A senior AI advisory practice, kept deliberately small.
We are Navitec, a founder-led AI advisory practice based in West Sussex. The practice was established in 2025, with a working ecosystem of associates and partners when an engagement calls for it.
We help organisations think clearly about AI in environments where getting it wrong has weight.
Our work sits where governance and execution meet. We are usually engaged because the question is not only “what is the policy” but “what would the policy actually do, in this organisation, this year, against these constraints, at the speed AI is moving.” Both readings of that question matter. Our practice carries both.
A typical engagement is a senior pair of eyes on AI strategy, governance, or a particular programme. We arrive with evidence of similar work, an honest read of the situation as we find it, and a recommendation we can defend. Recent work has included guiding a client through ISO 42001 certification.
Three shapes of work we increasingly see.
A growing share of our engagements starts after the first wave of AI work has already happened. The shapes below are the ones we see most often.
AI work that needs to land somewhere defensible
AI agents, copilots, and automations get built quickly when the tools are this good. They stop being quick when the questions become “who signs off on what it does”, “what happens when it gets it wrong”, and “how do we explain it to the board”. We are usually called in at that point. The work is to take what exists, set it on a defensible footing, and decide what to keep, what to rebuild, and what to retire.
Workflows that worked once and need to work every time
A proof of concept that runs in a demo is not the same thing as a workflow that runs in production for a year. The difference shows up as flaky outputs, drift, silent failure, and small problems that compound. The work is to identify where the workflow is fragile, where it is mis-scoped, and what has to change before it can carry the load it was built to carry.
Systems built quickly that need to be run for years
Code written under time pressure is rarely the code anyone would write deliberately. It tends to work, until it does not. The work is to bring the system to a state where the team running it can read it, change it, run it, and explain it. Not a rewrite from scratch. A measured pass to make the system production-ready, documented, and operationally defensible.
Small practice. Named ecosystem. No aspirational org chart.
We are the size we are. Most engagements are delivered directly. Where additional expertise is called for (public sector procurement, compliance and data protection work, specialised technical implementation) we bring in named associates from a working ecosystem. We tell you who is doing what before the engagement begins.
There are problems we do not solve. We will say so.
We do not run model training. We do not audit to a published standard. We do not certify. Where any of those is what you need, we will say so and point you somewhere better.
Our deference is part of our credibility. A practice that names its boundaries tends to deliver more inside them.
If any of this fits, the way in is a conversation.
A short call. Thirty minutes, no preparation needed. We listen first, then say where we think we can help, where we cannot, and what a useful next step might be.
Get in touch