Skip to main content

For nonprofits

Your donors did not lose interest. They never heard from you again.

Fewer than one in five first-time donors ever gives a second gift. That is rarely about generosity. It is capacity: the thank-you and the second ask land on a team already at its limit. We build custom AI systems that steward every donor on time and in your voice, then run them for you.

100 first-time donors. 19 give again.
Fundraising Effectiveness Project, 2025

the arithmetic

18.9%
First-time donors who give a second gift

FEP 2025

59%
Second-time donors who keep giving after that

FEP 2025

92%
Nonprofits now using AI somewhere

Nonprofit AI Adoption, 2026

7%
Nonprofits where AI meaningfully changed output

Nonprofit AI Adoption, 2026

Almost everyone has the technology. Almost nobody has the outcome. The gap is not the model. It is that no one is running it.

where it actually breaks

Four failures, and none of them are about caring less.

01

The second gift is never asked for

A first gift arrives on a Thursday. The thank-you goes out late, the impact update never gets written, and the second ask waits for a quarter that never has a spare week in it. The donor does not decide to leave. They are simply never spoken to again.

02

You already bought the AI. Nothing changed.

Almost every organization now has a tool with AI in the name. Very few report it changing what their team can actually finish. Software was never the missing piece. Somebody has to design the workflow, wire it to your donor data, and keep it running on the weeks everyone is underwater.

03

Stewardship loses to whatever has a deadline

Grant reports have dates. Board packets have dates. Thanking a donor properly does not, so it is the work that slips first, every single time, on teams that are already carrying more than they have support for.

04

Nobody trusts the numbers in the board packet

Retention gets recalculated by hand from spreadsheets that disagree. The figure in the deck does not match the figure in the database, so the conversation becomes about the data instead of about the strategy.

what we build and run

One donor, followed all the way through.

This is the sequence a single first-time donor moves through once the system is live. Your team writes the voice once and approves it. After that it runs whether or not anyone has capacity that week.

  1. 34 seconds

    Thanked

    The gift is matched to the right supporter and a personal thank-you goes out in your voice, not a receipt.

  2. Day 3

    Shown the outcome

    One short note about what the money did. No ask attached. This is the message almost nobody sends and it is the one that earns the second gift.

  3. Day 30

    Invited back

    A second ask timed to how this donor behaves, referencing the program they already funded.

  4. Day 90

    Offered the recurring path

    Donors who have given twice are invited to give monthly, where retention runs in the high seventies.

  5. Always

    Escalated when human

    A major gift signal, a complaint, or anything ambiguous stops the automation and goes to a person with the full history attached.

who we do this for

WORK+SHELTER

Delhi, India/Nonprofit with a manufacturing arm/All-volunteer staff

WORK+SHELTER employs and trains women in Delhi living in poverty, many of them survivors of domestic abuse. Paid work at fair wages, savings programs and interest-free loans, family counseling, English classes, and computer training. The nonprofit runs on volunteers, alongside a manufacturing business producing apparel, accessories, and reusable packaging.

Two organizations sharing one operation. Donors and customers, program outcomes and production orders, volunteers and staff, all tracked in different places by people giving their evenings to it.

What we built and run for them

A custom command center
One place to run the whole operation end to end, instead of the work living across separate tools that never agreed with each other.
Automation across the operation
The follow-ups, handoffs, and updates that used to depend on somebody remembering now happen on their own, including on the weeks nobody has an evening spare.
Design work
How the organization presents itself, so the public face matches the seriousness of the work behind it.
Operating strategy
What to run, in what order, and what to stop doing. The hardest part is rarely the tooling.

Built for a team with no spare capacity at all, which is the condition this whole page is really about.

the fair questions

What a board will ask before they say yes.

Will our donors be able to tell?

They should not, because the words are yours. Your team sets the voice and approves the templates once. The AI does the remembering, the matching, and the timing. It never invents a fact about your programs, and anything it is unsure about goes to a person instead of being guessed at.

Is our donor data safe?

Your data stays in your systems. We work alongside the CRM you already have rather than replacing it, every change is logged with who or what made it, and nothing is used to train a public model. Security is the sector's most cited AI concern and it should be.

We have no capacity to run another tool.

That is the entire point. This is not software we hand over with a login and a training video. We build the system around your operations and then we run it, which is why it keeps working in the weeks your team has nothing left to give it.

How is this a defensible use of donor money?

Because it is aimed at the cheapest revenue you will ever raise. Keeping a donor you already have costs a fraction of finding a new one, and the second gift is the single conversion that turns a one-time giver into a long-term supporter.

start

Bring us your retention number.

Twenty minutes. We look at your own donor file, work out what a first-to-second gift improvement is actually worth to you in a year, and show you the sequence that gets it. You keep the plan whether or not you work with us.

Free30 minutesYou keep the written planYou talk to John, not a rep