Prakhar Dixit

Prakhar Dixit is a partner in McKinsey’s Seattle office who helps product and engineering teams redesign how they build software with AI. For one airline, he designed and deployed a new product-development process. McKinsey reports that the airline brought products to market 60 percent faster. At a global bank, Prakhar led the work to identify where AI and other changes could improve software development, then helped put those changes into practice.

Before joining McKinsey, Prakhar worked in product management and software engineering. He worked on Firefox at Mozilla and held technology and product roles at Meta, British Telecom, and JPMorgan Chase. He holds a Bachelor of Technology from the Indian Institute of Technology Roorkee and an MBA from the University of Michigan’s Ross School of Business.

From a feature request to a specification

At Sonar, engineers could often make sense of a backlog item because they knew the product well enough to fill in what the request left out. Prakhar was part of the McKinsey team that worked with three Sonar engineering teams to make that knowledge explicit. Together they agreed on what a feature proposal should tell an engineer, including who would use the feature and how to tell when it was finished. They built a backlog agent that used those criteria to spot gaps and ask people for the missing information.

Prakhar’s 2024 research on software teams’ skills examined how employers could develop those skills as roles changed. By 2026, he and his coauthors were describing two emerging roles they called definers and builders. A definer brings product and design decisions together in a specification. A builder takes responsibility for implementing it, testing it, and running the software. Prakhar has seen clients organize that work in teams of two to four on one- or two-day cycles. Agents work overnight and people review the results the next day. As a team gains confidence in a particular step, it can move human review to a later point in the cycle.

What a player would miss

Prakhar also leads McKinsey’s video-gaming work in North America, advising clients on studio investments, growth strategy, and AI in game development. He and his coauthors propose a test for AI in game design. A studio can measure how much time AI saves in production. To judge what it adds to a game, they ask whether players would miss a feature if the studio removed it. They imagine an open-world game where reckless driving gradually makes other characters wary of the player. Designers could set the rules for how characters respond to a player’s history without having to script every encounter.