About
Richard Carter. Based in Leeds.
I've spent 30 years changing how work gets done inside businesses — small ones, large ones and a few very large ones. The technology has almost never been the hard part.
What actually goes wrong is this. Something gets set up, it works for a fortnight, then people drift back to the old way. Or one person understands it, and then that person leaves. Or half the team adopts it and half don't, so now there are two ways of doing the same job instead of one. Or it fixes something that was never really the bottleneck, while the actual bottleneck carries on untouched.
None of that is a technology problem. It's picking the right thing to change, making the new way easier than the old way and coming back three weeks later to check whether anyone is still doing it.
That's the part I'm useful for.
How I got here
I started as a developer in 1996 at Binary Vision in London, building websites and CD-ROM projects for clients including AstraZeneca, the Post Office and the Royal Air Force. In 2001 I moved to Leeds to head the design department at Coolpink, a team of eight.
From 2006 I was Development Director at bolser for eight years — a team of 30, fixed-price digital work for BBC, Sony, Virgin and EE, and something over 30 tenders won. The agency picked up a run of industry awards over that period, mostly for mobile work. I then spent 18 months as Development Director at Flashtalking, running a department of 80 and moving releases from an annual cycle to a weekly one. That work was shortlisted three times at the 2015 Agile Awards, including Most Influential Business Leader.
Since 2015 I've worked almost entirely on UK government digital delivery: HMRC, the Department for Work and Pensions, the Ministry of Justice, the Care Quality Commission. That included delivery lead on the Customs Declaration Service — the systems handling imports and exports after Brexit — inside a programme budgeted at £270m, and a prisoner education platform rolled out into two prisons.
I'm currently on a government pilot team building software AI-first.
AI in practice
What I build
I don't only advise on this. I build with it.
A live trading system running on my own money — a deterministic pipeline for the decisions that cannot be wrong, with agent workflows handling analysis, code and fixes around it.
A multi-agent editing system that proofs and edits against a defined voice, because "it doesn't sound like us" is the most common reason an AI draft gets binned.
A delivery forecasting assistant — Monte Carlo completion forecasts, throughput dashboards, scenario modelling — packaged so other people install and run it themselves, across three different model providers.
Everything those systems produce is a draft or a report for a person to approve. Nothing acts on its own. That isn't caution, it's the design.
Deliberate retraining
Trained, deliberately, for 16 years
I've retrained roughly every three years since 2010. Prince2 in 2010, Scrum Master in 2013, scaled agile in 2015, advanced Scrum and Scrum@Scale in 2018 and 2019, agile coaching in 2020, machine learning at Stanford in 2024 and Anthropic's engineering courses since. Each time into whatever was about to change how work got delivered.
On the Anthropic side I've worked through both ends of their course set — AI Fluency for Small Businesses at one end, Model Context Protocol, agent skills and subagents at the other. That combination is the job: understanding what a 40-person firm actually needs, and being able to build the thing that does it.
I've also taken Google's AI Professional Certificate — the one your team is probably being pointed at. It's good, and it will make individual people faster at their own jobs. It will not change how work moves through your business. That gap is the whole of what I do.
AI is the current one. It won't be the last, and I intend to still be here when it isn't.
The through-line
Before it was called AI, it was called flow
Fifteen years of formal training in measuring where work stalls, what a handoff costs and how to shorten a cycle time. That is what the Scrum, Kanban and agile coaching qualifications actually certify, and it is what the assessment does.
The assessment doesn't look for AI opportunities. It finds where work stalls, then asks whether AI is the cheapest way to remove the stall. Sometimes it isn't, and I'll tell you so.
The technology changed. The question didn't.
How I work
I bring the technical side. You bring your business. I work it out with you.
That isn't a fallback position, it's the method. I've never had prior knowledge of a client's industry and it has never been the problem. A plumbing firm and a law practice handle an incoming enquiry in almost exactly the same way once you strip out the vocabulary. It's why I don't specialise in a sector and why the assessment opens by asking you to walk me through how a job actually moves through your business.
In their words
What others say
Keith Varty
Ryan Yates
Jason Simpson