Fractional AI Director — AI-Si.com
Simon Steggles, Fractional AI Director, Birmingham UK
AI governance, AI strategy, and hands-on AI implementation for UK SMEs and local councils. Based in Birmingham. Operating UK-wide.
I spent the first six years of my working life in the Royal Navy. After that I built a second career in finance and technology directorships — Finance Director, Non-Executive Director, Chairman — across manufacturing, professional services, and tech businesses in the Midlands and beyond. What I kept seeing, once AI started becoming something organisations actually deployed rather than just talked about, was the same pattern repeating: a consultant would come in, produce a strategy document, and leave. The organisation would be no closer to having working AI than when they started, just considerably lighter in the budget. So about five years ago I stopped advising from the outside and started building from the inside. I work directly in client environments now. I write code, I deploy things, and I stay until they actually work.
AI-Si.com is the trading name I work under. Based in Sutton Coldfield, Birmingham. Engagements are UK-wide, mostly with SMEs and councils that need senior AI leadership without a permanent hire.
Here are three engagements that illustrate what the outcomes actually look like. The council work recovered over £300K in budget at risk, with no GDPR incidents across the full engagement. The West Midlands manufacturer hit a 43% operating-cost reduction. And the law firm's AI-assisted analysis turned up £2.1M in additional revenue they hadn't known about. Anonymised case studies for each; named references on request.
Board-level AI leadership delivered
UK public sector, NHS, legal, manufacturing, and recruitment
Saved for a UK metropolitan council
GDPR incidents across all client engagements
Why clients retain Simon
Why UK organisations retain Simon Steggles as Fractional AI Director
The clearest example I can give you: a West Midlands manufacturer had paid £120K to a boutique consultancy for an AI strategy. Properly researched, well-presented, eighty pages. Twelve months later they had the document and nothing operational. When I came in, the operations director's first question was why the document's lead recommendation — a computer vision system for production-line quality control — hadn't gone anywhere. The answer was that no-one had checked whether it could talk to their existing manufacturing execution system. It couldn't, not without a middleware project the strategy document hadn't costed. That's the gap I'm describing. Working out where AI creates real value — that part is maybe 20% of the engagement. The other 80% is the getting-it-working part: deploying something real, making sure the team can run it without me, and making sure it hasn't broken six months after I've left. That's what the eighty-page strategy document didn't cover.
Direct engagement — no junior consultants
Every engagement is run by me — not an account manager, not a junior who got the file handed to them after the first call. When we speak during diligence, that is the same person who builds the thing and delivers the board report.
Implementation included — no strategy-and-exit
I write the strategy and I implement it. Those are not separate engagements with a handover gap in between. When the strategy document lands, the implementation has already started.
No vendor bias — fully independent AI advice
Three vendors have approached me about referral agreements since I started this business. I turned all three down. It's not complicated: if I'm collecting a referral fee from a platform vendor, the recommendation can't be fully independant — even if I try to make it so. The fee isn't a lot of money relative to what an engagement is worth. The independence is.
Monthly rolling — no long contracts
Engagements run monthly with a 30-day exit clause. I've structured it this way deliberately — if the engagement isn't delivering, you should be able to leave without a fight. The contract continues as long as it's generating real value.
Fast deployment. Measurable milestones.
The first month ends with a board-ready report and at least one thing that actually works. The six-month discovery phase that produces a slide deck and no working software is not something I offer.
UK-based. Public sector security cleared.
I hold current NPPV3 clearance — the vetting level required for substantive work with police forces and local councils. I started my career in the Royal Navy vetted to Positive Vetting Category 3. For public sector clients, the security background is not cosmetic.
Professional background and AI credentials
Years of board-level leadership
Finance Director, Non-Executive Director, Chairman — across technology, manufacturing, and professional services. The P&L accountability and governance experience I bring to AI engagements isn't borrowed from a course; it comes from having sat on those boards during the decisions that actually mattered.
Years of hands-on AI implementation
The AI tools I recommend are ones I've used in anger, not ones I've seen demoed. I build prototypes against real client data, test whether they hold up under operational conditions, and document what breaks before I recommend anything. The case studies on this site are from actual engagements, not constructed as examples.
Hands-on AI builder
Most AI advisors have strong opinions about tools they've never built anything with. I can tell you exactly why a particular LLM integration breaks down at scale, why retrieval-augmented generation works for some document types and not others, and which vendors routinely oversell what their product actually does. I know because I've tested them.
Accountable through delivery
Systems that never go live don't save anyone money. I stay until yours does — and until your team can run it without me. The engagement closes when that is true, not when the document gets signed.
What makes Simon Steggles different as a Fractional AI Director
35 years of board-level experience
When I present an AI recommendation to a board, I'm drawing on having sat on those boards — the ones where you have to explain a £400K project to a CFO who needs to justify it to the audit committee. I know which questions come back and how to pre-empt them. Governance and risk framing, regulatory compliance language, commercial rationale — I've had to produce all of it under pressure. The AI layer is newer; the board dynamics are not.
Personal technical capability
I can build the thing I'm recommending. At board-level AI consultancy, that is rare. It means I can tell when a vendor's technical claim doesn't hold up — and it means no developer can tell me something is impossible when it isn't. The board-level judgment and the hands-on technical capability sit in the same person.
Stays through implementation
There's a moment in most AI implementations — usually around six weeks in — where it looks like the project is failing. I know this because I've been there. On a recent council engagement, the data that had been described to me as clean and export-ready turned out to need five weeks of remediation before the model could train against it. We weren't behind on the implementation; we were behind on a problem nobody in the council had known existed. I stayed. We finished. That's what staying through implementation actually means — not a policy, but a decision you make at week six when walking away would be easier than explaining to the board why it's taking longer than scoped.
Cross-sector AI experience
The manufacturing client with a paper-based process problem and the council with a benefits assessment backlog have different politics, different procurement rules, and completely different risk profiles. But the underlying AI challenges — data quality, staff adoption, governance gaps, vendor selection — overlap more than you'd expect. There are things you learn from doing this across five sectors that staying in one never teaches you.
What is a Fractional AI Director and how does the model work?
A Fractional AI Director provides board-level AI leadership on a part-time, retained basis. You get the same strategic oversight, governance accountability, and hands-on implementation expertise as a full-time Chief AI Officer — without the £150,000 to £300,000 annual salary commitment.
Why the fractional AI Director model works for UK SMEs and councils
A full-time Chief AI Officer costs £150,000 to £300,000 a year before benefits. Most UK SMEs and councils can't justify that commitment for something they're still working out. A fractional engagement gives you the seniority and technical capability at a fraction of the cost — and if it stops delivering, you're out in 30 days with no penalty.
How a Fractional AI Director engagement works in practice
One to two days a week on a monthly retainer, with clear deliverables and a 30-day exit clause. The time allocation shifts depending on what's happening — heavier during implementation phases, lighter during steady-state governance. There is no six-month lock-in. Read a full breakdown of what a Fractional AI Director actually does, including governance, vendor selection, implementation oversight, and board reporting.
What you get
- AI strategy development aligned to your business objectives and commercial reality
- Hands-on technical assessment of AI tools, platforms, and build-vs-buy decisions
- AI implementation oversight with board-level delivery accountability
- AI governance framework development and compliance guidance covering UK GDPR, EU AI Act, and ISO 42001
- AI staff training and internal champion development so your team builds lasting capability
- Board-level judgment on AI risk, investment decisions, and implementation timing
Expertise
AI credentials and regulatory expertise
ISO 42001 AI Management System
ISO 42001 is the international standard for AI management systems. I apply ISO 42001-aligned AI governance frameworks to client engagements. This is the standard now showing up in regulated-sector procurement frameworks — and it's the one that tells a regulator or insurer you took governance seriously before something went wrong.
UK GDPR and AI Data Protection
Every AI deployment is UK GDPR compliant from day one. Data handling is designed for auditability and ICO defensibility. I've held data protection responsibility at board level and I understand how an ICO investigation actually unfolds — the document requests, the timelines, the questions that catch organisations out when the governance paperwork isn't in order.
EU AI Act Compliance for UK Organisations
UK organisations supplying to EU markets or processing EU citizen data are within scope of the EU AI Act. I ensure AI systems are correctly classified, documented, and compliant before enforcement action becomes a real risk. The general-purpose AI provisions took effect in August 2025 — this is no longer a future concern.
AI Governance for UK Local Councils and Public Sector
Every AI governance framework I've delivered to UK local councils includes Freedom of Information Act defensibility, Public Sector Equality Duty compliance, and democratic accountability documentation built in from the start. I hold current NPPV3 public sector security clearance, which is required for substantive operational access at council level.
AI Board Reporting
Boards and senior leadership teams get clear, non-technical AI progress briefings each quarter. The format is designed for governance decisions — not to demonstrate technical depth, but to give the people accountable for the organisation what they need to discharge that accountability confidently.
End-to-end AI implementation delivery
A strategy document sitting in a folder isn't worth what you paid for it. The engagement closes when you have working AI, trained staff, and a number you can put in the board report. Every engagement is structured so that is what I am accountable for — not just the recommendation.
Evidence & verification
How AI-Si.com evidences every claim
Every claim on this site falls into one of four categories. Each states only what can be independently verified: formal certifications, documented board-level experience, the published engagement framework, or material available privately on request during diligence. Nothing on this site is aspirational.
1. Certified
Formal credentials, originals available for inspection: current NPPV3 (Non-Police Personnel Vetting Level 3) police clearance for council and public sector work; Royal Navy service 1984 to 1990, vetted to Category 3 Positive Vetting at the time (note: the PV scheme has since been superseded by Developed Vetting). Where this site references ISO 42001, UK GDPR, EU AI Act, PSED, or FOI, it refers to working knowledge of those standards, not certification against them. Profile verifiable on LinkedIn.
2. Documented experience
Specific delivered client engagements, written up as anonymised case studies with full methodology. The £300K, £480K, £2.1M, 43%, 60%, and 18 hours per week figures each map to one of the four published case studies: council, manufacturer, professional-services firm, and NHS Trust. Governance controls, metrics, and timeline are disclosed. Named references and underlying calculations available on request.
3. Advisory framework
The engagement methodology and pricing applied on every AI consultancy engagement are published openly for diligence purposes. See the five-phase approach, three retainer tiers (£3,500–£8,500/month), and 30-day exit clause. Sector breadth across technology, manufacturing, professional services, and public sector is drawn from 35 years of board-level roles.
4. Available on request
Named client references, signed certificates and security clearance evidence, the underlying calculation packs behind every quoted saving, copies of the AI governance and GDPR documents delivered, and original source URLs. All are available privately as part of engagement diligence on request. Request the evidence pack
Common questions
Questions about working with AI-Si.com
The questions that come up most often before an initial conversation.
Find Out Where AI Can Save or Generate Money in Your Organisation
Bring a real problem. Leave with a concrete first step.