Agents & MCP
Structured business context your AI agents can run on.
Most agents guess at your business from chat history and pasted docs. Canvasm gives them the actual operating map — via MCP and the API — so agent work is grounded, reviewable, and it compounds.
MCP and API access is live for early-access teams — capabilities are expanding.
Inspect the strategy-metric graph
Agents see the real structure — objectives, metrics, actions, evidence, and the causal links between them. Business context as a graph, not a wall of pasted text.
Create and update nodes & relationships
An agent can add a metric, connect an action to the target it serves, or update values — using the same primitives your team uses on the canvas.
Propose dashboards and checks
Agents can draft dashboards, suggest tracking checks, and flag metrics whose instrumentation looks off — proposals your team reviews on the map.
Act in the open
Everything an agent does lands on the shared map where humans can see, verify, and correct it. No black-box side effects.
Why it matters
An operating map agents can use — not just a UI humans click.
When your strategy, metrics, and evidence live in one structured graph, an agent's work doesn't evaporate at the end of a chat. It lands on the map: visible to the team, connected to the numbers, and there for the next agent — or the next human — to build on.
Agents operate inside the same visibility and access rules as people: what a token can't see, an agent can't read — and every change is attributable and reviewable.