A recent article in The State of Brand stopped me because it put hard data behind something we have believed at Slick+ for a long time: technology on its own does not transform organisations. People do. The article’s central point was clear: cutting people for AI does not appear to correlate with better financial performance.

The article was reporting on Gartner research showing that approximately 80% of organisations piloting or deploying autonomous business capabilities – including AI agents, intelligent automation and digital twins – had also reduced their workforces. But those reductions did not translate neatly into improved returns. Gartner’s Helen Poitevin put it plainly: “Workforce reductions may create budget room, but they do not create return.” 

That distinction matters.

Right now, too many AI business cases are being built backwards. They start with a cost line. They identify people. They estimate how much work AI can absorb. Then they present the savings as the return.

But savings are not the same as value. Fewer people do not automatically mean a smarter organisation. If the people leaving are the ones who understand how work really gets done, the organisation may be cutting away the very knowledge it needs to make AI useful.

At Slick+, we are not anti-AI. Quite the opposite. We believe AI will become one of the most powerful tools organisations have ever had to support learning, productivity, decision-making and role-based performance. Gartner forecasts AI agent software spending will rise sharply from $86.4 billion in 2025 to $206.5 billion in 2026 and $376.3 billion in 2027.

This is not a passing trend. But neither is it a magic trick.

Gartner describes the opportunity not as “humanless business” but as “human-amplified business”. That phrase matters because it changes the question. The best organisations will not simply ask, “Which jobs can we remove?” They will ask, “How do we help people guide, govern, improve and scale what AI can do?” 

That is a very different mindset.

It is also where tacit knowledge capture becomes strategically important. Every organisation has formal knowledge: policies, SOPs, manuals, training modules and process documents. Then there is informal, tacit knowledge: the judgement, shortcuts, stories, patterns, exceptions and “how we do things here” knowledge that lives inside people’s heads.

Michael Polanyi famously wrote, “We can know more than we can tell.” That is the heart of tacit knowledge. People often know how to solve a recurring issue, handle a difficult customer, onboard a nervous new colleague, avoid a safety mistake or make a judgement call long before that knowledge appears in a formal system. 

This is the space Slick+ has always cared about.

Long before AI agents became a boardroom topic, we were focused on helping organisations capture the lived experience of their people through short, practical, colleague-made videos. Not because video is fashionable. Not because another platform is the answer. But because so much of what matters at work is contextual, visual, practical and human.

Our own writing has often used the “Maureen from Accounts” example: the person everyone goes to because she knows how to fix the awkward system problem nobody else can explain. She may never have been appointed as a trainer. She may never have written a manual. But when she is on leave, everybody notices. That is organisational knowledge. And if it is not captured, it is fragile.

AI makes this more urgent, not less.

An AI agent is only as useful as the knowledge, context and judgement it can access.

If an organisation’s knowledge is incomplete, out of date or trapped in people’s heads, AI will automate the visible layer while missing the reality underneath. It may process the policy but miss the exception. It may summarise the manual but not understand what happens on the night shift, in the depot, on the construction site, in the care home or with a new colleague who needs reassurance before they can perform with confidence.

This is why people-centred AI has to start with knowledge retention and frontline knowledge transfer.

Before asking AI to manage work, organisations need to understand the work. 

Before deploying agents, they need to map the judgement calls, handovers, risks, local practices and hidden expertise that keep the business running. Before replacing people, they need to ask what knowledge those people hold – and whether losing it will make the organisation thinner, slower and less capable.

This is not a soft argument. It is a performance argument.

The World Economic Forum’s Future of Jobs Report 2025 found that 39% of workers’ core skills are expected to change by 2030, while 63% of employers identify skills gaps as a major barrier to business transformation. McKinsey makes a similar point about AI in the workplace: this is “not a technology challenge” but a business challenge requiring leaders to align teams and rewire companies for change. 

And transformation is hard. Bain & Company found that only about 12% of business transformations achieve their original ambition, with success strongly linked to how well organisations retain, develop and acquire the right talent and capabilities.

So the lesson is not: slow down AI.