AI tools and integrations

Internal AI toolkit

The internal toolkit connects AI agents to company systems and packages workflows for different roles. Teams can start with the tools and instructions relevant to their work. I guide its technical direction while the AI Platform team develops the shared components and integrations.

Server product integration

I built the first version of an integration that connects AI agents to a server product through the Model Context Protocol (MCP). A dedicated team now develops that work further.

Webmind: search for agents

I built Webmind to give my agents a shared way to search the web and read source pages. It searches across multiple engines and can add a paid search route when coverage is thin or a direct agent request needs current information. Its broader search pipeline can expand a query and combine engine rankings, while current-information queries receive different treatment from background reading. That includes lowering the priority of encyclopedia pages for breaking news.

When a query names a product, pages that repeat generic search terms can still outrank the source for that product. Webmind gives matches to those named entities more weight. Repository searches get a targeted follow-up. If the first results mention a file without locating it, a file-specific query can trigger another search. I compare the raw results with the final ranking to tell whether useful sources are missing or ranked poorly.

Direct requests from agents share a monthly allowance for paid search. I made that search selective and added usage counters with request limits. Article extraction has its own deadlines and response-length limits, with a shared cache that respects the length each caller requests. Failed extracts remain visible beside the source links, so an agent can use the available material and see what it could not read. I continue to test those boundaries as the service changes.

Document processing

I extended my personal data tools so agents could use information buried in email attachments. The extraction work covers PDFs and scanned images, along with Word documents and spreadsheets. A test email with real attachments confirmed that the extracted text reached the live service.

Later work added linked-document ingestion and backfills for older messages. I also developed calendar tools with source attribution and duplicate-event handling, so an agent can show where an event came from without listing it twice.

Memory and retrieval

An agent was retrieving too much stored context for a question. Tests showed that most of the excess came from retrieving too many distinct memories. Removing duplicate entries would have made little difference.

I chose a lower retrieval limit based on those results, without deleting stored memories. We checked that the answers in a small test set remained available and kept a way to reverse the change.

Switchyard: development and review

I designed Switchyard to assign development and research tasks to eligible AI models. Other model families review changes, with findings and test results kept in the work record. Defined fallback paths handle unavailable providers.

Related notes

• The cached page was the wrong answer
• An agent was recalling too much