I’ll build a secure AI agent connected to your business tools.

An AI agent that handles the repetitive parts of your work: inbox triage, document processing, reporting, or answering questions from your own data. Built with MCP, tool calling and RAG, and connected to the tools you already use.

Unlike most agent builds, mine are secure by design: the agent gets only the permissions it needs, treats emails and documents as data rather than instructions, and logs every action it takes.

• A working AI agent connected to 1 to 2 of your tools (email, CRM, documents or database) • Least-privilege access and protection against prompt injection • A log of every action the agent takes • Deployment to your environment • 2 weeks of support and adjustments after launch