Why AI Has Stopped at the Individual

AI's biggest barrier is organisational change rather than technological capability.

Almost every AI article today sits in one of three camps:

• ‘10 AI prompts to make you more productive'
• ‘AI is going to replace project managers’
• ‘Here's our new AI feature’

Over the past two years, AI has transformed the way individuals work. We summarise meetings, write reports, analyse information, generate ideas and automate countless manual tasks far quicker, with greater accuracy than ever before through the use of our preferred AI tool. The technology has evolved at remarkable speed, and almost every week brings another announcement promising to revolutionise the workplace.

But there’s not a lot of information out there to cover what AI can do beyond us, as individuals.

At Verto, we’ve spent the last two years not only developing AI capabilities within our own platform, but listening carefully to how organisations respond to them. We’ve seen curiosity, excitement, hesitation and caution in almost equal measure. Which has prompted us to ask:

Why has AI become a personal productivity tool when its real potential is organisational?

A typical maturity curve takes users on a bumpy ride through initial interest, and eventual acceptance, but there is often a stage where the hype begins to deteriorate.

User adoption for any kind of transformative work is a challenge that can be made easier with the help of a team of change managers. But without that team in place, it can be much harder to drive the change forward to improve organisational outputs.

Here, we have explained this from our viewpoint.

Level 1 – Individuals

This level has high use and low risk. Here, people use AI in earnest, for things like:

• Summarising meetings
• Writing reports
• Generating presentations
• Research
• Writing code

We suspect this is in fact where the vast majority of adoption sits because there is a low permission threshold, no data is shared, the change is simple and benefits are clear.

Level 2 – Team AI

Now things become harder. It takes collective engagement and enthusiasm to collate a working model for AI when more people are involved. Teams start asking:

• Can we use AI together?
• Can AI read our project data?
• Can AI summarise delivery?
• Can AI analyse risks?

And suddenly there are questions asked and decisions to make:

Ownership and autonomy

“Who is leading the ‘project’ to initiate AI into our workflows? Do we all feel the same way about how it’s done? If so, what is the best way to implement this, and will the ultimate process be right for us all? We will have to check in and see how it’s going after a few weeks. What if there’s drop off? Who will manage that? Is it part of my job now?”

It’s now a thing and people need to be responsible and answerable for its success or failure. It feels like more of a challenge as, despite the data being shared is still minimal, it now warrants a discussion. The benefits are still obvious, but the question is, are they worth the change? That’s because the biggest difference here is the change, and if you don’t have that managed by someone, then it just won’t stick.

Processes and change

This is where change management enters the conversation. Some people see AI as an opportunity to modernise how work gets done while others ask a perfectly reasonable question: “Our current process works. Why should we shift?”

Neither perspective is wrong. But introducing AI almost always means changing established ways of working, and organisational change is rarely successful without leadership, communication and ongoing support. A change of process can knock the toughest and tightest of teams by throwing in uncertainty.

Policies and Security

“What are the risks?”

This level of process change will spark some debate over what is allowed and what isn’t. But whether it’s worth developing an internal policy is another reason that makes this level of AI the hardest to implement successfully. It’s everyone’s and no one’s job. There has to be a champion in the mix to get this right.

Then there’s security. Which AI tool is right for the work you’re doing and how is the data handled? This is a case of expertise. Does anyone in the team understand the different ways in which the AI handles the information it is provided by the team, or has the capabilities to find out and understand?

This one is a hard challenge to solve.

Trust and data quality

“Can I give this to the board?”

Finally for this level, you need to trust the data output. For the clients we serve at Verto, they need true analysis of their data, correct production of research and support, and accurate reporting of data.

Yes, the human team need to check and review the AI output, as we should all be doing on everything created and produced by AI tools. But without confidence in the underlying data, confidence in AI quickly disappears.

Level 3 – Organisational AI

This is where almost nobody has reached and where the greatest opportunity lies. At this level AI can be put to even better use.

AI starts helping organisations decide, which programmes should continue, where should investment go, which portfolio delivers the greatest value, which projects are likely to fail, where is capacity constrained and where are dependencies hiding.

The technology is increasingly capable of supporting these conversations and an entirely different proposition. Here, it is no longer about productivity. It's about decision-making.

But the stakes are high. Permission is required across multiple departments. Sensitive information must be managed appropriately and governance frameworks need to evolve. Change is significant and will affect multiple teams. It will be hard and might be some time before the benefits are seen.

The real adoption gap

Perhaps we’ve been asking the wrong question. Rather than asking why organisations aren’t using AI more, perhaps we should be asking why AI has stopped at the individual and how to tackle moving it up into the team level.

We’ll do more to think about, and try, to make sense of the organisational science of AI adoption over the coming months. We’ll look beyond the technology itself to examine governance, trust, change management, data quality, leadership and the role AI could play in transforming project, programme and portfolio delivery.

Our view is that the future of AI won’t be defined by how quickly individuals can complete tasks, but by how confidently organisations can use it to make better decisions.

We’ll also be taking this conversation to the AI Stage at DigiGov this September 2026, where our Chief Product Officer and Co-founder, Ross Mardell, will join a panel discussion exploring why AI adoption has stalled, what meaningful return on investment really looks like, and how organisations can move beyond isolated use cases towards AI that supports delivery at a strategic level.

If you’d like to join this conversation, head to our landing page dedicated to DigiGov 2026 and register to attend. Or contact us for a conversation about your thoughts on bringing AI benefits to your organisation at a systems level.