When I finished building JobPilot at the ANCHR AI Native Bootcamp in March 2026, something struck me that I have not been able to stop thinking about since.
JobPilot is an AI agent that automates end-to-end job search — resume tailoring, application tracking, follow-up management, the whole thing. It works well. But what made it work was not any technical skill on my part. I do not write code in the conventional sense. What made it work was that I understood the job search process deeply. I knew where the friction was. I knew what a recruiter actually wants to see. I knew which steps people skip and why those skipped steps cost them.
That knowledge — workflow knowledge — is what turned a generic automation into something genuinely useful.
The developer assumption
Most conversations about AI agents start from a technical framing. Which model to use. Which API to call. How to handle memory and context. These are real questions, but they come second.
The first question is always: what problem are you actually solving, and what does the solution look like from the inside?
A developer who has never hired anyone can build a job application tracker. But they will track the wrong things. They will optimise for the steps that are easy to automate rather than the steps that actually move the needle. They will miss the nuance — the difference between a follow-up email that feels human and one that reads like a bot wrote it.
A recruiter who has reviewed ten thousand CVs, on the other hand, knows exactly what good looks like. They know what triggers a rejection in the first eight seconds. They know how to frame a career gap, when to lead with impact rather than responsibilities, which roles are worth a speculative application and which are not.
If that recruiter learns to build agents — even basic ones — they will build something a developer could not imagine.
What changes when domain experts build
I have seen this pattern now across multiple industries. When I deliver digital transformation work for manufacturers, the people who make the biggest impact are not the IT managers. They are the line supervisors who have run the same process for fifteen years and can immediately tell you where the data is wrong, where the system fails at 3am, and why the third shift always produces different numbers than the first.
Give those people the right tools, and they do not just automate tasks. They redesign processes from the inside. They catch the edge cases the consultants miss. They build systems that actually survive contact with the real world.
AI agents are the same. The leverage is not in the technology. The leverage is in the expertise that gets encoded into the agent.
What no-code actually means
No-code does not mean no thinking. It means the thinking stays where it belongs — with the person who understands the problem.
When I work with business owners and non-technical professionals in my AI Agent Workshop, the biggest shift is not learning a new tool. It is realising that they already have everything they need to build something powerful. The tool just needs to be pointed in the right direction.
The participants who build the most useful agents are the ones who come in with a clear picture of a workflow they know deeply — a sales follow-up sequence they have run manually for three years, a client onboarding checklist they have refined across forty projects, a content approval process they have fought with every week.
Those people build agents that work on day one. Not because they are technical. Because they are expert.
Where to start
If you are sitting on years of workflow knowledge and have been waiting for someone to tell you that you are allowed to build with it — consider this that permission.
Pick one workflow. The one that takes you the most time and that you know better than anyone else. Map out every step, including the ones that only exist in your head. Then come find me.
That is where we start.
Want to talk about this?
Whether it sparked a question or you want to explore how it applies to your work, I am happy to chat.
Get in touch