How to build an AI strategy that actually works

Right now, there’s a lot of talk about AI — but not nearly as much about how to actually move from pilot projects to a functioning, company-wide AI strategy.

I (Angelika) often meet organizations that want to do more with AI but get stuck right at the start. The problem isn’t lack of ambition — it’s lack of structure. I came across a great report from MIT Technology Review Insights. It contains concrete insights that many companies would benefit from.

Here’s your four-step roadmap for AI, with my own reflections included:

Step 1: Build a sustainable data foundation and think beyond numbers

AI runs on data — but data isn’t just numbers. It’s also texts, emails, reports, meeting notes, and other information often scattered across systems or documents. Focusing only on technology is too narrow. We also need to ask: how do we manage information overall?

✔️ Organize everything from system data to loose documents

✔️ Digitize what’s still manual or verbal

✔️ Think beyond automation — think transformation

My reflection: This is where we need to look up and see the bigger picture. It’s not just about transcribing meetings faster, but about rethinking meeting culture. Which meetings can we skip entirely? How can we prepare better, follow up smarter, and create greater clarity? AI creates real value when it helps us work differently — not just faster.

Step 2: Choose the right models and partnerships

Most organizations don’t build their own AI models from scratch — and you don’t need to either. The smart move is to use what already exists and adapt it to your context.

✔️ Fine-tune prebuilt models and choose what works best for your tasks

✔️ Let AI become part of your existing tools

✔️ Focus on business problems, not technology first

My reflection: This is often about daring to say no. Just because something is possible doesn’t mean it’s relevant for your business. A good AI strategy focuses relentlessly on what actually makes a difference in daily operations.

Step 3: Expect investments — but see the full value

Building AI solutions costs money. But you can’t measure it as an efficiency project from day one — and that’s not the point.

✔️ Budget for both technology and cultural change

✔️ Measure value in terms of time freed up for strategy and development

✔️ Be patient — experimentation is needed to find what works

My reflection: AI isn’t about running faster right away — it’s about giving people the space to think further. By freeing up time from repetitive tasks, we can focus more on what truly matters: innovation, relationships, strategic decisions. That’s where AI’s real value shows.

Step 4: Think safe and ethical from the start

AI also brings new risks — legal, ethical, and technical. This should never be an afterthought.

✔️ Build in governance and control mechanisms

✔️ Stay updated on new regulations, like the AI Act

✔️ Always ensure human oversight

My reflection: Many feel resistance here because it seems complicated. But the truth is, if you take it seriously from the start, you’ll save both time and trust in the long run.

Conclusion: You don’t need to know everything — but you need to start somewhere

The biggest blocker I see? Many wait because they don’t know where to begin. And that’s okay.

Here’s my advice: Start where you are. Run a small test. Map your data. Involve a team. Choose one business challenge and see how AI can contribute — not just to doing things faster, but to doing them better and more sustainably.

That’s how it all starts. 

/Angelika