Who it's for
Companies with a workflow that's too specific for packaged software and too expensive to keep doing by hand.
- A document your team rebuilds from scratch dozens of times a month
- A field process that produces the same data twice, once for compliance and once for billing
- A business development routine that depends on someone checking permit sites every morning
- A modeling or scenario tool that lives in a fragile spreadsheet
- A vendor tool that covers most of the job, with the rest done in email
What you get
Workflow design
We map the process step by step with the people who do it, and decide where software helps, where AI helps and where a person stays in charge.
Working software
Web apps, progressive web apps for field use, internal platforms and dashboards, built and deployed in your environment or ours, as agreed.
AI agents
Agents that carry out multi-step tasks, such as monitoring sources, gathering records, drafting documents and flagging exceptions, with a person approving what goes out.
Integrations
Connections to the systems that hold your data: ERP, CRM, PIM, document storage and e-signature, so the tool works from live records.
Admin and audit layer
User roles, change history and audit logs, so you can see who did what and when.
Documentation and handover
Setup, architecture and operating documentation. Ownership and licensing of the code are agreed in writing for each engagement.
How it works
Step 1: Scope
We define the workflow, the deliverables for each phase, and what we need from you: data access, subject-matter time and decision-makers.
Step 2: Prototype
A working prototype on real or representative data, early, so you can react to something real before the full build.
Step 3: Build in phases
Fixed-fee phases with defined deliverables. Anything outside scope goes through a written change order.
Step 4: Launch and hand over
We deploy, train your users and hand over documentation. Ongoing support is available through Managed AI Services.
Where it fits
A few things we've built, client names withheld.
- Declaration of Conformity generatorFor an industrial manufacturer: a progressive web app that looks up a product by QR code or serial number, pulls data from a Salsify-class PIM, produces the EU Declaration of Conformity in multiple languages, and keeps an admin and audit trail.
- Obligation engineA prototype for an erosion-control contractor: a multi-state stormwater rule engine, rain-triggered inspection scheduling, and one field capture that produces both a compliance record and a billing record, with GC signature capture. The ledger layer isn't tied to any one ERP.
- Agentic lead findingFor contractors: agents that monitor land disturbance permits, ITBs and approved projects, identify the owner and contacts, build a call list and draft intro emails that reference relevant past work.
- ESOP scenario-modeling platformA scenario-modeling platform built for an ESOP advisory firm.
- Commercial insurance sales hubA sales hub MVP scoped for a commercial insurance agency on its existing platform.
Frequently asked questions
Who owns the software you build?
It depends on the engagement. Some builds are fully client-owned, some use components we reuse across clients, and some are a mix. We agree ownership and licensing terms in writing before the build starts, so there are no surprises.
How do you keep a custom build from going over budget?
Fixed-fee phases, defined deliverables, and your dependencies written into the scope. When something changes, we write a change order and you approve it before we do the work.
What happens if the AI gets something wrong?
We design for it. Agents draft, a person approves. Outputs that go to customers, regulators or the ledger have a review step, and every action is logged.