Anthony has been running Vanda Coatings, a Cardiff-based commercial recoating business, since completing a management buyout in 2003. Over more than two decades, the company has worked on high-profile contracts including sites for BAA, Canary Wharf, and Google.
In parallel with running the business, Anthony completed an MSc in Computing at Cardiff University. That combination of hands-on business management and technical education meant that when AI tools became genuinely useful, he was well-placed to assess and build them.
Before working as an AI consultant for other businesses, Anthony built and deployed Python-based AI tooling inside Vanda Coatings. The tools address real operational problems and are in daily use by the team. That matters because the consulting comes from someone who has done the work and measured the outcomes, not someone who has read the case studies.
Read the full Vanda Coatings case study
Read the research report on founder stress as a business risk
Most AI consultants come from one of two places: technology or strategy consulting. Very few come from running an SME. That gap shows up in the work.
Decisions made under time pressure. Limited IT support. Staff who need tools that are simple enough to use without training. These constraints are obvious to someone who runs a business, and not always obvious to someone who consults for them.
The AI tooling built at Vanda Coatings has to work in the real world. If it does not save time or causes problems, it does not get used. When Anthony evaluates a tool or approach for your business, it goes through the same test: will it actually work in a business like yours, or does it only work in a demo?
The MSc in Computing and hands-on Python and cloud experience means the advice is grounded technically. But it is always filtered through "does this make sense for a business with your resources", not "does this work in an enterprise context".
Most SMEs that need digital improvement do not have an automation problem. They have a process problem, a data quality problem, or a people adoption problem. Automation applied to a broken process produces a faster broken process. The diagnostic has to come before the build.
Business process automation (BPA) works where a process is clearly defined, consistent, and repeated often enough to justify the cost. Robotic process automation (RPA) is the right tool where two systems that should connect do not: where someone copies data from one place to another every day because nothing joins them up. AI adds value where the input is unstructured: documents, images, incoming queries that need interpreting before anything else can happen.
Getting that diagnosis right before recommending anything is what separates a system that saves 5 hours a week from one that nobody uses three months after delivery. The work starts by establishing which of those is actually in play, which processes are stable enough to automate, and which need redesigning first. That thinking comes from running a business with real operational complexity, not from reading frameworks about it.
Digital maturity, how well a business uses technology to deliver consistent and scalable operations, almost always breaks down in the same places: operations and delivery, financial visibility, systems integration, and people capability. The diagnostic maps all four before anything is recommended.
"I spent years watching businesses get sold AI tools they were not ready for, or avoid them entirely because nobody could explain what was actually worth doing. The gap is almost never the technology. It is the process thinking that should happen before you touch the technology. That is what this practice is built around."
This AI consulting practice is based in Cardiff with a primary focus on UK SMEs. Most client work is delivered remotely. Location is rarely a barrier. On-site visits across the UK are available for implementation projects where they genuinely add value.
Based locally? See the Cardiff AI consultant page for more specific detail on local engagements.
I write about AI and automation for small business owners on Substack. Practical notes from real projects, with no hype.
Read on SubstackAn initial call takes about 30 minutes and costs nothing. It is a straight conversation about your business and what you are trying to achieve.
Get in touch