Markus Green
Started in 1980, self-taught on my first computer. The first thing it ever taught me: GIGO - garbage in, garbage out. Never once stopped being true, and it's still the whole reason AI projects succeed or fail.
Had a spell selling door-to-door and on the phone. Then years working with some of the biggest names in tech - Microsoft, Applix, Salesforce - and later hands-on with everything from email marketing and CRM systems to building software from scratch.
The tech is only half of it, though. What actually matters is that I know how business works - real business, not theory. Combine that with genuinely hands-on, current knowledge of AI, and that's what lets me bring solutions that actually work to a local business, not just a demo.
Founder, Home Counties AI. AI guru? Not quite - and neither is anyone else. Thirty years in business, then AI - not the other way round. Here's where that actually started.
Read the full story - and why GIGO still matters more than any AI buzzword
Where this all starts
A ZX80 and a cassette tape for storage - not much to look at, but that's where GIGO first stuck.
Why that still matters
Most people selling AI to local business in 2026 won't tell you this: the model doesn't need to be smarter. It needs better inputs.
Understanding your actual business - what's real, what's broken, what "input" even means when you run on phone calls and habit, not spreadsheets - is the entire job. Anyone can type into a chatbot. Knowing what not to feed it is the actual skill.
Before any of this
Door-to-door sales, then telesales, then corporate life at Microsoft, Applix (later sold to IBM), and Salesforce - learning enterprise software from the inside, not just as a user.
Building things
First employee, then General Manager, at Selling People - a UK agency taking overseas tech companies to market here. Then Ellery Communications, running offshore customer service at real scale, up to 8,000 agents worldwide.
None of that is really about the size of the operation, though. Running something at 8,000 agents just makes the same problem impossible to hide from: bad information going in wrecks everything downstream, whether it's one shop or a thousand seats. That's the part that carried over - not the scale, the pattern.
Since then: a lead-generation consultancy, email marketing, and a CRM I built from scratch myself - which is exactly the muscle that later went into designing and building the bespoke B2B platform on the References page. Different industries. Same question every time: what's actually going into this, and is it any good?
Still GIGO
That's the whole reason Home Counties AI exists. An AI agent is only ever as good as what you feed it, and whether a human actually checks what comes back out. Decades of watching real businesses - including my own - get that input right or wrong is worth more than any amount of prompt engineering.
I'm no AI guru, and neither is anyone else - not really. The tools are too new for that word to mean anything honest yet. What I've got instead: a long run of getting inputs and outputs right in businesses that had nothing to do with AI, plus hands-on, current practice with the tools that exist today. Read the full GIGO article for why that matters more than a fake years-of-AI-experience number.
AI helps. The human in the loop keeps it real.
That's not a strapline I picked because it sounded good. It's how Markus IT Support runs, how the AI-assisted B2B platform I designed and built - then sold - was built, and it's the rule for anything I'd ever build for you: agents draft, agents prepare, a human decides. Never the other way round.
Chesham, Buckinghamshire.