Stop falling for AI washing and start making investment decisions that actually improve your bottom line.

The Great AI Debate: Bubble or Revolution?

You’re sitting in yet another vendor demo, watching someone promise that their AI will “transform your entire operation.” Sound familiar? If you’re feeling overwhelmed by the endless stream of AI pitches whilst trying to figure out what’s actually worth your budget, you’re not alone.

The tech world is having a heated debate about whether AI is the next dot-com bubble or something more sustainable. But here’s the thing: whilst everyone argues about analogies, you’ve got real decisions to make with real money.

The AI Marketplace Feels Like a Chaotic Kitchen Exhibition

Picture this: you’re wandering through a massive kitchen equipment exhibition. Every vendor swears their gadget will revolutionise your bakery. The spiraliser promises to “transform customer experience.” The molecular gastronomy kit claims to “unlock unprecedented efficiency.” The smart oven guarantees “10x productivity gains.”

Meanwhile, you’re thinking: “I just need something that helps me bake better scones and serve customers faster.”

That’s exactly where we are with AI. The market is flooded with solutions looking for problems, with 79% of executives reporting exposure to generative AI, but many struggling to show concrete returns. Recent studies suggest that enterprise GenAI implementations often show zero or negative productivity returns, whilst vendors keep pushing “all-you-can-eat” pricing models they’re losing money on.

Your No-Nonsense Recipe for AI Investment Decisions

Forget the bubble debate. Here’s your practical framework for cutting through the noise and making AI investments that actually matter:

Step 1: Start with the Problem, Not the Tech

Before you evaluate any AI solution, write down your top three operational pain points. Be specific. “Improve efficiency” isn’t good enough. Try “Reduce average customer wait time by 2 minutes during peak hours” or “Cut inventory forecasting errors by 15%.”

Step 2: Apply the “Boring Tool Test”

Ask yourself: “Could I solve this problem with a spreadsheet, a better process, or existing software?” If yes, do that first. AI should solve problems that genuinely require pattern recognition or prediction at scale.

Step 3: Demand Transparent Pricing Models

If a vendor can’t explain their pricing structure clearly, run. OpenAI reportedly loses $2 for every $1 in revenue, which means those “affordable” rates won’t last. Ask for cost projections at different usage levels and get everything in writing.

Step 4: Pilot with Clear Success Metrics

Define exactly what success looks like before you start. Set a time limit (90 days maximum), track specific KPIs, and be brutal about calling failures early. Your success metrics should directly tie to business outcomes, not just “user engagement” or “time saved.”

Step 5: Calculate Total Cost of Ownership

Factor in training time, integration costs, ongoing maintenance, and the inevitable price increases. A £100/month AI tool that requires 20 hours of staff training isn’t really £100/month.

Step 6: Have an Exit Strategy

Before you sign anything, know how you’ll extract your data and processes if the vendor goes bust or prices become unaffordable. The AI landscape is moving fast, and not all players will survive.

The Real Question Isn’t Bubble or Breakthrough

Whether AI is a bubble or the next cloud computing doesn’t matter for your day-to-day decisions. What matters is building a sustainable approach to evaluating technology that serves your customers better and improves your bottom line.

The companies that thrive won’t be the ones who jumped on every AI trend. They’ll be the ones who asked the right questions, measured the right things, and invested in tools that solve real problems.

Stop chasing shiny objects and start building a tech strategy that actually works. Your future self (and your CFO) will thank you.