"We urgently need AI" - and then the project fails

The Scene That Keeps Repeating

The CEO comes out of a conference. Or has read an article. Or heard that the competitor is "doing something with AI."

"We urgently need AI."

Budget is released. A team is assembled. Nine months later: A model lies around that predicts something. But no one uses it.

Why? Because the prediction is of no value to anyone. Or because the quality is inadequate. Or because there was no plan for how people are actually supposed to use the thing.

(Does that sound exaggerated? It's not. We see this every few months.)

The Uncomfortable Statistic

About 70% of company-wide AI initiatives do not deliver the expected business value.

Not because the models are bad. Not because the data scientists are incapable. But because companies start with the wrong question.

"How do we use AI?"

That's the wrong question.

The right question: "What specific business problem can we solve with it?"

That sounds trivial. But it's not. The first question leads to a bottom-up approach: You gather tech experts and they search for problems for their solution. That's backwards.

The second question leads to a top-down approach: You ask where the pain points are, where costs arise, where customers are frustrated. And then you check if AI can help.

The Five Reasons Why AI Projects Fail

1. Hype instead of Strategy

The pressure to "not be left behind" leads to investments without a clear goal. "We must implement generative AI" is not a strategy. That's panic.

2. Data is not what companies think

"We have so much data!" Yes. But having data in the system is not the same as having clean, structured data usable for machine learning.

Most companies massively underestimate the effort required for data preparation. It's not sexy. But it's crucial.

3. No Business Ownership

Many AI projects are driven by IT. That's a classic mistake. If the business owner is not on board - if the person who should use the outcome is not actively involved - then you are building a solution that no one wants.

4. Change Management as an Afterthought

Building a model is the easy part. Getting people to use it and adjust their processes - that's the hard part.

We've seen highly accurate prediction models that no one uses. Because users weren't trained. Because integration into their workflows is painful. Because no one explained why it should be valuable.

5. ROI is not defined

In the end, someone asks: "Was it successful?" And no one knows. Because it wasn't defined beforehand what "successful" means.

Was it time savings? Revenue increase? Error reduction? If that's not clear, you can't measure anything. And the next manager says: "The project was expensive and brought nothing."

What AI Strategy Really Means

AI strategy is not "We implement AI." That's like "Our strategy is to use software." That doesn't say anything.

AI strategy means:

A clear goal. Not "we want to use AI," but "we reduce the average response time in customer service from 24 hours to 2 hours through automatic ticket classification."

Realistic expectations. ChatGPT is impressive. But it hallucinates. It sometimes doesn't understand nuanced business logic. It may not work on your specific data.

Honest readiness assessment. Do you have the data? Do you have the skills? Do you have the infrastructure? These aren't theoretical questions - these are practical blockers.

Clear metrics. Before you start: What does success mean? How do you measure it? When do you know if it's worked?

Generative AI: The Reality Behind the Hype

Generative AI (ChatGPT, Claude, etc.) is not a miracle cure. It is a powerful tool for specific use cases.

Where it works well:

- Summarizing long documents

- Content generation (with human review)

- Q&A based on knowledge databases

- Code generation (the code must be reviewed)

- Classification and extraction

Where it works poorly:

- Financial or critical decisions where accuracy is crucial

- Highly specialized domains where the model could hallucinate

- Privacy-critical use cases when data protection is not ensured

What Most Underestimate:

Hallucinations are real. Generative models can present invented facts as if they were true. That's not a minor issue.

Data privacy is complex. If you use ChatGPT directly (not the enterprise API), your data can potentially be used for training. That's a GDPR issue.

Cost scales. "Generative AI is free" is a myth. With a million API calls per month, it quickly becomes five figures.

The Pragmatic Approach

If you want to seriously tackle AI:

1. Start with the problem, not the technology

What are the three biggest pains in your company? Where do most costs arise? Where do customers lose time? Where are error rates high?

Only when you know that, do you ask: Can AI help here?

2. Honestly assess your readiness

Do you have the data? Is it clean? Do you have someone who can drive this? Is the business behind it?

If the answer to several of these questions is "No" - then start there, not with the AI project.

3. Start small

Not "we transform everything with AI." But: A pilot. A problem. Clear metrics. If that works, scale. If not, learn.

4. Define success upfront

What does it mean if the project is successful? How do you measure that? Write it down before you start.

5. Take governance seriously

The EU AI Act is not optional. If your model runs in sensitive areas - recruiting, credits, medicine - then there are regulatory requirements. Ignoring them is expensive.

What You Can Do

1. List the three biggest business problems - not IT problems, business problems. Where are you losing money? Where are customers frustrated? Where are errors occurring?

2. Ask yourself for each: Could AI help here? Be honest - sometimes a simple rule-based system is better than machine learning.

3. Assess your data situation - If you wanted to train a model tomorrow, could you provide the data to the data scientist? Is it clean? Structured?

4. Talk to the business - Not just IT. Would someone from the specialty area use the result? Is there genuine interest?

The Point

AI is not a magical thing. It is a tool. A very powerful, but a tool.

AI creates value when you solve the right problems. When you are realistic about capabilities and limits. When your data is in order. When the business is behind it.

Companies that are successful with AI do not have the best technology. They have the clearest goals. They know what they want to achieve. They measure if it works.

That's not sexy. That's not hype. But it works.

Do you want to approach AI strategically - not as a hype project? We help Swiss companies find the right problems and achieve real ROI. Talk to us.