Services

Five ways in. One layer.

An intelligence layer built into how a business already operates: the data and context foundation, retrieval, reasoning and agents, the interfaces and workflow integrations people use, and the evals, governance and cost control that make it safe to run. At whatever stage the client is at. And the client's own AI engineering capability, so their team can run and extend the layer without us.

Not a strategy deck
Not a chatbot bolted on the side
Not a platform licence
Not a pilot that never ships
Not a body shop billed by the hour

The five

Each one is a path into the same layer.

Every service carries the offers that fit it and the demos that show it running. The bands are published on the approach page.

01DiscoveryWhere AI pays off in your business, and where it won't. A ranked list with the business case, the eval design and the cost model, before anything is built.AI Signal ScanSignal SprintWhat we build and what it costs 02AI product engineeringProducts with AI inside: planners, writers, analysts and assistants that pass an eval bar and stay inside a cost ceiling.Shape PilotShip BuildShip RetainerWhat we build and what it costs 03Agents and automationAgents that read, check, propose and wait for a person before they act. Built into the systems you already run.Shape PilotShip BuildSustain OpsWhat we build and what it costs 04Data, search and retrievalCited answers over your documents and data, with an honest "not in the policy" when the answer is not there.Shape PilotShip BuildSustain OpsWhat we build and what it costs 05Enablement: BenchYour own AI engineering bench. We set up the stack, then your team ships its first evaluated system during the engagement.Bench SetupBench EnablementWhat we build and what it costs The Layer Path and the bandsDurations, prices and terms