Most companies today use AI for operational tasks: translating texts, drafting emails, summarising documents. That’s a start – but it’s rather like driving in first gear on the motorway. The often overlooked potential: Modern AI systems such as Claude or ChatGPT can do far more than handle routine tasks. They can serve as strategic sparring partners – for decision-making, product development, negotiation preparation, and the critical reflection of business strategies.
This article demonstrates how we at NordAGI deploy AI in our daily work, what prerequisites are necessary, and where the limitations lie. This isn’t about efficiency gains in routine tasks, but about a fundamentally different approach to AI when addressing matters of strategic importance.
AI as a sparring partner – an often untapped potential
When the only tool you have is a hammer, every problem looks like a nail. The same applies to AI. If you’re content with AI translating texts or drafting emails, you’ve quickly defined its role – and drastically underestimated its potential.
The difference lies in the prompt. Instead of “Please translate this PDF into German”, the request could be:
“Support me in developing a go-to-market strategy for our new product. I’ve shared the technical description with you. Let’s proceed iteratively. Ask me all your questions – one at a time – so you have all the information you need to support me optimally.”
This is a fundamentally different approach. The AI isn’t being used as a digital assistant for simple tasks, but as a partner on equal footing. As a challenger who examines a problem from different perspectives and questions assumptions to uncover blind spots.
AI in practice
At NordAGI, we work with a capable AI model as our sparring partner. The model knows our company structure, understands who handles which tasks, and recognises where our strengths lie – and where we need support.
Product development is my domain. Marketing, however, I find more challenging because I’m not particularly keen on being visible. The AI model knows this because I’ve told it – and it takes this into account in its recommendations.
Wording and communication: Together with AI, we developed our wording through an iterative process, which now forms the foundation for all our corporate communications. Every blog post, every LinkedIn article is reflected upon with AI beforehand. This ensures we communicate consistently and our messages land clearly.
Product development and positioning: The AI model regularly supports me with strategic questions. Is our value proposition communicated clearly enough? What expectations do our customers have? We work not only on USPs but also on the sales process. AI helps us identify blind spots in our argumentation and prepare counter-arguments.
This means: We’ve critically worked through and challenged our product development before any real customer contact. Gaps, errors, or false assumptions become visible earlier. After the first customer interactions, their feedback flows into the process – and the next iteration begins.
Preparation for meetings and negotiations: Which arguments do we take for granted because we have an information advantage – which our conversation partner doesn’t have? AI helps us play through various scenarios. It’s not always about “winning”. Sometimes the greater value lies in both parties finding a good balance. AI also supports us with difficult decisions – such as whether we should accept a commission that carries certain risks.
Team usage: We use AI at a strategic level – as a team, both together and independently. Through iteration, a multifaceted picture emerges that leads to more well-founded decisions.
An interesting observation from practice: Many find it easier to have a business case challenged by AI than by colleagues. The threshold is lower, the openness greater.
AI as a sparring partner: The AI isn’t used as a digital assistant for simple tasks, but as a partner on equal footing – a challenger who examines problems from different perspectives and questions assumptions to uncover blind spots.
AI as a Sparring Partner: The Four Success Factors
From operational assistant to strategic thinking partner – how to unlock the full potential of AI systems.
Provide Context
The more relevant information AI has, the more precise its contributions. Without context, results remain generic and mediocre at best.
Work Iteratively
Rather than one monolithic prompt: step by step, the solution develops – making the thinking process visible and refinable.
Activate Challenger Mode
AI systems are inherently polite. Explicitly request critical questioning – not a yes-man, but a critical sparring partner.
Have Uncertainties Named
AI statements often sound supremely confident. Have uncertainties explicitly disclosed – this helps contextualise answers properly.
Prompting AI properly: “Create a PowerPoint for me” isn’t enough
For AI to function as a sparring partner, it needs more than a brief command. Three factors are decisive – plus a bonus:
1. Context: AI needs appropriate context to be truly helpful. Without this context, the result will be generic and mediocre at best. The more relevant information the AI has, the more precise and useful its contributions become.
2. Iteration: Rather than a monolithic prompt containing everything at once, an iterative approach is often more effective. Step by step, the solution develops – and the thinking process becomes visible.
3. Challenger mode: AI systems are inherently polite and restrained with criticism. That’s not particularly helpful. It’s better to explicitly tell the AI to question assumptions. Its role is that of a critical sparring partner, not a yes-man.
Bonus – Having uncertainties named: What AI produces often sounds supremely confident. It’s worth asking the AI to clearly identify uncertainties. Where is it confident? Which aspects raise doubts or open questions? This helps enormously in properly contextualising the answers.
With these adjustments, instead of superficially convincing-sounding statements, you receive differentiated, critical feedback. That’s the difference between “improve my text a bit” and systematic use at a strategic level.
Limitations of AI: Why decisions remain with humans
All this sounds almost too good to be true. And indeed, there are clear limitations.
AI is a coach for us – but not the decision-maker. No AI can assume responsibility for a final decision. And it shouldn’t.
Gut feeling: What AI cannot do is listen to gut feeling. Experienced decision-makers know this: everything looks good on paper, but something doesn’t feel right. AI cannot deliver this signal – nor replace it.
The echo chamber: Just like humans, AI can also become trapped in an echo chamber. With each iteration, the echo grows louder. A wrong approach at the beginning can steer the entire process in the wrong direction.
Choosing the right model: The more complex the question, the more important it is to work with a capable model and incorporate additional functions such as web search or document analysis.
A practical tip: We occasionally have the results from one AI challenged by another AI. Four eyes – even if they’re virtual – see more than two. This has given us interesting insights into how different AI systems evaluate and argue the same starting position.
What has changed in my way of working
Previously, I “wrote to think”. The writing process helped me gather arguments and contextualise perspectives. The result was long emails to the team, asking for feedback and validation of my thought processes.
With the AI model, this has fundamentally changed. I consistently use it to structure my thoughts and have decisions challenged. AI has become a highly available advisor for me. I know its limitations – and that’s precisely what makes the collaboration easier.
The chats are shared with the team. We’ve maintained the final validation of important decisions by the team. But the path to the decision – that has fundamentally transformed.
Using AI strategically: A closing thought
It’s worth trying to address strategic topics with AI as well. Current models are good enough to offer genuine support.
Anyone who only uses an LLM like Claude or ChatGPT to translate texts or summarise emails is driving in first gear on the motorway. The available potential remains largely untapped.
And the decision about what to do with AI’s insights? That always remains with humans. But a well-prepared human makes better decisions.
Curious about how AI could be effectively deployed in your company?
The examples in this article come from our daily work at NordAGI.
In our Leadership-First Programme, we pass on this experience: not as a PowerPoint training session, but as practical work on your real strategic questions. You bring your specific topics, and we show you how AI becomes a partner.
