Case study

Rewiring the way McKinsey works with Lilli, our generative AI platform

On this page: The opportunity | The impact | The rewiring | What’s next

Learn how we built a next-gen platform to transform our client work

But why stop there? Testers of only a primitive Lilli prototype immediately leveraged it for everything from data analysis to planning and creative problem solving. This indicated that Lilli could become a platform on which we build gen AI–based applications that completely rewire the way we operate.

“We chose to take a page out of our own book, quite literally, and approach this as part of our own Rewired transformation with clear targets,” says Rodney Zemmel, the senior partner who leads McKinsey’s technology practices, including the firm’s own AI transformation.

Such a transformation would enable colleagues to work more effectively and efficiently together in the emerging AI-first world and deliver the best of the firm to clients to maximize their outcomes.

“We recognized the potential of gen AI to transform not only how we operate internally but also how we create deeper relationships with business leaders and deliver even greater value to our clients,” says Erik Roth, the senior partner who leads the firm’s development of Lilli. “However, to do it, we had to set up a new operating model that brought together parts of our firm that historically had not collaborated deeply.”

The Impact

Much more than a research partner

Since Lilli’s firmwide rollout in July 2023, the platform has been widely adopted, with 72 percent of the firm active on the platform, and colleagues reporting up to 30 percent time savings in searching and synthesizing knowledge. More than 500,000 prompts every month (and growing) help colleagues around the world and across functions in numerous ways, including the following:

And they can engage with Lilli confident that all interactions and data are secure thanks to McKinsey’s relentless focus on risk controls and security from day one.

While it’s still early days, we see the beginnings of a shift in how we operate. For example, colleagues and leaders now think more critically about how they create and codify knowledge, with an eye toward how it can enhance Lilli’s capabilities. And colleagues have reshaped their workflows to incorporate the power of Lilli.

Perhaps more important than the benefits we’ve realized internally are those we’ve been able to share with clients. The hands-on experience we’ve gained from developing and evolving Lilli provides valuable guidance for hundreds of clients pursuing their own gen AI endeavors. And we offer them a version of the architecture underlying our gen AI platform that is customizable for their organization’s specific work and industry—a key to both platform adoption and competitive advantage.

“Our clients are getting value from specialized knowledge agents similar to Lilli that are tuned to assist their employees with tasks specific to their workflow and industry. They appreciate learning from our experience creating agents in a scalable and responsible way that drives adoption,” says Delphine Zurkiya, the senior partner responsible for extending such solutions to our clients.

_{1. Blind scoring of the quality of outputs from participants with and without Lilli, using a predefined five-point scale for accuracy, content richness, and distinctiveness.}

The Rewiring

Learning by doing

When we kicked off the Lilli project, few—if any—large organizations had built anything like it. The ongoing development journey has been exciting, challenging, and filled with lessons. Following our Rewired framework has proven critical in guiding us through the process.

Our business-led roadmap isn’t about building a tool; it’s about transforming how we manage and access knowledge and expertise, which is the foundation of our organization and competitive advantage.
Rodney Zemmel
Senior Partner and leader of McKinsey’s technology practices and firm’s AI transformation

What’s next

Innovating to meet new challenges

A year after launch, the Lilli team continues to deliver a robust pipeline of improvements, both in the core experience and in feature innovations. The most anticipated new features? The addition of agents built to automate specific time-consuming tasks and expansion of Lilli’s slide-building capabilities.

While adoption continues to grow, there’s more work to do to boost regular engagement with the platform. Lilli upgrades make up only a fraction of the effort to that end—providing continued education, fostering gen AI enthusiast communities, and employee skill development will continue to stand front and center to drive our transformation.

Meanwhile, our colleagues at QuantumBlack, AI by McKinsey, are working with clients to bring the power of Lilli to their own organizations to help them with everything from creating a request-for-proposal engine to speeding their drug development operations. These colleagues bring with them the innovations of QuantumBlack Labs, which regularly contributes its code to open source and has developed more than 25 proprietary tools and industry-specific applications to help organizations bring AI to bear at speed and scale.

While the future opportunity for generative AI is still being written, Erik reflects on the journey McKinsey has taken with Lilli so far: “We have created something that I believe will change the way we serve our clients. We have the opportunity to set up our firm for success in its next century and perhaps redefine our industry.”