Kathleen Mitford: „The technology is here, but the culture needs to catch up“

Ad – Kathleen Mitford, Corporate Vice President of Global Industry Marketing at Microsoft, sits down with us to discuss the real-world adoption of generative AI, separating the hype from the practical strategies that define today’s leading companies. We met Kathleen at the Microsoft AI Tour in Frankfurt*.

Kathleen, thank you for being with us. To start, let’s look at the big picture. From your vantage point at Microsoft, which industries are leading the charge in AI adoption right now?

What’s exciting is that we’re seeing adoption across all industries. And if you zoom out a bit, you really see it happening in four big areas.

First, enriching employee experiences. This is the productivity layer, like M365 Copilot, which is seeing broad adoption everywhere.

Second, personalizing customer experiences. This is naturally gaining traction fastest in consumer-facing industries. A great example is Ralph Lauren. We worked with them to create a personalized AI-powered shopping assistant on their website. You can ask, “I’m speaking at an event in Frankfurt, what would you recommend?” and it provides outfit ideas, much like an in-store personal stylist.

Third, reinventing business processes. This is where things get more complex.

That sounds significantly more complicated to implement.

It is. From my conversations with customers, I see the financial services industry as the furthest along here, followed by manufacturing and industrial. For instance, Krones, which makes bottling equipment for companies like Coca-Cola, is using AI to create digital twins of their equipment. This is helping them reduce waste, such as liquid splashing out during the bottling process, and improve overall efficiency.

The fourth area is bending the curve on innovation. Healthcare is a standout here. Dragon Copilot, for example, is fundamentally changing the patient-doctor interaction. Instead of typing notes, the doctor can be fully present with the patient. Dragon Copilot drafts the clinical notes and, crucially, automatically integrates with electronic health record (EHR) systems. Doctors tell us this eliminates “pajama time”—the hours they used to spend at home writing up notes. It’s so impactful that some have said it kept them from leaving the profession.

We’ve seen the Dragon Copilot demo and it’s impressive. Given the strength of German industry, particularly in manufacturing, what do you see as the primary barrier to adoption? Is it technology, or something else?

Honestly, both in Germany and globally, the biggest challenge is the same one we saw with digital transformation: the people and change management. The technology is here, and now the culture needs to catch up.

So it’s a cultural shift. How does leadership effectively drive that? Must the C-suite become AI experts themselves?

Where we see the most successful adoption, the change is driven from the top. And it’s hard to advocate for change if you don’t use the technology yourself. We see many customers running specific, hands-on sessions for their executive teams to get them using the tools directly.

“It’s hard to advocate for change if you don’t use the technology yourself”

Microsoft uses the term ‘Frontier Firms’ for companies that are getting this right. What defines a Frontier Firm, and what practical lessons can other organizations learn from them?

A Frontier Firm is one that uses AI to transform every part of its business: every process, every function, every role. They believe AI will fundamentally transform their entire operation.

Rolls-Royce, the turbine maker, is a perfect example. They use generative design in the engineering and simulation process. They use AI on the factory floor to run more efficiently. And they use it for predictive maintenance on engines in service, which keeps airplanes in the air longer. They’ve reduced around 400 unplanned maintenance events annually, saving millions in repairs. That is a frontier company.

For a company that isn’t yet ‘frontier,’ what is the single most important piece of advice you’d offer?

Just get started. AI is still new. Rather than waiting for the perfect use case, experiment, fail fast, and move on.

This is what I do with my own team. I am naturally a very analytical person who likes a structured project plan. But with AI, we allow it to be “messy.” We let every employee experiment with building agents and using AI in new ways. This generates so many ideas that a central team would never have thought of.

That idea of ‘allowing it to be messy’ sounds like a significant shift in leadership style.

It is exactly a change in my leadership style. I am a very buttoned-up, structured person. My leadership has had to change to allow for less structure and more experimentation. We still have top-down directives for specific AI use cases, but we allow this bottom-up creativity to happen simultaneously.

“I am a very buttoned-up, structured person. (With AI) my leadership has had to change to allow for less structure and more experimentation.”

So you encourage experimentation, but at some point, you have to formalize it. How do you scale those ‘messy’ ideas?

Exactly. For example, my team built an agent to reduce the time of legal reviews for marketing materials. Anything we publish must go through our legal team, which often would require multiple review cycles. My team built an agent trained on Microsoft’s guidelines. Now, they run materials through the agent before meeting with legal, and it’s reduced the number of reviews down to one or two.

This seems to lead to Microsoft’s concept of the ‘Agent Boss.’ Could you define that for us?

Think of it this way: Today, I manage employees. In the future, I will also manage agents. An “Agent Boss” manages a team of both humans and AI agents, delegating the repetitive, administrative tasks to the agents so people can focus on high-value work. For example, my team, we develop many stories about customers in the AI transformation journey. The fun part is the interview. But sometimes, you get stuck getting started on the story writing. We built an agent that helps us beat the “blank page problem” and gives us ideas to start with. Then, the agent helps us make sure the story is formatted correctly and following the company’s editorial style, which can be the tedious part of publishing customer stories. That saves an enormous amount of time.

This raises a provocative question. Some experts theorize that if individuals can manage powerful AI agents, the very concept of the firm becomes obsolete. Does AI kill the traditional organization?

I don’t believe AI kills the firm, but I do think firms are going to look different. The roles within them will change. Frontier firms are already reimagining their organizational structure.

I think you will see more managers, but they will be managing agents in addition to people. You may see professionals managing AI agents earlier in their careers. Also, if you look at the statistics, we have a global workforce gap. We need AI to help fill that gap because the amount of work isn’t decreasing, but the number of people entering the workforce is.

But if young people can be so effective as self-employed ‘Agent Bosses,’ why would they choose to join a large, bureaucratic organization at all?

We cannot underestimate the human connection. As humans, we get joy from engaging with other people. The culture of a team, the sense of a shared mission—that is not going away.

When you think about the industries we serve - healthcare, education, citizen services - the impact comes from people working together. As leaders, we must continue to reinforce and enable that human connection, even as we adopt non-human colleagues.

As we conclude, you’ve been at the forefront of major tech waves, from IoT to the cloud. Now we have generative AI. From your personal perspective, what makes this transformation feel different?

It is, without question, the rate of adoption.

I’ve been in this industry for more than 25 years. With IoT, we talked about “pilot purgatory” for years. I am seeing AI adopted at a pace I have never witnessed. A large part of this is the foundation that has been laid. Companies have spent the last decade on their digital transformation, moving to the cloud and getting their data estate in order. AI is only as good as its data. Because that foundational work is done, they are able to move so much faster with AI than with any technology I’ve ever seen.

__________

* We attended the AI Tour Frankfurt at the invitation of Microsoft. There was no editorial influence; this piece was produced independently.