Why AI Without Strategy is Just Expensive Guesswork

Every resources company is being told they need AI. The pitch decks arrive weekly. The consultants follow. The board asks "what's our AI strategy?" and suddenly a team is assembled, a budget is allocated, and a pilot is launched.

But here's the uncomfortable truth: most AI initiatives in the resources sector fail not because the technology doesn't work, but because nobody asked the right question first.

The Question That Matters

Before you evaluate vendors, before you hire data scientists, before you allocate a single dollar, you need to answer one question:

What decision are we trying to improve, and what would improving it be worth?

This sounds simple. It isn't. Most organisations jump straight to "what can AI do?" when they should be asking "what decisions cost us the most when we get them wrong?"

The Signal-to-Noise Framework

At Rekon, we use a three-part framework to separate genuine AI opportunity from expensive noise:

1. Decision Audit

Map the 20 decisions that drive 80% of your value. For a mining company, this might be fleet deployment timing, maintenance scheduling, or grade control. For an energy company, it might be demand forecasting or network investment sequencing.

2. Data Reality Check

For each high-value decision, assess: do you actually have the data required to build a model? Not "could we theoretically collect it", do you have it now, is it clean, and is there enough history to train on?

3. Human-in-the-Loop Design

The best AI systems don't replace human judgment, they augment it. Design the workflow so that AI provides the analysis and humans make the call. This isn't a compromise; it's how you avoid the hallucination problem that makes pure AI solutions dangerous in high-stakes environments.

What We've Learned Since 2020

We've been building computational models with Wolfram since 2020, well before the ChatGPT wave made AI a boardroom buzzword. Here's what that experience has taught us:

1. Start with the decision, not the data. Data-first projects produce dashboards nobody uses.
2. Prototype cheap, validate expensive. A $50K proof of concept that fails is a success. A $2M platform that nobody trusts is a disaster.
3. AI literacy matters more than AI capability. If your leadership team can't interpret model outputs, the model is worthless.

The Bottom Line

AI is not a strategy. It's a tool. And like any tool, its value depends entirely on the skill of the person wielding it and the clarity of the problem they're solving.

If you're being told you need an AI strategy, push back. What you need is a decision strategy, and AI might be part of the answer.

---

Peter Winnall is the founder of Rekon Group and has been building AI-augmented decision systems for resources and energy companies since 2020.