Lilli: The Next Generation of Consulting

What happens when McKinsey hands the keys to artificial intelligence?

McKinsey & Company just introduced their most productive team member to the world. Bob Sternfels, McKinsey's global managing partner, recently began talking about Lilli, their internal generative AI platform. He acknowledged that Lilli agents are already making many of the slides and charts that the firm used to rely on junior associates for. The expectation is for Lilli only to get better. Is one of the most influential organizations in the world beginning to hand over the keys to AI?

The scale of McKinsey's AI push is impressive, and no doubt the collaboration with AI will help McKinsey deliver more consulting work in less hours – does that mean fees are coming down? Yeah right.

So what does this mean for the future of consulting? And what can the rest of us learn from watching one of the world's most prestigious firms navigate the same technological disruption that's reshaping every industry?

Is It the End of Consulting as We Know It?

Let's be honest – this headline is pure clickbait. The actual answer is much more mundane.

While industry observers debate whether AI represents an existential threat to consulting, the people actually running these firms seem remarkably calm about it. Sternfels has been clear that McKinsey still needs its associates, and their people are going to have bigger impact and solve more complex problems than ever. The AI isn't replacing consultants – it's just handling the more routine parts of their jobs.

This makes sense when you think about it. Consulting has always been about solving problems that clients can't figure out themselves. If a chatbot could replace that kind of work, clients probably wouldn't need to pay McKinsey's rates in the first place. What's changing is how consultants spend their time, not whether they're needed.

The real impact is likely to be on the most junior roles – the analysts who traditionally spent their days formatting slides and crunching data. But even here, the smart money isn't on mass layoffs. It's on teaching these junior people to work alongside AI tools and focus on higher-value activities from day one.

Concern About Automation Eliminating Jobs Isn't New

This hand-wringing about technology replacing workers has been going on forever. I've been recommending that people read Kurt Vonnegut's "Player Piano" to help them think about how automation actually affects jobs. In that novel from the 1950s, workers worried about machines taking over, only to discover that people were still needed to maintain and manage the automation. Jobs changed, but humans remained essential.

More recently, Andrew Yang built his entire 2016 presidential campaign around the idea that technology would eliminate millions of jobs. He predicted that one in three American workers would lose their jobs to automation within 12 years. Yang was genuinely concerned about mass unemployment from AI and robotics.

How did that prediction work out? Current unemployment is sitting around 3%, suggesting Yang may have been a bit pessimistic about our robot overlords.

Is Today Different?

I can't say whether AI will be fundamentally different from past waves of automation, but I can tell you that most of the work I did as a junior consultant had already disappeared by the time I left the firm.

In one of my first training courses, I learned how to properly leave a voicemail. We also got lessons on formatting written emails and subject lines – in case an assistant was printing them out for our client. Somehow I still managed to work pretty long weeks.

The Life of a Slide

When people think about what new McKinsey associates do, they picture someone hunched over PowerPoint all day. That's not entirely wrong, but it misses the point. Creating slides wasn't just about formatting presentations – it was about turning raw insights into something a client could actually use.

By the time I arrived at McKinsey, the technology was already pretty sophisticated compared to the early days when the firm pooled spreadsheet and presentation tasks to a small office of experts trained in Lotus 1-2-3 and Harvard Graphics. We had moved well beyond that.

My first manager explained that our job was to generate at least one valuable insight per day, measured in slides. I'd spend most of my time getting data from clients, interviewing users, and doing analysis in spreadsheets. But turning that information into a clear, compelling slide was a big part of the work that happened after clients went home.

The process was elaborate. We'd hand-draw our charts and send them to the visual graphics department. They'd send back PowerPoint slides that usually had a few typos and needed some grammar cleanup. I'd print everything out and give it to my manager for review. My managers always carried red pens because they marked up everything by hand. I'd either make the changes myself or send the slides back to graphics with the hand-drawn edits.

This system seems almost quaint now, but it worked. The question is what happens when AI can handle most of this process automatically.

Technology Changed Everything, and Nothing

The early 2000s brought some big changes to how we worked. Email became standard, and by the time I was teaching new recruits, we'd stopped the voicemail training entirely. The email lessons got updated too – executives were now reading messages on their BlackBerrys instead of having assistants print them out.

Slide creation evolved dramatically. Visual Graphics became VGI – Visual Graphics India – so we could send work overseas at night and have it back the next morning. Managers got touchscreen laptops and could mark up slides digitally instead of with red pens. We got think-cell for charts, which made creating our beloved McKinsey graphs much faster. At some point, the firm even decided it was okay to use color in presentations.

You would think with all that efficiency in slides, we wouldn't still be working 80-hour weeks. Somehow the hours didn't change, but we would get more done. Sometimes it would just be more slides, but often we would try to "fly the rocket higher with the same amount of fuel," as one of my colleagues liked to say.

Bob Sternfels is still aiming higher. He recently posted on LinkedIn that AI will bring "More creativity. More ability to drive real change. And more time helping our clients across virtually every industry better use gen AI to improve their business." If he's right, that's genuinely exciting. The risk, of course, is that firms will just use the efficiency gains to produce more of the same work rather than tackling more meaningful challenges.

Does Up or Out Apply to Lilli?

Associates join McKinsey with the expectation that they'll learn a ton and grow – either to become a partner one day, or to become a business leader at another big company. With Lilli doing the slide work of many associates and helping partners with proposals that junior partners used to draft, what's happening to all the learning?

There's real value in what people learned from working on slides. For me, taking a slide from scratch to inclusion in a deck gave me a much better understanding of the business. Having edits to words and alignments helped me get better at explaining things to other business leaders. When my manager marked up my work with that red pen, I wasn't just learning PowerPoint – I was learning how to think clearly and communicate effectively.

But now that AI can handle much of this work, where do junior people get that experience? McKinsey has always operated on an "up or out" philosophy – you either advance or you leave. The question is whether this applies to Lilli too, or whether AI agents will just keep getting better while junior humans get fewer opportunities to develop critical skills.

This isn't just a McKinsey problem. Consulting, coding, accounting, lawyering, and many other professions rely on apprenticeship models where junior people learn by doing routine work under supervision. When AI can handle the routine work, what happens to the apprenticeship?

There are now so many things that big enterprises don't need to teach junior people. But when you're using AI to accomplish what you need, the AI is also gaining experience. Thinking big picture, what experiences do we want – and need – the next generation of business leaders to be trained on?

The real question isn't whether consulting is changing forever. It's whether apprenticeship is going extinct.