Why So Many Organisations Struggle to Get Value from AI in Business And What the Successful Ones Do Differently

I spend a lot of time inside businesses and talking to senior leaders across retail, travel, financial services, B2B and the public sector. AI comes up in almost every conversation. Everyone is experimenting, everyone is curious, and everyone has a handful of pilots underway.

But when you dig beneath the surface, only a relatively small number of organisations are genuinely getting value from AI. Not demos. Not prototypes. Not internal showcases. Real value that shows up in commercial performance, customer experience, or operational efficiency.

Deriving value from AI

And over the past couple of years, you start to notice consistent patterns. The organisations making progress aren’t necessarily the ones with the biggest budgets or the most sophisticated tech. They’re the ones that have quietly figured out how to bring AI into the heart of the business without overcomplicating it or turning it into a science project.

Here’s what we see time and again.

The organisations that succeed treat AI as a business change, not a technical exercise

Whenever I meet leaders who are getting somewhere with AI, it’s striking how differently they frame the conversation. They’re not talking about models or embeddings or training pipelines. They’re talking about availability in stores, improving guest experience, reducing claim processing times, strengthening the sales pipeline, or solving very specific pressure points in operations.

In other words, the business problem comes first. The technology is simply the tool.

You see this most clearly in retail. The teams who are starting to win with AI aren’t obsessing about the model itself; they’re looking at how it helps stores run more efficiently, how it supports merchandising decisions, or how it strengthens loyalty propositions. That makes the work much easier to adopt internally because everyone can see the link between the experiment and the outcome.

The organisations that struggle tend to do the opposite: the conversation starts with the technology, and they hope the business will want it once it exists.

They bring business and technical people together early, and keep them together

One of the quiet breakthroughs we see is how successful organisations organise the work. They don’t have “AI happening over there” and “business happening over here”. They bring people into the same room early, and they work through problems together.

Trading managers sit with data engineers. Product owners and marketers sit with analysts. Operations teams help shape the direction from the start. This mix does two things. It brings the customer or colleague perspective directly into the design, and it forces the technical teams to stay connected to the real-world conditions they’re building for.

I’ve sat in enough steering groups to know that when these two worlds don’t mix, the project almost always slows down, gets stuck, or quietly fizzles out.

They spend far more time on trust than most people expect

This is probably the most consistent predictor of success. In organisations where AI works, people trust it enough to use it. That doesn’t mean blind trust or hype-driven enthusiasm. It means informed trust — the kind that comes from understanding what the system is doing, what it isn’t doing, and what role human judgement still plays.

That trust is built through conversation, transparency, and iteration. When teams understand why a forecasting model behaves a certain way, or why a recommendation engine adjusts outputs week to week, they’re far more willing to adopt it. You see this clearly in travel and hospitality, where trust in demand and pricing models is essential for revenue teams to act on them.

Without trust, it doesn’t matter how good the model is. It won’t get used.

They measure value simply and consistently, without turning it into an academic exercise

The organisations making progress have very down-to-earth ways of measuring value. They look at the metrics that matter to them already. If the model improves availability, they track availability. If it reduces the time it takes to produce a report, they track hours saved. If it lifts customer engagement, they track that.

They don’t over-engineer the ROI analysis. They use simple, honest measures, check progress regularly, and adapt quickly.

The organisations that struggle tend to swing to extremes. Either they try to measure everything in a way that becomes paralysing, or they don’t measure anything at all because they’re unsure where to start.

They treat AI as an ongoing portfolio of work, not a series of shiny experiments

The strongest organisations create a rhythm. They have quick wins that show immediate value, mid-term improvements that build momentum, and longer-term initiatives that reshape how parts of the business operate.

Crucially, they are willing to stop things that aren’t delivering. They move resources to where the value is. They don’t cling to ideas out of ego or sunk cost.

This mindset shift is what separates “innovation theatre” from genuine transformation.

So as a leader how do you extract value from ai in business?

If there’s one thing I’ve learned, it’s that AI success isn’t random. It’s the result of clear thinking, the right conversations, and a willingness to treat AI as something that belongs to the business, not the IT department.

The leaders who make progress are the ones who:

• talk about outcomes rather than technology

• bring mixed teams together early

• invest time in trust and adoption

• measure value in a grounded, practical way

• manage AI as an evolving capability, not a collection of experiments

None of these require huge budgets. They require clarity, persistence and a willingness to learn as you go.

If you’d like to explore where your organisation sits on this journey — or what it would take to move from experimentation to meaningful impact — we’re always happy to share what we’re seeing across the industry and help you shape a path that works for your context.