Built on doing it, not just advising on it.
Human–AI Systems exists because AI adoption is hard in ways that theory doesn't prepare you for. The practical knowledge comes from having led these programmes directly — in commercial settings, in public sector governance, and at board level.
Why Human–AI Systems exists
Most organisations are at the early stages of AI adoption. They're running pilots. Some of them are working. Fewer are changing how work actually happens at scale.
The gap isn't technology. Organisations have access to excellent tools. The gap is operational adoption — the hard work of redesigning workflows, building governance that is rigorous without being obstructive, and developing the leadership capability to manage teams where AI is part of how work gets done.
Human–AI Systems was founded to help organisations close that gap. The approach is practical and evidence-based, shaped by direct experience of what it takes to move from experimentation to operational adoption — and what goes wrong when organisations try to skip the difficult parts.
"I work with organisations that are serious about adoption, not just experimentation. That usually means starting with an honest diagnosis of where AI can genuinely create value — and designing programmes that are measurable, governable, and built to last."
Human–AI Systems is itself built and run on AI — a Chief Marketing Officer, a COO, a CTO, a CFO, all operating as AI roles within a shared organisational context. Mike has written about what that actually looks like: what works, what doesn't, and what it revealed about AI adoption that advising alone wouldn't have taught him.
Leading it at Verimatrix
At Verimatrix — a publicly listed cybersecurity company — Mike led the company's AI transformation programme. This was not a pilot project or an innovation initiative. It was a programme designed to change how engineering teams actually worked: how they wrote code, reviewed it, managed technical debt, and shipped product.
That meant confronting what most AI transformation programmes avoid: the workflow design questions, the accountability questions, and the governance questions. What happens when AI makes a mistake in production code? Who owns the output? How do you measure productivity when the nature of the work has changed?
Outcomes — Verimatrix AI Transformation
~£1M engineering productivity improvement
~£3M revenue impact attributed to the programme
AI governance framework designed, built and approved at board level in a listed company
The board approval process for the governance framework was, in some ways, the most instructive part. It required translating the operational reality of AI adoption into language that directors could interrogate, challenge and ultimately stand behind. That experience — of making AI governance real at the highest level of a listed company — is directly relevant to any organisation facing similar scrutiny.
Cabinet Member at Cotswold District Council
As a Cabinet Member at Cotswold District Council, Mike holds the digital and climate strategy portfolio — including responsibility for the council's approach to AI. He co-authored the council's AI Policy and oversaw the rollout of Microsoft Copilot to around 250 officers and elected members.
AI adoption in local government is different from commercial settings in almost every dimension: the governance environment is more complex, the consequences of getting it wrong are more public, the resource constraints are tighter, and the workforce is more varied in its starting point with technology. None of that makes it slower because people are resistant. It's slower because the design constraints are genuine — and need to be taken seriously.
Outcomes — Cotswold District Council
AI Policy co-authored and adopted by the council
~250 officers and elected members onboarded to Microsoft Copilot
Governance framework designed for a politically accountable environment
That combination of commercial and public sector experience is not common. It means the governance instincts developed at Verimatrix translate directly into the scrutiny environment of local government — and vice versa.
Enterprise technology at scale
Before founding Human–AI Systems, Mike spent over two decades in enterprise technology — including a senior leadership role at Cisco, where he held responsibility for a £430M EMEA Public Sector P&L. That background — building and leading large commercial technology organisations across major enterprise and public sector accounts — provides the business context within which AI adoption advice is grounded.
The career has run across enterprise SaaS, cybersecurity, public sector digital, and AI transformation. It is not a background in AI research or software development. It is a background in making technology work inside real organisations — and in understanding what leadership, governance and operational design are required to do that.
AI adoption through the eyes of the people doing the work
These are the kinds of problems that get solved when you stop adding tools and start redesigning how work actually happens.
Your best people are buried in bid admin instead of talking to customers.
The proposals take five days. The CRM data is three days stale. Reps prep for discovery calls in the twenty minutes before the call starts. Everyone knows the work should be sharper — but there's no time to make it so.
One mid-market technology company changed how its entire deal cycle operated. Win rate improved roughly 10%. Revenue impact: around £3M.
The tools were already there. What changed was how the work moved between people.
The queue never gets shorter, no matter how hard the team works.
Half the tickets are variations on solved problems. The answers exist — somewhere in a knowledge base with four hundred articles, or in a Slack thread that got buried last Tuesday. The best engineer spends her mornings answering colleagues' questions instead of the hard cases.
A SaaS company redesigned how support actually operated. Low-complexity tickets dropped 30%. Resolution times fell by a quarter. Churn linked to support experience dropped measurably.
They didn't hire more people. They changed how knowledge reached the point of work.
Every engineer has Copilot. Nothing ships any faster.
Individual developers are faster, but cycle times are flat. Tech debt is growing. Nobody can answer whether AI-generated code in production carries copyright risk. The roadmap is slipping and the competitors aren't waiting.
A SaaS company redesigned how engineering work flowed — from code review to architecture decisions. Feature velocity improved 25%. Engineering productivity gained roughly an hour per developer per day.
They stopped adding tools and started redesigning how the team worked.
How the work gets done
Three principles that shape every engagement, regardless of size or sector.
Start with diagnosis, not prescription
Every organisation is different. The right starting point for AI adoption depends on what workflows exist, what governance environment applies, and what the organisation is genuinely ready for. The Radar exists because good advice has to be grounded in reality, not in a standard model.
Governance is a design problem, not a compliance exercise
Well-designed governance enables adoption — it doesn't obstruct it. The organisations that get AI wrong usually have governance that arrived too late, was too vague to be useful, or was never connected to how work actually happens. Getting it right from the start is the whole point.
The goal is an organisation that works without us
Every engagement is designed to leave the organisation more capable, not more dependent. That means building internal understanding, designing governance that the organisation can sustain, and developing the leadership capability to manage human-AI work going forward.
Training and speaking
Alongside consulting engagements, Mike delivers practical AI training workshops and speaks at industry and company events.
AI training workshops
Sector-tailored workshops that give teams practical AI skills — not theory slides. Three standard programmes cover AI essentials, hands-on skills, and adoption planning, each designed around the workflows your people actually use.
Available as half-day and full-day sessions for up to 20 participants, with bespoke programmes for organisations with specific needs.
Explore training →Presentations and events
Mike is available for keynotes, panel discussions and introductory presentations at industry conferences, leadership away-days and company events. Topics cover practical AI adoption, governance, and the realities of building a hybrid workforce — drawn from direct experience, not research summaries.
Sessions typically run 20–45 minutes and can be tailored to your audience and sector.
Get in touch →Human leadership, AI workforce
Human–AI Systems operates the same model it helps clients build. Mike provides the leadership, judgement and accountability. An AI team — each named after a computing pioneer — handles defined roles under his direction. It's a practical demonstration of the Worker model in operation.
Meet the teamStart with a conversation
If you'd like to talk through your situation — where you are with AI adoption, what you're trying to achieve, and whether there's a fit — let's have a conversation.