What to ask before you approve an AI project
You are shown a working prototype and asked to approve a budget. This session gives you the small set of questions that separate a project that will reach production from one that will not, and the cost model to check the number you are being given.
- €2,000 per workshop
- Half a day
- Up to 12 people
- English or German
Nobody expects you to evaluate the model. You carry the risk either way
Nobody expects a decision maker to evaluate an architecture. But you are the one who signs for it, explains it to a board, and answers for it if it goes wrong, so the relevant skill is not technical depth. It is knowing which questions have answers and which do not.
Most AI proposals that reach an executive have the same three gaps: no named user, no cost at real volume, and no measure of success agreed before the build. Each of those is a question you can ask in one sentence. Each of them, unanswered, predicts a project that quietly dies.
The session also covers what you are legally on the hook for. The EU AI Act assigns obligations by role and by risk class[4], and ISO/IEC 42001 exists precisely because organisations needed a management system to demonstrate they are handling it[1]. You do not need to read either. You do need to know which of your systems fall inside them.
The column you own, and the two you have to fund
The model is the same in every track. What changes is which column you own and what you are expected to hand over.
The business
“What is it worth, what is in the way, and who signs?”
- The value chain and where the money actually is
- Constraints that are real, not imagined
- The number, checked rather than accepted
- The measure of success, agreed up front
In the room: You, finance, legal, the sponsor
The user
“Who is this for, and what does their day look like afterwards?”
- A named user, not a department
- The journey as it should be
- What breaks today
- Where AI is allowed to act
In the room: Product owners, business analysts, designers
The technology
“Can it be built, run and defended?”
- Architecture and the layer decision
- Cost at real volume
- Security and certification evidence
- Who is on call, and what that costs
In the room: Engineering, platform, security, operations
What you take into the next steering meeting
Everyone leaves with the same two documents. The optional ones are for the projects already on your desk.
The question setIncluded
Roughly a dozen questions, in order, with what a good answer sounds like and what an evasion sounds like. One page, usable in the next steering meeting.
Cost and value modelIncluded
What AI costs at your volume, the three places the number is usually wrong, and how to sanity-check a proposal without an engineer in the room[14].
Portfolio readOptional
Your current AI initiatives sorted into: real, needs work, and demo. With the reason for each, written down.
The questions, and what a real answer sounds like
A sample. The full set comes with the session, but these four alone will change most of the conversations you are having.
| The question | Why it works | What an evasion sounds like |
|---|---|---|
| Who is the named user, and what does their day look like afterwards? | Forces a person rather than a department. Most failed projects cannot answer it. | "It will help the whole organisation be more efficient." |
| What does one transaction cost at full volume? | Unit economics are where AI projects break. Demo volume hides it entirely. | "The pilot costs are minimal." |
| What measure did we agree before we started, and who agreed it? | Success defined afterwards is always achieved. | "The feedback has been very positive." |
| What happens the first time it is confidently wrong? | Separates people who have run one from people who have demoed one. | "The accuracy is very high." |
| Who is on call for it, and does that person know? | Operating cost is the part nobody presents. | "It's fully managed." |
Half a day, in three parts
No slides for the sake of slides. Two of the three parts use your own initiatives.
The questions
The question set, with real examples of projects that passed and failed each one. This is the part people quote afterwards.
Your portfolio
Your own current AI work, sorted. Uncomfortable, useful, and usually the reason the session was booked.
One workshop, the results you choose
Same model as every other workshop: €2,000 for the session, €1,000 per written result. Net, plus VAT.
AI for decision makers workshop
Pre-session intake, the live workshop with up to 12 people, written results within 5 business days.
€2,000 each, 3 to 4 hours, remote or on-site
Delivered within 5 business days after the last session.
AI for decision makers
What people ask before booking.
Do I need any technical background?
No. The session is built so that the questions work without it. If you can read a budget you can use this.
Is this just a risk session?
No. Half of it is about finding the projects worth doing. The question set filters in both directions.
Can I bring my leadership team?
Yes, and it works better. Up to 12. The uncomfortable conversations are more useful when everyone hears the same version.
Will you tell me if a project should be stopped?
Yes. That is usually the most valuable thing that happens in these sessions, and it is why I price by the workshop rather than by the outcome.
Is it AWS-specific?
No. The cost model uses real AWS numbers because those are the ones I can source, but the questions are vendor-neutral.
Linda Mohamed
I design and build AI and cloud systems on AWS and hybrid platforms, and I teach how they work. I have organised the AWS User Group Vienna for more than seven years, co-organise AWS Community Day DACH, teach AI architectures at Hochschule Burgenland, speak at conferences in Europe and the US, and maintain ai-solutions.wiki. Based in Vienna. Remote or on-site, in English or German.
Where the facts come from.
Prices, service names and limits change. The linked official pages are the reference; figures on this page were checked in September 2026.
- ISO/IEC 42001:2023, AI management systems
- ISO/IEC 27001, information security management
- ISO 9001, quality management
- Regulation (EU) 2024/1689, the EU AI Act
- NIST AI Risk Management Framework
- CNCF TAG App Delivery: Platforms White Paper
- CNCF TAG App Delivery: Platform Engineering Maturity Model
- DORA: DevOps Research and Assessment
- AWS Well-Architected Framework
- Amazon Bedrock AgentCore: observability
- AWS: AI services overview
- ai-solutions.wiki: from AI proof of concept to production
- ai-solutions.wiki: build vs buy for AI
- ai-solutions.wiki: total cost of ownership for AI
- Training Magazine: 2025 Training Industry Report
Start with one conversation.
30 minutes to look at your idea, your data and the right starting point. Or book 15 minutes if you only have a question.
- Which of your AI initiatives are real
- What the first question should be
- Whether your governance position is a problem yet
- Who should be in the room
Vienna, Austria · remote across Europe · [email protected]
