In the breathless pace of today’s business environment, timing is everything—especially when you’re walking into a client meeting. Having the right information at your fingertips can mean the difference between a good conversation and a game-changing one.
Now senior colleagues have a powerful new tool, the McKinsey Value Intelligence (MVI) Assistant, developed on Google Cloud using Gemini, which can surface essential news, insights, and trends from a vast repository of 50,000 public companies. It has been thoughtfully designed to capture the kind of actionable information that can help leaders build trust, deliver high-value advice, and deepen client relationships. This effort builds on McKinsey’s alliance with Google Cloud to help clients catalyze large-scale, innovative change through the power of cloud and AI.
The best information at the right time
Whether preparing for a high-level meeting or a discussion of strategic opportunities, a user can:
- review key topics—and tough questions asked—from the latest earnings calls
- compare points of strength and weakness across company peer sets
- help prepare their clients based on potential investor questions
It offers a nuanced understanding of competitor dynamics and the market reaction and can help inform strategy.
A new approach to navigation
This new tool extends the capabilities of an existing asset, the McKinsey Value Intelligence Platform, which has produced insights for the Strategy and Corporate Finance Practice for over a decade. But it also reflects the firm’s broader Rewired transformation using gen AI to reduce the time it takes to extract information from vast amounts of unstructured data and deliver meaningful insights to clients.
The team designed and built an early version of MVI Assistant in less than four months and took a fundamentally different approach to usability. “We realized that traditional prompt-based tools often left users unsure of where to start: what to ask, how to phrase it, or even what information they needed to upload,” explains Bridget Haby, an asset leader at McKinsey. “So, we designed a streamlined, click-first experience that puts the most relevant content just one search away.”
McKinsey colleagues can simply enter a name into a company search bar and quickly see key topics, recommended analyses, financial benchmarks, news, and related firm content.
Within the first three months of launch
The tech behind the “click”
A key factor of success was the ability to leverage an existing set of curated financial analyses (i.e., charts) and a robust set of APIs enabling easy retrieval of company documents. The ability for Gemini’s model to handle a higher context window—or larger amount of input data (text, code)—also significantly accelerated the team’s timeline by allowing non-technical team members to conduct most of the prompt testing and curation.
The platform is built using a series of LangServe prompt workflows to deliver curated content at the click of a button. API calls are used to fetch the appropriate data from the platform based on a company ID, and then a series of prompts can extract key insights from the unstructured call transcripts, and then recommend the best financial analyses, news, and McKinsey content to show based on a theme.
MVI Assistant is just one example of how McKinsey is using gen AI to support smarter, faster decision making across the firm and beyond.