Find the right people.
See why they are the right people.
The Prospect Agent answers one question three different ways: who should your team be talking to. It generates new leads from public sources, discovers candidates inside the constituent base you already own, and recommends the audience for a campaign — each one ranked, explained, and waiting for a person to approve it.
AI Lead Gen
Builds a pipeline that does not exist in your CRM yet.
Describe the kind of prospect you are looking for as a search recipe rather than a query: the type of donor or organization, the roles, the giving scenario, the regions. The agent expands that into search combinations, works through public search engines, and streams what it finds straight into your Lead object.
- Recipe dimensions: donor types, organizations, roles, scenarios, regions and free-text terms
- Runs across public search engines, up to 200 search combinations per task
- Results import live — created, merged and failed counts update as the task runs
- Contacts are matched on email so a repeated run updates a lead instead of duplicating it
- Every task keeps its own import batch, so you can always see where a lead came from
Constituent Discovery
Turns a fundraising goal into a ranked shortlist.
State the goal in plain language. The agent proposes a reviewable analysis plan, narrows the full constituent base down to a candidate pool with normal database rules, builds an evidence package for each candidate, and only then spends AI on the strongest ones. The ranking that comes back is stable and repeatable.
- Four scoring dimensions — affinity, capacity, readiness, momentum — with weights your team sets
- Signals drawn from gifts, event attendance and registration, interactions, open opportunities, volunteer and mentor activity, and research indicators already in the system
- Automatic exclusions, including deceased constituents and anyone already in active major-gift cultivation
- Thresholds you control: minimum recommendation score, minimum confidence, and how many people to return
- Save the run as a daily, weekly or monthly schedule and let new recommendations arrive on their own
Campaign Audience
Recommends who a campaign should reach, before it exists.
Describe the campaign and the agent proposes the audience alongside the rest of the plan: the existing segments that fit the brief, plus a new set of conditions drafted from your description. Every headcount on the screen is calculated by the segmentation engine against live data — the AI never estimates a number.
- Matches the brief against segments your team has already built, and drafts new criteria when nothing fits
- Estimated reach comes from the segment engine, not from the language model
- Include several segments, exclude the ones you do not want, or refresh the recommendation
- A drafted audience with no matching constituents is never offered as a recommendation
- Nothing is created and nothing is sent until someone accepts the draft
Scoring & Evidence
Every recommendation shows its work.
A score nobody can explain is a score nobody will act on. Each candidate carries a recommendation score, a confidence score, the four dimension scores behind it, and the individual records that produced them. The AI summary is required to cite evidence that actually exists on the record; anything it cannot support is dropped.
- Recommendation score and confidence score, side by side and never merged
- Dimension breakdown across affinity, capacity, readiness and momentum
- Traceable evidence: which gift, which event, which interaction, which opportunity
- Data gaps stated openly, so a thin record is not mistaken for a weak prospect
- Staff decision on every candidate: qualify, request research, or disqualify
- Six lifetime gifts, most recent 8 months agoDonation
- Attended the regional alumni reception and the engineering showcaseEvent attendance
- Two staff interactions logged in the last 90 daysInteraction
- No open opportunity and no active cultivation planOpportunity
- Capacity indicator on file, recorded by advancement researchResearch indicator
What the agent is not allowed to do
Prospect work touches the most sensitive data in the institution, so the boundaries are part of the design rather than a setting. The agent reads the data your team already governs, keeps its reasoning tied to records that exist, and stops at a recommendation.
Where the Prospect Agent stops
It decides who deserves attention and why now — new leads, existing constituents, campaign audiences — and hands over a ranked list with the evidence attached.
Where the Gift Officer Agent starts
What to say, how to cultivate the relationship, and what to ask for. Once a candidate is qualified, that work belongs to the next agent in the workforce.
Meet the Gift Officer Agent →