Agents & Tool Use

This section collects research and evaluations of how our agents choose and use tools, and how well they complete end-to-end tasks with them.

The work here covers tool-call accuracy and search-scope selection, the effect of large tool sets on tool selection (tool overload) and how orchestrators and dynamic skills mitigate it, agentic table workflows, and document generation with skills such as client-ready PowerPoint decks.