One command center for every Innovation Accelerator agent.

Select an agent, pull its instruction set and input files from SharePoint, execute it on Claude, and publish the output artifact back — without copying a single file by hand.

Open command center View execution history

SharePoint is the source of truth

Agent folders, instruction sets and input files are discovered dynamically. Add a new agent folder and it appears — no code change.

Claude does the processing

Pick Sonnet or Opus per run. Instructions and every input document are packaged and sent to the Anthropic API automatically.

Output lands where it belongs

Generated artifacts are previewed in-app and written straight back into the agent's Output folder in SharePoint.

Full execution audit

Every run records the agent, model, files used, token usage, duration, status and operator.

Execution pipeline

Instruction Set  +  Input Files      (SharePoint, discovered dynamically)
        |
        v
   Claude model                       (Sonnet / Opus, selected per run)
        |
        v
   AI processing                      (instructions applied to every input file)
        |
        v
   Output artifact                    (previewed, then written to /Output)