Why hybrid intelligence is the future of artificial intelligence at McKinsey

In 2015, McKinsey acquired QuantumBlack, a sophisticated analytics start-up of more than 30 data scientists, data engineers, and designers based in London. They had made their name in Formula 1 racing, applying data science to help teams gain every possible advantage in performance. Healthcare, transportation, energy, and other industry clients soon followed.

Many times, acquisitions melt quietly into the parent company. This isn’t the case for QuantumBlack; it has been an accelerating force for our work in analytics. Today, it enters a new chapter by officially becoming the unified AI arm of McKinsey. “When we talk about helping our clients achieve sustainable and inclusive growth, AI is naturally part of the conversation. It’s transforming all businesses, including the way we, as McKinsey, serve organizations,” explains Alexander Sukharevsky, who along with Alex Singla leads QuantumBlack.

Grid wall of multiple group photo shots
Our QuantumBlack community through the years
Grid wall of multiple group photo shots

Over the past seven years, the QuantumBlack community has helped McKinsey achieve a number of feats: building and then donating Kedro, an industry-leading developer tool, to the open-source community; being named a Leader in AI; and supporting women in technology through community efforts and mentorship. The team grew quickly, to 400 in 2020, and now has more than 1000 technical practitioners across the globe today.

Along the way, QuantumBlack has been a critical part of many digital and AI transformations across industries. “We have now brought together all of our analytics colleagues under one umbrella called QuantumBlack, AI by McKinsey,” says Alex Singla, “sharing a single culture and strongly defined career pathways, and using common methods and tools.” 

Team members range from deeply experienced data scientists and engineers to AI-fluent business consultants. The firm has also undertaken intensive training and certification in all aspects of AI and machine learning, including digital and analytics risk.

“One thing that hasnt changed: our original principle of combining the brilliance of the human mind and domain expertise with innovative technology to solve the most difficult problems,” explains Alexander Sukharevsky. “We call it hybrid intelligence, and it starts from day one on every project.”

AI initiatives are known to be challenging; only one in ten pilots moves into production with significant results. “Adoption and scaling aren’t things you add at the tail end of a project; they’re where you need to start,” points out Alex Singla. “We bring our technical leaders together with industry and subject-matter experts so they are part of one process, co-creating solutions and iterating models. They come to the table with the day-to-day insights of running the business that you’ll never just pick up from the data alone.”

Our end-to-end and transformative approach is what sets McKinsey apart. Clients are taking notice: two years ago, most of our AI work was single use cases, and now roughly half is transformational.

Yetunde Dada and Kat Shenton of QuantumBlack
QuantumBlacks Yetunde Dada and Kat Shenton prepping for their SXSW presentation on our open-source tool Kedro.
Yetunde Dada and Kat Shenton of QuantumBlack

Another differentiating factor is the assets created by QuantumBlack Labs. “We capture the insights we have learned over the years with industries and fuse them with the best technologies to stay at the forefront,” explains Matt Fitzpatrick, a senior partner who leads QuantumBlack Labs with Jeremy Palmer. These tech assets can solve up to 70 percent of the work that used to be done on a bespoke basis.

“We already know how to tie the analytic model into the client’s data pipelines. Now we have industry models that are plug-and-play with security, scalability, and risk management already baked in,” says Paul Beaumont, a senior principal data scientist based in Singapore. For example, the CustomerOne toolkit for telecommunications companies can reduce time to market for analytics campaigns by 75 percent.

QuantumBlack Labs will expand significantly over the next year: “We want to become a magnet for the best technologists in the world and create assets that bring together all of our knowledge, so we can take this to our clients,” says Matt.

Today, our experts work in major cities around the globe, but one tradition from QuantumBlack’s early days remains. “There has always been beautiful art on the walls, a deep commitment to high-quality design, and a fantastic community and culture,” recalls Kat Shenton, who has been with QuantumBlack since 2017.

The team recently engaged Sougwen Chung, an AI artist and researcher, to create a painting that brings to life the concept of hybrid intelligence and will form the basis for QuantumBlack’s new visual identity.

As a first step in the artistic process, QuantumBlack data scientists processed data from a river to train its CausalNex machine learning model. “It was the ideal tool for this project because it intrinsically requires humans and tech to work together,” explains Paul. Sougwen further developed the model, adding her own biofeedback.

The model guides the movement of two robotic arms that paint alongside her to create a beautiful swirling visual. The result, as you can see in the film on this page, is a meld of artistry and technology that expresses the cutting-edge work we do for our clients.

Never miss a story

Stay updated about McKinsey news as it happens