Krnic · For companies · one owner for your whole AI agenda

AI is on your agenda. Who owns it?

One senior person who owns the outcome, from the first decision to production.

Six questions. Nobody is watching.

1. AI came up at your last management meeting. What exactly was decided, and who is doing it this week?
2. You have been trying things in ChatGPT yourself, in the evenings. That is how most owners start. Has anyone turned what you learned into a plan for the company?
3. How many AI pilots did you start this year? How many does someone use every day?
4. You work with an agency, a software vendor, an IT partner and a consultant. Who holds the whole picture? Name the person.
5. The people who build your software are good. How many of them have put an AI system into production? Not a demo: real users, real data, still running a year later.
6. If your AI initiative fails, who is accountable? If the answer is nobody, that is the first thing to fix.

Two landed. You don’t need another tool or another agency. You need one senior person who owns the outcome.

Where these questions come from

1. Half of Belgian SME owners say they do not know how to put AI to use in their business. Exact KMO Barometer 2026, over 1,300 owners and managers of companies with 2 to 250 staff. Exact KMO Barometer 2026 (summary by 4BS, 22 April 2026)
2. In Flanders, almost half of the companies that use AI (48.9%) say the use grows bottom-up, from individual employees, rather than from a decision at the top. AI-barometer 2025, ECOOM for the Flemish government, March 2026. Departement EWI, Vlaamse overheid, 11 March 2026
3. By the end of 2025, over half of generative-AI projects had been abandoned after the proof of concept, according to Gartner (2026). One in four AI initiatives delivered the expected return, according to 2,000 CEOs surveyed by IBM (2025). Gartner, Why 50% of GenAI Projects Fail (2026); IBM CEO Study, 6 May 2025 · IBM CEO Study, 6 May 2025
4. Fourteen percent of Flemish SMEs use AI in a structured way; in six out of ten, employees largely decide for themselves how they use it. Teamleader and iVOX, 297 respondents, June 2026. HR Magazine on the Teamleader/iVOX survey, 4 June 2026
5. About 5% of custom enterprise AI tools reach production; 95% of organisations see no measurable effect on the P&L. MIT NANDA, The GenAI Divide: State of AI in Business 2025. Fortune on the MIT NANDA report, 18 August 2025
6. The EU AI Act applies to companies that only use AI: since February 2025 you must see to your staff’s AI literacy, and since August 2026 customers must be told when they are talking to a machine. The high-risk rules were postponed to December 2027 by the Digital Omnibus. The law names no one in your company to own any of this. In practice, that lands on the general manager. European Commission, AI Act and its application timeline

In one sentence

Krnic takes over your company’s relationship with AI, end to end. The strategy and the decisions, what to build and what to buy, the prototypes and the production systems, the agencies and vendors, the AI Act paperwork, and the running of it afterwards. One senior person accountable for all of it, not a committee.

What owning it means

Decide. Week one. I look at your company, not your wish list, and set the AI agenda: what is worth building, what to buy, what to stop. Including “nothing yet”.

Show. Within days. You see a working prototype, not a slide deck. Built on your data, in front of the people who will use it.

Carry. Until production. I stay accountable until it works in production, and the agencies and vendors on your AI work report to me on it. Every phase has an outcome, a capacity in days and a fixed monthly fee. If I underestimate a phase, that is my problem, not your invoice.

Anyone who has built a house knows the difference between the architect and the contractor. The contractor builds what the plans say. The architect answers for the plans, and for the works, until the keys are handed over.

How it starts

The first week: Fixed price, delivered in week one. An analysis of how your company actually works, one decision on what to prototype first, and the first working version of it on your screen. If the honest decision is “nothing yet”, you get that in writing, and we stop there.

Three examples of what a phase looks like

• One day a month. Keep the picture, take the decisions, unblock the people doing the work.
• One day a week. A phase in motion: a prototype becoming a system, with me on the build.
• Three days a week. A quarter to production. The seat, the build and the accountability, most of the week.

You will know which one you are before you commit, and you will always know how many days you get, what they deliver and what they cost that month. The proposal you receive after we talk states all three, phase by phase.

• You talk to me. I do the work. No account manager, no handover to juniors.
• When I recommend building something, you can build it with me or with anyone else. I oversee it either way.

Why me

• 25 years: of building software, twenty of them at the pace of the American market: decide in the morning, ship in the evening.
• AI in production: Pipelines that extract with Gemini and verify with Claude, a generator that writes whole websites from registry data, AI-rendered characters in a children’s app. Model-agnostic, chosen per job. Running today, on this site.
• Built for: Bird Buddy, Journey, Loop Africa, Mealmate, CityX. A smart bird feeder with AI recognition, a hotel loyalty platform, a wellness membership, a food platform, a smart-city platform. All shipped, all on this site.

Send me your situation. You don’t need a brief.

In twenty minutes you will know whether I can help, and so will I. Write to [email protected] or call +32 490 36 49 17.