Anthropic’s AI tools
“We pay for Claude, or want to, and it is not connected to anything we use.”
I build assistants, agents and automations for small businesses on Anthropic’s Claude models and the tooling around them.
What it is
Building assistants, agents and automations on Anthropic’s Claude models and the tooling around them.
Tools I use
Claude models
The Model Context Protocol
Anthropic’s open standard for connecting AI to a business’s tools.
Claude’s coding agents
When it suits a small business
- You want capable assistants and agents quickly.
- You are comfortable with a managed service.
- You want the assistant connected to your own tools.
What I do with it
- I design and build connections between the AI and your tools.
- I set up agents with confirmations and limits.
- I run several agents in parallel on real work.
- I put checks around what they produce.
Why me for this
I built about 30 tool servers on the Model Context Protocol for the private assistant.
That project’s development and its daily review are run by Claude coding agents working in parallel.
In the pitch-sequencing app, a Claude model selects and phrases stat angles for broadcasters, and a check removes any percentage or speed that is not in the source facts.
I built a mail connector, as a Model Context Protocol server, that lets an AI assistant search, read and draft email.