The story · Layer 02 of 04
The context layer.The company remembers.
Once the data agrees, the company can start to remember. One context layer for the whole company, organised by client, project or team, filled on a schedule from everyone's work, filtered, dated and sourced, and read by everyone and by every agent.
How it works
Created once. Filled on autopilot. Read by everyone.
The space exists.
A new client, project or team gets its own space in the context layer, from your template, with access set by role. Nobody sets it up by hand.
Everyone’s work fills it.
Calls, emails, documents, chat and data from the whole team are read on a schedule and filed with a date and a source. Private conversations and sensitive data are filtered out before anything is written. It is always up to date.
The team asks it.
Plain English in, the answer with its source out. In Claude, ChatGPT or a bespoke web app. Each person sees only what their role allows.
Every agent reads it. And writes back.
The monthly report, the reply to the client and the next proposal all start from the same facts. The next project reads the last three.
What you get
- Your context-layer template, shaped with your team
- Access by role: people and agents see only the clients, projects and teams they are allowed to
- Scheduled filling, with private conversations and sensitive or personal data filtered out before anything is written
- Piloted on two or three live clients or projects
How long
- 10 to 50 peopleabout 3 weeks
- 50 to 150 people3 to 4 weeks
- 150 to 300+ people4 to 6 weeks
Alongside data foundations. Everything else runs on top of it.
Price
Agreed after discovery, before we start. Your cloud and AI usage costs are paid to your vendors directly.
People stay in the loop. You decide what belongs in the context layer and who sees what. Nothing sensitive is pushed into it: the filter runs before anything is written, and access is set by role.
