Hip of happening? | Everyday AI.

According to Jim, the initial sense of fascination has largely worn off. “The novelty has started to wear off. We have a reasonable idea of what AI can do, but it becomes far more interesting once people are allowed to use it as part of their day-to-day work.”

That shift is significant, because AI is no longer something people simply experiment with. They want to use it as part of their everyday work to:

• find information more quickly
• spend less time on manual tasks
• make better use of the knowledge that already exists within the organisation.

None of that happens automatically. Many employees still use AI much like a search engine: question in, answer out. That’s understandable, but it only scratches the surface of what’s possible. As Jim explains: The real value only emerges when AI has access to the right context, whether that’s documents in SharePoint, project information, customer data or internal knowledge.” At that point, AI moves beyond providing generic answers and starts delivering support that is genuinely relevant to the work people need to do.

That’s why AI literacy is becoming increasingly important. Not as a box-ticking exercise or a one-off training session, but as a fundamental requirement for using AI effectively across the organisation. Employees need to understand not only what AI can do, but also where its limitations lie.

As Jim puts it: “AI isn’t magic. At its core, it’s mathematical and built on historical data.” That means AI can produce impressive results, but it can also make mistakes, hallucinate or reinforce existing biases. Without that understanding, AI quickly becomes a risk. Employees may place too much trust in inaccurate outputs or share sensitive information in situations where they shouldn’t.

Others may expect AI to fix a process when the underlying data or way of working still isn't fit for purpose. Building AI literacy helps reduce those risks. Not by making people wary of AI, but by giving them the knowledge and confidence to use it responsibly.

As Kees points out, “Many people equate AI adoption with giving employees access to a tool. But purchasing a licence isn’t the same as adoption.” In the podcast, he explains that the real benefits only begin to emerge once the right foundations are in place, because understanding AI is only the first step. Employees also need to know how to apply it effectively in their own work and processes.

That requires much more than a brief introduction.

• Which processes are well suited to AI?
• Where can AI save time?
• Where can AI improve quality?
• Where do new risks emerge?
• How do you prevent employees from experimenting independently with disconnected AI tools?

AI may be one of the first technologies employees actively ask for. They see colleagues benefiting from new tools and want access to them as well. That’s a positive development, but it can also lead to fragmentation if left unmanaged. Without clear guidance, organisations can quickly find themselves dealing with shadow IT, or in this case, shadow AI.

Shadow IT rarely emerges because employees deliberately want to work irresponsibly. More often, it develops when people have a genuine need for certain tools and capabilities, but the organisation hasn’t yet provided a clear and practical alternative.

That’s why governance matters. But effective governance is about far more than maintaining a list of prohibited apps and tools.  As Jim puts it: “A good AI policy doesn’t just tell people what they can’t do. It also shows them what they should do.” Organisations still need to define which tools may be used, which data require protection and which use cases carry unacceptable levels of risk.

Good AI governance should also provide direction. Ask yourself:

• What are we trying to achieve with AI?
• Where do we want people to use it?
• How does it help employees do their jobs more effectively?
• How do we ensure AI supports the organisation’s objectives?

If all you do is tell people what they can’t do, they’ll find workarounds. Make it clear what they can do, and AI adoption becomes both safer and more effective.

Alongside adoption, data governance is one of the most important factors in successful AI implementation. After all, AI can only deliver meaningful value when it’s clear which data it can access and use. As Kees puts it: “Adoption matters, but data governance matters just as much—perhaps even more.”

For many organisations, that remains a significant challenge. They possess vast amounts of data, but often lack a clear overview of what is stored where, who has access to it and which information is sensitive. Without the right controls, introducing AI into that environment can create unnecessary risks. Confidential documents, payroll information, customer data and strategic plans all require careful handling.

Organisations therefore need to decide which data AI should be allowed to use, which information must never leave the organisation, and which safeguards need to be put in place. Employees also need to understand why those measures matter and how they support responsible AI use. Data quality and data classification are not issues to address at a later stage. They determine from the outset whether AI can be used safely, effectively and with confidence.

Many conversations about AI start with productivity gains. That’s understandable, as AI can help employees summarise, write, research and prepare more efficiently. But if you judge AI solely by the time it saves, you risk missing the bigger picture. In the episode, Kees poses an important question: should AI always be viewed through the lens of cost savings?

After all, time saved can be reinvested in better customer engagement, more informed decision-making, new propositions or work that previously never made it to the top of the priority list. The real business value lies not just in working more efficiently, but in what you choose to do with the time you gain.

AI becomes truly valuable when you’re clear from the outset about the problem you’re trying to solve. Do you want to:

• give employees more time?
• accelerate processes?
• make knowledge easier to access?
• improve quality?
• unlock opportunities that previously weren’t possible?

Start with the outcome you want to achieve, not the AI tool itself.

Many AI conversations start with the promise of saving time. That’s hardly surprising, given how effectively AI can help employees summarise, write, research and prepare. But focusing solely on minutes saved risks overlooking the bigger opportunity. Everyday AI requires experimentation. And as with any form of experimentation, not every initiative will deliver the expected results. As Jim explains: “That doesn’t matter. An experiment can’t really fail if you learn something from it.” There is an important caveat, however: experimentation without direction is something entirely different.

• Start small, with clearly defined use cases and clear ground rules.
• Focus on processes where AI can deliver measurable value.
• Make sure employees understand what they're doing and why as they experiment.
• Put strong data governance in place from the outset.
• Define clear ownership for governance, adoption, maintenance and evaluation.

This helps prevent AI from becoming a collection of disconnected initiatives and creates a clear path from experimentation to lasting value. Not because AI is new, but because it helps move your organisation forward.

Ask yourself one final question: Are your employees using AI primarily as a useful standalone tool, or do you already have the knowledge, governance and adoption strategy needed to turn it into lasting value?