C $250K–$999K · D $100K–$249K · E $25K–$99K | A jump of two levels or more needs the Director of Prospect Research |
| Protected details | Health, religion, sexual orientation, political giving, minors | Always removed; only the reason is kept |
| Student records (FERPA) | Never read | Degree and class year from the degree record only |
| Alumni in the EU, EEA and UK (GDPR) | Legitimate-interests assessment on file | An objection stops research; data requests answered within one month |
| Clocks | Visit brief due 1 business day before the visit | Full profile 10 business days · quick question 1 business day · unassigned within 24 hours of due escalates |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
Sources: CASE, Insights on Voluntary Support of Education 2025 (giving to US colleges about $78.8 billion in fiscal 2025, up 4%; about 2% of donors gave 89% of gift dollars) · FERPA, 20 U.S.C. 1232g and 34 CFR Part 99 (education records) · GDPR Articles 21 and 12(3) (the right to object; requests answered within one month) · "demo" = seen in the working solution, on its sample data · "typical" = published figures · "estimated" = our arithmetic, kept modest. Halden State University, its donors, companies and the people named on this page are fictional: characters and sample data in the working solution.
Prospect research is the work an advancement office does before a visit to learn who a donor is to the institution, what they could give, what they care about and what not to raise. Prospect Research gives gift officers that as a cited, one-page brief built by agents from the CRM, giving history and public filings, with every weak or protected detail left out and shown, and an analyst approving any rating change.
Wealth screening scores a list of constituents for capacity, usually from a vendor's data. This solution uses the screening file as one source among several and turns it, the CRM and public records into the page a gift officer reads before a visit. A wealth figure stays only when two independent sources agree, and everything else is listed under Removed with its reason.
The Verification agent keeps each asset only when two independent sources agree, for example county property records matched by the screening file, or an SEC Form 4 matched by the proxy statement. The Rating agent then applies the institution's method, set in the settings: 7.5% of verified assets at the low end and 10% at the high end, with private holdings counted at 60%. The arithmetic is shown on the capacity page.
No. The Privacy agent removes health, religion, sexual orientation, political giving and details about minors before anyone sees the brief, and keeps only the reason. Education records of alumni and their children are never read under FERPA; degree and class year come from the degree record only. An objection from an alumnus in the EU, EEA or UK stops research.
No. The agents gather, check and propose; they never write a rating, change a record, share or send anything, or read the student information system. A rating reaches the CRM only after an analyst approves it with a reason code, and a contact report only after the officer approves it.
Yes. An analyst approves every rating change, a jump of two levels or more also needs the Director of Prospect Research, and an officer can never approve a rating for her own prospect. Sharing a brief outside advancement needs the Director of Prospect Management, and the officer can flag any fact on a brief for an analyst to check.
Yes. The two-source rule, the identity confidence threshold, the capacity method, the rating scale, the protected categories, retention and the request clocks are all settings, changed by the administrator and kept on the audit trail.
The Agentic Solution Engine builds and deploys it from your requirements and documents, such as your research policy, your rating scale and a sample of past profiles, and it goes live once every quality gate has passed. We will show it to you on a sample of your own prospects first.
Every charge checked against the donor's own words, and idle payout put to work.
1-min demoGift Acceptance & Donor Due DiligenceEvery large gift and naming decided on a cited file, with returnable gifts flagged early.
1-min demoGift Processing & Donor AcknowledgmentGifts booked and receipted within 48 hours, each coded to the fund the donor named.
1-min demoWe'll build briefs for a sample of your own prospects, from your CRM and public sources.