The truth about AI maturity (beyond the hype)

Many of the companies we work with think they're behind on AI. In reality, many are behind on something else entirely – and that's a very different challenge.

The anxiety behind the "we're too late" narrative is usually about speed:  who's ahead, what's been deployed, what opportunities might already be slipping away. AI has quickly become associated with urgency, experimentation, and visible progress. Boards want strategies, leadership teams want use cases, and employees are already testing tools on their own.

But when we look at what actually blocks results, it's rarely the lack of AI tools, ideas, or even technical capability.

Most organizations already have plenty of proofs-of-concept. Plenty of demos. Plenty of enthusiasm. In many cases, the issue is the opposite: experiments have multiplied faster than the organization’s ability to operationalize them.

The real problem is often the groundwork beneath the AI itself.

It's the data that was never structured for anything beyond reporting, or data that was never captured in the first place, but is now suddenly essential. It's operating models and value creation logic still built around how humans (both customers and employees) operate. As a result, AI gets applied to individual tasks without anyone rethinking what AI should do across the end-to-end value creation process or how human behaviour may need to change alongside it.

None of this is new, of course – these are the same dilemmas we've been dancing around for years in digitalization. But the conversation becomes much harder to avoid when companies want to build new AI-driven capabilities on top of non-existent data and unclear ideas of how end-to-end value creation should actually work with AI.

It's a hard conversation to have because many companies come looking for shiny AI, and end up needing unfashionable groundwork instead. 

FMCG, media, and retail are especially good examples of where this plays out because of the sheer volume of the data, systems, and customer interactions involved. The cracks in the groundwork become visible fast.

We asked three of our experts what they're currently seeing play out across these industries.

How far has AI come in the FMCG industry?

Karoliina Seppälä,
Associate Director, Knowit

AI in FMCG is moving from experimentation to real business impact in selected areas. It is already improving forecasting, pricing, and supply chain decisions where data is strong. Companies like Unilever use AI efficiently to accelerate innovation, for example, in product and recipe development. In contrast, in areas like creativity and brand building, AI is still searching for clear value and justification. AI is also beginning to reshape roles such as key account management, as decision-making becomes more data-driven and transparent across the industry.

What separates the AI leaders from the rest?

Leaders integrate AI into core processes with clear ownership and measurable outcomes. They combine strong data foundations with a commercial focus and scale use cases across functions. The difference is not access to technology, but the ability to turn AI into consistent business value.

How far has AI come in the media industry?

Paula Hernetkoski ,
Director, Knowit

AI in media has moved beyond experimentation into everyday use in content personalization, audience analytics, ad targeting, and production workflows. Many media companies already use AI to recommend content, optimize subscriptions, automate tagging, and support editorial work. Generative AI is also speeding up ideation, formatting, and localization. At the same time, its role in original storytelling, trust, and brand voice is still evolving, making maturity across the industry uneven.

What separates the AI leaders from the rest?

AI leaders in media treat AI as a strategic capability, not just a tool for efficiency. They combine strong first-party data, editorial judgment, and clear governance to create better audience experiences and faster workflows. They know where automation adds value and where human creativity and trust remain essential. While others run isolated pilots, leaders embed AI into products, processes, and decision-making in ways that drive measurable business impact.

How far has AI come in retail?

Elsa Nurmi,
Senior Advisor, Knowit

Further in adoption than in impact. Globally, AI has delivered most clearly in back-end operations: supply chains, demand forecasting, and inventory management. Customer-facing applications are progressing too, but remain relatively thin, especially in-store.

The two worlds rarely connect: retailers now operate more efficiently, but the customer experience often feels much the same. Many retailers are now looking into answer-engine optimization (AEO) and generative-engine optimization (GEO), yet few are exploring how consumer behavior itself may change in response to these technologies.

What separates AI leaders from the rest?

In many cases, it comes down to decisions made two or three years ago about data. The leaders began cleaning and structuring their data before they even knew exactly how AI would use it – deciding what to collect, where to store it, and how to organize it.

Unglamorous work, but it now allows them to move faster: training models on high-quality data and experimenting ideas without constant input roadblocks. The leaders aren't necessarily more skilled in terms of AI, but they are months, if not years, ahead of the companies that are sitting on scattered or missing data. 

Here's what we suggest

Collect data smartly. The winners with AI aren't the companies sitting on the biggest data lakes. They're the ones who decided early what data actually matters, in what form, and for what purpose. Smart collection beats heavy collection every time. And no, synthetic data won't save you, real data still does most of the heavy lifting, and the gaps in your data foundation will show up as gaps in your AI outcomes. The companies that started organizing their data two or three years ago aren't smarter; they're simply further ahead. If you haven't started yet, start now.

Think end-to-end, not desktop-by-desktop. The biggest gains from AI don't come from giving every employee a clever assistant or building flashy customer-facing gadgets. They come from rethinking entire processes, from production to the customer's everyday life. Isolated tools create isolated wins. Integrated AI changes the economics of how work gets done. Before you ask what AI can do for one role, ask what it can do for one process, end to end. That's where the real value sits, and how the leaders are winning. 

If any of this sounds familiar, there's a good chance we've already worked through it with someone in your industry. We work with companies from the first strategic conversation all the way through to delivery to help them derive value from AI. Give us a call if you'd like someone to spar with.