A senior engineer works alongside your subject-matter experts — remotely, through workshops and working sessions — until we know the work as well as they do. Then we build the AI-powered system around how your people actually operate. Workshop, proof of concept, production — in weeks.
Most consultancies run a workshop to decide which fraction of your wish list is affordable. We run one to capture the whole thing, then build it. Forward-deployed means embedded in your problem, not in your office: workshops and working sessions are remote, and we go as deep with your experts as it takes to get every detail that matters.
A remote working session with the people who do the work, and one rule: describe the system you would have if nobody said no. We map how the work actually happens, then capture every screen, rule and shortcut your team wants. Not the three or four we can afford — all of it. AI-accelerated engineering is what makes that a promise we can keep.
The pieces that depend on AI — summarising, matching, extracting, routing — run as a proof of concept against your real cases and are judged on measured results. The system you asked for isn't up for debate; how each AI piece earns its place is.
Requirements as a clickable application, with your data in it, that your users try before anyone commits. It becomes the blueprint for the build — or the baseline you hold every vendor to.
The winners become AI-powered production systems: integrated around what you own, extended with bespoke modules, or replacing the legacy system outright. On your stack, ours, or Palantir.
From prototype to production
A Living Requirements Document starts life on day two as a rough prototype in your users' hands. Their comments land on it; by day five the comments are the requirements. In week three it runs against real files as a proof of concept and is judged on measured results. The screens your people shaped are the screens that go live.
Nothing is thrown away between stages. The prototype becomes the POC becomes the production system, which is why a Treefrog build lands in a quarter rather than a year.
And if you end up buying instead? The same document becomes your scorecard →Doing your first AI build in-house?
AI tools have made it possible for an operations team to build the system they would once have bought. The difference between a build that works and one that quietly becomes the next key-person risk is usually an experienced engineer in the room at the right moments.
Our guided-build engagement puts a forward-deployed engineer alongside your team: architecture and data design up front, review of what your team builds, and the security, testing and hand-over discipline that turns a clever prototype into something the business can rely on.
Read the story behind this: Is Software as a Service Dead? →Data model, integrations and AI boundaries designed by someone who has done it before, so your team builds on solid ground.
Weekly reviews of what your team has built: security, performance, maintainability, and the edge cases that bite in production.
Tests, documentation and deployment so the system survives the person who built it.
No. Forward-deployed means embedded in your problem, not your office. Workshops and working sessions run remotely and go as deep with your subject-matter experts as the problem requires.
A first Living Requirements Document — a working prototype with your data in it — is typically in your users' hands within two to five days of the workshop.
That is what the proof of concept is for. The AI pieces are tested against your real cases and judged on measured results before the build commits to them. The system you asked for is built either way; how each AI piece earns its place is tested.
Your own stack (Azure, .NET, SQL Server, Python, low-code), Databricks, Palantir Foundry and AIP, Power BI, and our Foundation platform for (re)insurance pricing. The choice is made for the problem.
About a quarter from first call to live system: workshop in week one, prototype by week three, proof of concept by week four, production build over the following six to eight weeks.
Yes. Our guided-build engagement puts a senior engineer alongside your team for architecture, weekly review and hand-over discipline, so a clever prototype becomes a system the business can rely on.
Thirty minutes with a senior engineer. You leave with a scoped first build and a clear read on build vs buy, whether or not you hire us.