← All work
Grants · our own product

The whole grant in one place, for every role

A grant passes through a grant writer, a program director, leadership, and finance, and each keeps their piece of it in an email thread, a spreadsheet, or someone else's system. Granary gives every role the same story, from proposal to renewal.

GRANT WRITER LEADERSHIP PROGRAM TEAM THE GRANT ONE RECORD EVERY PART EVERY PART OF THE GRANT · EVERY ROLE · ALL THE WAY THROUGH ✕ NOT IN EMAIL THREADS, SPREADSHEETS OR EXTERNAL SYSTEMS
Case study

How it broke down

Problem

Nobody owns the thread from proposal commitment to work plan to delivery to funder report to renewal. The proposal lives in the grant writer's files, the award letter with finance, the deliverables in a pasted spreadsheet, outcome data in an external system, and the stories in email. Each role sees its own fragment, so the deliverables go unrevisited until a report is due, and the writer spends the final weeks reassembling the grant from everyone else's pieces. Funders judge renewals on the relationship, and a program that learns it is off track at month eleven can only explain.

Constraint

Each role needed a different view of the same object, not another system to update. A program director sees one plan for their program. A grant writer sees, per funder, what that funder is owed in that funder's phrasing. Leadership sees what is at risk across every grant. A consultancy sees across all of its clients, with every organization its own tenant. And program staff who would never log in had to be full participants from their inbox.

Approach

One record per grant, built as an append-only story from proposal through renewal, with each role's screen a view over that same record. Commitments are drafted from the proposal and award letter and corrected by a grant professional rather than written from a blank template, and the corrections are kept as data. The program owns its plan; the product sets its own checkpoints between the funder's dates; evidence is captured at the point of work, in the app or by replying to an email, and lands on the commitment it answers. Reports assemble from that trail, and anything a funder will see passes an approval gate first. Freshness is surfaced before progress, so the record says "nobody has entered the last six weeks" before it ever says "40% behind".

Outcome

The emails, spreadsheets, and exports that used to carry a grant between roles collapse into one record that every role reads and writes in its own way. In production for an invite-only pilot with a grants consultancy and the nonprofits it serves. The report is a read of what was captured along the way, not a scramble in the final month.

Built the way we work

This one is ours

Granary is a product we built end to end rather than a client engagement, which makes it the clearest example of the process on How we build: scoped with a grant professional, every screen approved as a mockup before code, built in waves by our agent team, with every change passing an expert review gate before it reached production. Start from the constraint, measure before believing, ship.

Start

Have a process scattered across inboxes and spreadsheets?

Your domain expertise ⊕ our engineering team. A 30-minute call is enough to tell whether this is a fit — bring the problem, no prep needed.

Book a call

Founded by scientists turned AI leaders — national-lab research backgrounds, then senior data science and AI roles shipping production systems across logistics, security, and adtech.