How we run agents after they go live
The first 30 days of agent care are included: we watch the agent's runs, fix what breaks and tune it on your real cases. After that you can keep us on a monthly care plan, build out the rest of the workflow, or run it yourself.
Why agents need care
Models update and workflows drift. An agent that worked last quarter can start missing cases this quarter unless someone watches its runs.
What care covers
- We watch every run.
- We fix what breaks and tune the agent on your real cases.
- A person on your side approves anything the agent is unsure about, while the rules are still being tuned.
After the first 30 days
- Agent care: a monthly plan, sized to the system.
- System build-out: €5,000 to €15,000 for the full agent team around the workflow.
- Run it yourself: you own the code, the prompts and the accounts.
When Henry is away
Nicholas Vantieghem, our Head of Operations, is trained to make every fix and update Henry makes, so care does not stop when one person is out.
We run our own companies this way
Our cybersecurity company SYBA.IO and the iOS App Factory run on the same agents we build for clients. See what is a Personal CEO Agent.
Related
- How we build an AI agent in 30 daysKickoff, build on your real cases, a supervised run, then live. How a Phase 1 agent goes from first call to production in 30 days.
- What is agent memory?Agent memory is the notes an agent writes while it works and reads at its next run, so corrections stick. What is kept and what never is.
- What is an Agent Enterprise?An Agent Enterprise is a full team of AI agents set up and run for a company, with a person accountable for each job and us tuning the runs.
- How AI agent projects are pricedThe pricing model in words: a fixed Phase 1 from €2,500 with a 30-day guarantee, then a build-out and a monthly care plan sized to the system.