VetBid
An AI business-development platform that reads federal solicitations and tells veteran-owned firms which ones they can actually win.
Visit vetbid.usThe problem
Service-disabled veteran-owned small businesses compete for billions in annual federal set-aside awards, almost always as owner-operators without a dedicated capture team. The bottleneck is not finding solicitations — SAM.gov publishes them all — it is reading them. A single RFP runs to hundreds of pages, and deciding whether to bid means matching dense compliance language against your own past performance and personnel. Get it wrong and you burn weeks on a bid you were never positioned to win.
What we built
- 01
Continuous ingestion of SAM.gov opportunity and entity data through scheduled background jobs, normalised into a Postgres model built around the way capture teams actually work.
- 02
An RFP analysis agent that parses uploaded solicitations up to 200MB, extracts requirements and evaluation criteria, and matches them against the firm's stored past performance and key personnel.
- 03
A win-probability model that scores fit rather than guessing outcomes — surfacing which specific requirements a firm meets, which it does not, and where a teaming partner would close the gap.
- 04
Generated artefacts that shortcut the manual work: capability statements, personnel bios, incumbent intelligence, and a recompete calendar built from award history.
- 05
A provider-agnostic AI layer so the underlying model is a configuration value, not an architectural commitment.
Decisions worth noting
Swappable model layer
Every AI call routes through an internal abstraction with per-provider adapters. Switching the reasoning model across the whole product is an environment variable, which matters in a field where the price-performance frontier moves every few months.
Durable background jobs
Solicitation ingestion and document analysis run as durable workflows rather than request-bound work, so a long analysis or a third-party outage never drops a job or times out on a user.
Tier gating built into the domain
Usage limits, seat counts, and feature access are enforced in the data layer rather than the UI, so entitlement logic has exactly one source of truth.
More work
Next Door Notice
A geospatial monitoring platform that watches every planning application in London and tells people when one lands near them.
AwesomeGene
A fully automated video production pipeline: one brief in, a finished, published video out, with no human in the middle.
Olim Paveway
An AI planning tool that turns a fragmented, high-stakes immigration process into a personalised, document-backed plan.
Tell us what is slowing you down
A short conversation is usually enough to tell whether this is a build, an automation, or something you should not do at all. We will tell you which.