Beyond Cool Tech: How to Unlock Real Business Impact with AI

The obsession with technology, first thinking is holding AI back from delivering real ROI. Businesses are falling into the trap of building AI solutions in search of problems rather than solving real customer and business needs. In this post, we explore how the Product Triad; Desirability, Viability, and Feasibility, can help refocus AI efforts for tangible success. We’ll also dive into the critical role of people and process change in AI adoption and why old-school best practices matter more than ever in the GenAI era.

Stuart Winter-Tear

3/14/20253 min read

man in black suit jacket and black pants figurine
man in black suit jacket and black pants figurine

"If we focus on technology, we end up with more technology, if we focus on the business and customers, we end up with more business and customers." Stuart Winter-Tear

This was the self-quote I used to kickoff and validate my own talk entitled: "From Cool Tech to Real impact". There is a lamentation across industries that GenAI is still not giving the ROI they hoped for, and much of that can be blamed on throwing out the ‘old wisdom’ and best practices, as if it doesn’t apply to AI which is somehow a ‘special category ’.

It isn’t.

From security wisdom, to established product practices, commercial approaches, and pretty much anything else you can think of, the focus is almost entirely on technology, as we all forget the adage “build it and they will come” never did work. This is the “cool tech syndrome”, and I would like to take us back to basics and cover two points from the talk, that could save us a lot of pain and wasted resources.

The Problem is not the Problem, That’s the Problem

I was boring my wife in the car with one of my product monologues, and she asked why there was so much AI technology in search of problems to solve. I said “the problem is not the problem, that’s the problem”. Now, she has been married to me for 25 years and understood what I meant, but for you dear reader, I shall explain.

We’re looking at all this cool tech and research, and thinking up all the cutting-edge things we could build, rather than spending time in the customer problem-space, deeply understanding the problem, and then weighing up if AI is the right solution. We’re putting the cart before the horse, in other words.

In order to counteract this strong temptation it’s good to use a simple, well established framework, to help us ask the right questions.

The Product Triad is a perfect example:

Desirability

This focuses on the customer and the problem. Do we understand the problem, the context of the problem, the problem under the problem? Is the problem worth solving, and will it create value to the customer and market, if we do so?


Viability

This focuses on the business. Does building this accord with the larger business goals and objectives? Can we create a moat, and will there be measurable ROI?


Feasibility

Finally, this focuses on technology. Can we build it, scale it, support it? Is AI the right technology to throw at the problem?


This is not a technology problem, it's a change problem.

The myopic focus on technology takes our eye off the fact this is in fact a change problem, and more specifically, a people and processes change problem.

On the people-side, we’re asking employees to adopt a technology they’re being told could displace them. The flipside of this is employees using GenAI to augment their work, but not telling their employers - again - for fear of being displaced, causing an avalanche of “shadow AI”. Much of this is the fault of hype and can hinder the people-change problem. There is a huge educational part needed here, and a lot more grounded, sensible, business advice, and social media posts.

On the process-side, the great hope for Agentic systems is capturing the power of LLMs inside knowledge workflow process automation. The difficulty is most businesses don’t know their processes intimately, and if you don’t know your processes, how can you expect AI to?

So, this necessitates the old-fashioned, boring, grunt-work of process mining and mapping. But who wants to do that when we can chase cool tech instead? Then we wonder why we’re not getting ROI.

The irony is, those who do this grunt work will be best placed to take advantage of the technology as it matures, and outpace their competitors.

There’s much more I could add from the talk, and if there’s interest, I’m happy to do so, but for now I hope you can get the flavour, which is…

We must not chuck out the ‘old wisdom’ and best practices for AI, as they apply more than ever!