Everyone wants to know the secret to scaling AI adoption.

The truth is that nobody has fully solved it yet, including many of the technology companies building the models themselves.

As I’ve been preparing for Media Tech & AI Week, I have been speaking with people working at the sharp end of AI development. When it comes to rolling out AI across a business, it’s clear adoption is far less about technology and far more about helping people redesign how work gets done.

Teams need help incorporating AI into workflow

Some employees have embraced AI completely

They’re using it daily, experimenting constantly, and redesigning how they work. Or as tech companies say, they’re using AI to “10x” their work. Others are still largely using it as a better spellchecker. The result is a bimodal distribution of adoption, with relatively few people sitting in the middle.

What’s striking is that access to tools is rarely the biggest problem.

Most organisations already have access to AI capabilities. The challenge is helping people translate those capabilities into their specific workflows. Knowing that ChatGPT, Gemini, and Claude exist is one thing. Understanding how they change the way a finance team closes monthly accounts or how a newsroom manages production workflows is something else entirely.

This is one reason engineering teams often move fastest. Their use cases are easier to identify, easier to test, and easier to measure. The connection between the tool and the outcome is more obvious.

Business functions can struggle more. Even inside AI-native companies, including those producing the foundational models, teams need help understanding which tools are best suited for which tasks. Add the pace of change — new models, new interfaces, new capabilities arriving every few months — and sustained adoption becomes even harder.

What’s working

The organisations seeing the most progress tend to focus on practical enablement rather than abstract AI education. In practice, that means scaling adoption team by team rather than trying to transform the entire company at once.

The most successful approaches share a number of characteristics:

• They use hands-on workshops and working sessions to show teams how AI tools apply to their actual jobs rather than relying on generic training programmes.

• They focus on demonstrating immediate value. Employees are far more likely to adopt new ways of working when they can see a task becoming easier, faster, or better within minutes.

• They provide practical guidance on which tools to use for which tasks. As the number of AI tools continues to grow, clarity often matters more than access.

• They work with real examples from people’s day-to-day responsibilities, allowing teams to experiment on their own workflows and experience the benefits directly.

• They create bridges between business teams and technical teams, translating business requirements into prototypes, workflows, and practical use cases employees can understand and test.

Most importantly, they let subject matter experts — whether in HR, finance, legal, marketing, or the newsroom — experiment inside their own work.

This is why we see AI adoption increasingly looks like an organisational transformation challenge rather than a technology deployment project.

Rethinking how the organisation works 

The shift isn’t simply about introducing new tools. It’s about moving people from manual execution toward process design, orchestration, judgment, and higher-order thinking.

That inevitably creates uncertainty. Some routine work will disappear. Existing processes will change. Roles will evolve. Organisations that communicate clearly about what is changing, and why, appear to bring people along faster than those that leave employees to fill in the gaps themselves.

For media companies, the parallels are obvious.

Media is not typically the function driving core technical innovation. Yet it faces the same pressure to adopt rapidly.

The most effective starting point is usually not an overarching grand AI strategy. It’s workflow strategy.

Where are the repetitive tasks? What are the bottlenecks? Which processes create frustration for journalists, product teams, commercial teams, or audience specialists every day?

Those are often the first opportunities.

The organisations making progress are starting with practical workflow improvements, working closely with subject matter experts, and focusing on demonstration rather than persuasion.

Show, don’t tell.

Ultimately, the bigger question isn’t “How do we use AI?” It’s “What does the future organisation look like?” 

If you’d like to subscribe to my bi-weekly newsletter, INMA members can do so here.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .