An AI mentor for businesses that actually get things done.

A fractional AI CTO is a part-time Chief Technology Officer who owns the AI side of your business without joining the payroll. The role covers three things: deciding which AI work is worth doing, running the implementation, and training the team that inherits it. It suits companies too small to justify a full-time CTO salary but too committed to AI to leave the decisions to a vendor. In practice that means four to eight working sessions a month, plus async support between them. Ivan Nordgaimer works this way with UK and EU small businesses, drawing on more than fifteen years of delivering business systems and a verified record of 56 completed projects and 2,615 billed hours at a 100% job success score. The difference from a consultant is accountability: the same person who writes the roadmap is the one who ships it.

This is for owner-led small and medium businesses in the UK and EU, typically between 10 and 200 employees, that are already trading and now spending money on AI without a clear view of the return. The common starting point is a company paying for three or four AI subscriptions where nobody can say which one earns its keep. It also fits businesses with obvious manual bottlenecks — quotes assembled by hand, enquiries routed by memory, reports rebuilt every month in a spreadsheet. You do not need an in-house technical team; that is the point of the arrangement. It is a poor fit for pre-revenue startups looking for a technical co-founder, and for enterprises with an existing AI function who need extra hands rather than direction.

Tool choice follows the problem, never the other way round. The recurring set is: Claude and ChatGPT for reasoning and drafting workflows, n8n, Make.com and Zapier for process automation, OpenAI Whisper for speech-to-text, and retrieval-augmented generation over your own documents when staff keep asking the same questions. Custom dashboards and CRM integrations get built when an off-the-shelf tool would cost more to bend than to replace. Ivan has shipped fourteen systems on this stack, including Zvukio, which runs nine independent analyses on every business call across more than ninety languages, and Polychka, which generates styled product imagery at roughly four cents a shot against photoshoots costing £150 to £500. Selection is driven by projected return, and tools that cannot show one are not recommended.

Quick wins land within two to four weeks, and a full transformation takes eight to twelve weeks. The first fortnight is deliberately unglamorous: an audit of what is actually happening in the business, followed by one or two automations narrow enough to finish. Those early builds exist to prove the approach on your own data before anyone commits to a larger programme. The eight to twelve week window covers the wider work — five to ten automated workflows, KPI dashboards, and the training that lets your team run them without you. Most clients see a return inside the first month of implementation, because the first thing built is chosen for payback speed rather than ambition. If there is no measurable progress in the first two weeks, the engagement is refunded in full.

No technical knowledge is required from you or your team. The technical strategy, the build and the integration work all sit on Ivan's side of the line. What your team receives is training on the finished tools, which are chosen to be operable by the people who already do the job — no code, no prompt engineering, no new specialism to hire for. The measure of a good handover here is that the business can keep running the system after the engagement ends. What is needed from you is different and harder to outsource: an honest account of how the work really gets done, including the workarounds nobody documents. Those workarounds are usually where the automation opportunity is hiding, and no amount of technical skill substitutes for someone who knows them.

An NDA is signed before any commercial detail is discussed, on day one rather than at contract stage. Implementations are built to align with UK GDPR, which shapes concrete decisions: what data leaves your systems, which model providers process it, how long anything is retained, and whether a provider may train on it. Where the answer to that last question matters, the architecture uses providers and settings that contractually exclude training on your data. Sensitive records are kept out of prompts entirely wherever the result can be achieved without them. Your data stays yours throughout, and the systems are documented so an auditor can follow what happens to it. This is treated as a design constraint from the first session, not a compliance review bolted on before launch.

A consultant is paid to deliver a recommendation; a fractional AI CTO is paid for the outcome. The practical difference shows up after the strategy document exists. Consulting engagements typically end at the roadmap, leaving the business to find someone else to build it and to discover the gaps between plan and reality on its own budget. Here the same person writes the roadmap, builds the systems, trains the team, measures what happened, and revises what did not work. Ivan has shipped fourteen AI systems that are live in real businesses, and his delivery record is externally verifiable: 56 completed contracts, 2,615 billed hours, and a 4.97 average rating across 32 client reviews. The other difference is candour — part of the job is saying which parts of AI would not pay for themselves in your business.

There are two commercial models: fixed-scope projects and monthly retainers. Projects suit defined pieces of work such as an AI audit, which runs two to three weeks, or an implementation programme of eight to twelve weeks. Retainers suit ongoing leadership and come in two tiers — an advisory tier of four strategy sessions a month, and a hands-on tier of eight or more sessions that includes implementation oversight, vendor management and attendance at key meetings. Pricing depends on company size and project complexity, so it is quoted after the discovery call rather than from a rate card. Invoices are settled in GBP or EUR by bank transfer. The free thirty-minute discovery call carries no obligation and is not a sales demo; its purpose is to establish whether there is a worthwhile project at all.

Yes. If there is no measurable progress within the first two weeks of an engagement, the fee is refunded in full, provided the agreed steps were followed on both sides. That last condition is not fine print: the work depends on access to the people and systems being changed, and an engagement stalled waiting for either cannot be judged on results. The guarantee exists because the first two weeks are deliberately structured to produce something checkable — an audit finding, a working automation, a measured baseline — rather than a plan whose value only becomes clear months later. It also reflects how the engagement starts. The initial thirty-minute call is free and frequently ends with the conclusion that a business does not need the work yet, which is a legitimate outcome rather than a failed sale.