06:40Morning
What we do
Agents
We find the tasks an agent can do well and build it to run inside your systems. Each agent knows when to hand the work back to a person.
Inference
We choose the models and decide whether to host them or call an API. Then we keep cost and response time inside the limits you set.
The company brain
We connect the places where your company keeps what it knows. People ask it questions. Agents act on the answers, with the same permissions a person would have.
12:00Noon
A company brain
Most of what a company knows sits in threads, documents and people’s heads. A company brain gathers it in one place, links every answer to its source, and shows each person only what they may see.
What it knows
The written record, kept current and searchable in one place.
What was decided
Choices and the reasons for them, so nobody has to ask twice.
What is happening
Open work across your tools, visible to people and to agents.
16:20Late light
How we work
- 01
A first call
We talk about your work and where AI might help. You don’t need to prepare anything.
- 02
A short study
We spend time with your team and your data, then write a plan that spells out what it costs and what could go wrong.
- 03
Building
We build the first version with your engineers, inside your systems.
- 04
Handover
Your team owns the code and knows how it works. We stay on call if you want us.
18:50Evening
Notes
Essays and small experiments, published as they are written.
20:30Dusk
Find a time to talk.
Tell us what your team is working on. In half an hour we can say whether AI will help, and where we would start.
Book a call