How McKinsey Quietly Built the Most Powerful AI in Consulting

What happens when a consulting giant stops chasing trends and starts writing the playbook? You get Lilli. And she’s not your average chatbot. McKinsey didn’t just adopt AI; they built their own, trained it on a century’s worth of proprietary wisdom, and named it Lilli. Now, this digital powerhouse is transforming consulting from the inside out—replacing weeks of research, accelerating client prep, and setting a benchmark that’s got the rest of the industry scrambling for their toolkits.

The McKinsey Whisperer That Doesn’t Sleep

McKinsey & Company, known for billing clients in six digits and sipping intellectual espresso, has turned its 100-year brain bank into something smarter than a room full of Ivy League MBAs on Red Bull. Meet Lilli—the firm’s homegrown generative AI platform that’s currently answering over 500,000 prompts a month. Lilli was launched firmwide in July 2023. Fast-forward a year, and 72% of McKinsey’s 45,000 employees are using it like their new favorite colleague, one who never takes coffee breaks or forgets where the slide deck is saved.

Source: www.mckinsey.com

According to internal scoring, Lilli has boosted content quality and accuracy by 20% and cut down research and synthesis time by 30%. That’s not just productivity, it’s a McKinsey-style mic drop. So, what does Lilli do? Well, imagine you’re a fresh associate trying to find analogies in obscure industries. Before Lilli, that meant weeks of schmoozing your way through the internal Rolodex. Now, type a prompt, and bam, Lilli gives you summarized insights, source links, and expert names faster than you can say “double-click.”

It’s an Operating System

This isn’t about slapping ChatGPT into a PowerPoint. Lilli was built from the ground up using McKinsey’s 100,000+ internal documents, interviews, playbooks, and IP. It’s trained on 40+ curated knowledge sources, covering everything from supply chain woes to pharmaceutical analogies involving... paint thickness? Yes, that happened. It is not just answering questions, she’s helping write project plans, coaching team discussions, suggesting comparable case studies, and even picking apart flawed logic before client meetings.

“We almost have AI in the room with us,” said Delphine Zurkiya, a McKinsey senior partner. “We often say, What does Lilli think?” They even built a team of 70 experts from cybersecurity, legal, and talent to ensure Lilli didn’t just work, but worked securely. The tool now operates in dual mode: one for McKinsey’s internal treasure trove and another for external sources. And in true McKinsey fashion, they didn’t stop using it. They're now helping clients build their own Lillis, tailored to their industry needs. Think RFP automation, drug development acceleration, and internal agent platforms. McKinsey isn’t just riding the AI wave; they’re surfing it while handing out custom boards.

Other Firms Are Playing Catch-Up

Boston Consulting Group (BCG) is also playing in the AI sandbox with tools like Deckster—a PowerPoint whisperer that 40% of associates now use weekly. They've created over 18,000 custom GPTs since rolling out ChatGPT Enterprise and introduced GENE, a charmingly robotic brainstorming chatbot. Over at Deloitte, the vibe is more Fort Knox. External AI tools are locked down tighter than a tax loophole, and their in-house creation, Sidekick, is strictly for non-client work. But they’re not slacking. Zora AI, their newest fleet of AI agents, is designed to think like humans, and their partnerships with Google Cloud and ServiceNow are pushing toward a multi-agent future.

PwC and KPMG are also in the mix, rolling out platforms like agent OS and Agentspace to corral all their smart bots into unified “AI armies.” But let’s face it—McKinsey isn’t just leading. They’re making donuts in the parking lot while others are still figuring out where the keys are.

The Numbers Tell the Story

This isn’t just about flashy apps. McKinsey estimates that AI can cut market research time by up to 80%, transforming weeks of digging into hours of synthesis. Their broader AI predictions are even more ambitious: $13 trillion in global economic impact by 2030, with consulting positioned to grab a high-margin chunk of that value. Meanwhile, BCG’s Deckster is used weekly by 40% of associates, proving that even a deck-polishing bot can bring real ROI. Deloitte’s AI teams are tackling predictive analytics, operational forecasting, and client-side customization, all at scale. This isn’t GenAI for vanity metrics. This is AI that hits the bottom line.

Duolingo’s AI Power Move

Just when you thought AI couldn’t flex any harder, Duolingo chimed in with a mic drop of their own. The edtech unicorn replaced many of its human course developers with generative AI and rolled out 148 new courses in less than a year. For context, it previously took them 12 years to launch 100. CEO Luis von Ahn made it official: the company is now taking an “AI-first” approach, shifting work like course creation and even performance reviews to AI. That’s not just efficiency, it’s warp speed. And for the consulting world? It’s a shot across the bow.

The New AI Colleague at Work

Here’s the deal: AI isn’t coming for consultants’ jobs. It comes with a keyboard and calendar invites. Tools like Lilli aren’t replacing McKinsey’s sharpest minds; they’re making them faster, sharper, and far less dependent on frantic late-night searches. McKinsey’s next move? Training consultants in prompt engineering, knowledge curation, and smarter documentation practices. This isn’t some plug-and-play software- it’s a full-on system reboot, built around proprietary IP and strategic deployment. Let’s talk numbers for the finance crowd: this isn’t just a digital makeover. It’s IP monetization + productization + margin expansion—in one elegant AI wrapper. So don’t wait around for someone else’s AI playbook. McKinsey didn’t, they built their own. A century of insight transformed into a platform, not just for productivity, but for redefining the consulting model itself. Because of the future of work? It’s already here. And it goes by Lilli. Get smart stories like this delivered straight to your inbox. Subscribe now!