Your software as tools for AI agents
Engineer ApS turns the software a company already runs into tools AI agents can call: rules out of the screens, described APIs and MCP servers, and no rewrite.
Photo: PIX1861 / Pixabay 
Photo: ddzphoto / Pixabay
Most business software was built to be clicked: a screen, a form, a button and a person in front of it. An agent can click too, slowly and by guesswork, and it breaks the day a button moves. Its better path is a described tool: an operation with a name, typed inputs, a clear answer and a permission check. We take the software you already run — the one your business depends on, however old — and give it that path, without stopping the business to rewrite it.
Unpicked: the rules out of the screens
The rules that run a business hide in the software: in a form’s validation, in a stored procedure, in a report nobody dares to change, in the head of the person who built it. We find them, write them down and move each into one layer — often the database itself — where every caller gets the same answer. The screen, the agent and the partner integration then ask the same question of the same rule, and cannot disagree.
Exposed: described tools an agent can call
Each operation an agent should have is offered as a tool named for the work — book a slot, quote a price, release an order — with typed inputs, errors an agent can act on, and idempotent calls so a retry never books twice. The description is an OpenAPI document generated from the code, and an MCP server offers the same operations to the agents that speak it. Authentication is OAuth, so an agent acts for a user without a password changing hands, and only within that user’s rights.
Kept running: no big-bang rewrite
A rewrite stops the business for the length of the project and bets everything on the last day. We open the software one operation at a time, behind the system that already works, and check the new path against the old one until they agree. Slow work moves onto a queue so the agent gets an answer while the work goes on. The screens your staff use keep working throughout, and the old code is retired only when nothing calls it.
An agent believes every word of a tool's description — so the description has to come from the code.
A person notices when a manual is out of date; an agent acts on it.
A contract that cannot drift
A tool description written by hand is out of date the week after it ships, and an agent trusts it anyway. Ours are generated from the code outward — the database, the API, the description, the client — with a check at every step that fails the build when two of them disagree. An agent reading the description reads what the software does today, and a change that would break a caller is found before it ships rather than after.
The work behind it
We have opened up large systems this way. We generated an API contract from the database out, with a guard at every hop, and migrated an API of 760 operations at 100% parity with its published description. We designed a JSON passthrough from PostgreSQL to the API, built a platform in clean layers with its business rules in one place, moved slow work onto a job queue, and built the operator back office that runs it without database access.
How an engagement runs
It starts with a list: what your people do in the software every day, ranked by how often and how much it costs them. We take the first few operations, find their rules, and expose them as described tools behind tests and permissions, running beside the screens rather than instead of them. Then the next few. Your team keeps the code, the tests and the description; the tools can be called by an agent built for your trade, by the assistants your staff already use, or by a partner’s system. It is bought like the rest of our work, on the terms in what it costs.
If your software only works when somebody clicks it, tell us what it is.
Your software already does what an agent needs — it has to be asked in words an agent can read.
It starts with a list of what your people do in it every day, ranked by what it costs them.
