AI agent development
I help you decide whether an AI agent fits the work, and which approach: a custom-built agent, an open-source framework or a managed platform. I then design and connect the solution to your existing systems. A chatbot answers questions. An agent can also plan and carry out multi-step work. On the steps that matter, the agent asks a person to confirm.
A good agent has one job, defined tools, a trail you can read and a person who confirms consequential steps. I do not represent any platform: I choose the one that fits your task.
What the development includes
- Choice of one bounded job and a map of its risks
- Integration with your systems (for example Microsoft 365, CRM or a document store)
- Human approval on consequential steps
- A work log, monitoring and observability in production
- Handover: how the agent works and who maintains it
Three paths
Custom-built agents
I build the agent around your task and systems when a ready-made platform does not cover the need.
Open-source frameworks
I work with open-source frameworks such as Hermes Agent and OpenClaw when you want control, customisation and the option to run them on your own server.
Managed agent platforms
I help you put managed platforms such as Microsoft Copilot Studio into use when you want a ready environment and less upkeep.
Who it is for
Organisations with repeating, text-heavy work such as classifying requests, preparing documents, looking up data or filling forms, where a person should confirm the consequential steps.
How I work
I put an agent into use only once it has an owner and confirmation points.
Book a 30-minute call
In a half-hour conversation we look at where AI would help your work, and agree the next step.
Book a 30-minute call