How I work

1. Understand the process and the result you want. I talk to your team and see how the work really gets done day to day, with your tools, your data and those exceptions that are rarely written down.

2. Pick the right task. I map your systems and flows to pick the task worth handing to AI. And along the way, that same map shows where there is risk under the AI Act, GDPR and security.

3. Agree the scope and the acceptance criteria. We put in writing what gets built and how we will know it is finished.

4. Build and integrate. I build what the task needs on your real data: an AI agent, an automation or software. And I connect it to the tools your team already uses.

5. Validate it with you. We check together that the system meets what we agreed before leaving it to run on its own.

6. Maintain and measure. The build ends on the date we agree, and from then on I maintain it with the scope each project needs. If we also agree on it, you get a monthly report on results and cost.

Until now, buying software meant buying a tool, and someone on your team had to learn it, feed it and check what it said. And when that did not happen, the tool ended up sitting in a drawer.

Now AI can do that work end to end, from qualifying incoming leads to writing the report or checking the case file. And that is how I deliver it, so what you get is the work already done.

This model is called Service-as-Software, a name Foundation Capital gave it in 2024. And in March 2026 Julien Bek, a partner at Sequoia, took it further in “Services: The New Software”.

Sequoia put it in one line: “A copilot sells the tool. An autopilot sells the work.”

An AI system can work really well in a demo and still be of little use if it does not understand how your company works.

So rather than working from the outside, I work as a forward deployed engineer. In other words, alongside your team I get to know how the work is really done and build the solution inside your systems until it works.

The term comes from Palantir, where it named the engineers who work inside a single client, build whatever that client needs and are judged by its result.

And that is also how I work inside Sesame HR, where every team is an internal client.

Before building, we put in writing what counts as done and how it is checked. It could be, for example, what a contact must have to count as a qualified lead, or which questions a report must answer.

We use that standard to validate the system with you before it runs on its own, and it is the same one it is measured against while it runs.

Each proposal separates three things, so you know what you are paying for in each.

The fees, which cover building the system and integrating it into yours.

Third-party costs, that is, the AI models and tools the system uses. And since AI is not free to run, these costs are itemised separately in the proposal.

Maintenance, with the scope the project needs.

And both prices and conditions are set in writing in each proposal, before starting.

Keys stay out of the code, and each agent gets only the permissions it needs. And whatever arrives in an email or a web page, the agent treats as information and not as orders to follow.

GDPR and the AI Act are part of the design from the start, which is when we decide which data is used and where it travels.

There are automatic tests and usage limits, and every change passes a set of checks before it goes live.

What each agent does, why it does it and what it costs is logged. Every piece is documented, tested and connected to the others, and what no longer serves is retired.

Repetitive tasks whose result can be checked, like the ones we saw above with leads, reports or case files.

Work the company already pays someone else to do, or that its own team does, and where there is a clear idea of what doing it well looks like.

Tasks with low liability, because where the law asks for a professional's signature, a person signs.

And bear in mind that an agent is only one of the options. Sometimes an automation or software is enough, and other times what is needed is a rule and a person in charge.