Advent International LabsProving the value of connected data by building it first, then gaining buy-in to ship it across the platform

1. The Context

Advent International is a global private equity firm. Its internal Labs team built digital tools for the firm’s Deal Teams, the people responsible for sourcing, evaluating, and managing investments. I joined as a Principal Product Manager on contract in 2021, leading a cross-functional innovation team. I was responsible for setting team OKRs, defining product direction, chairing team meetings including off-site planning events, and engaging directly with Deal Teams to identify opportunities and shape solutions.

The team operated differently from any I had worked with before. Instead of two-week sprints, they ran six-week cycles. Planning happened during whole-team away sessions in different countries. Between those sessions, individual squads had significant autonomy over how they worked. The only requirement was to deliver something of demonstrable value by the end of each cycle.

2. The Problem

The Labs ecosystem had three core products. The Advent People Network mapped relationships between internal staff, their contacts, and external individuals, giving Deal Teams a way to find warm introductions. Company Insights aggregated data about firms relevant to the firm’s investment activity. Knowledge Center, my product, was designed to organise and surface internal documentation so Deal Teams could find the latest version of a document and understand its context.

The problem was that these three products were entirely siloed. The interface had a left-hand menu where you picked a product, which opened its own search bar. Searching in one tool returned no results from the others. There was no combined results page. Searching dropped you directly into a specific record within whichever product you were in.

Most users did not navigate beyond the first search bar on the home page. The other products were effectively invisible.

Product thinking: A feature is only as useful as its visibility to the people who need it. The problem was discoverability. The products were invisible to the people who needed them.

3. What I Built

Knowledge Center had a foundational problem to solve before any of the cross-product work was possible. Files, whether pitchdecks, financial models, PDFs, or spreadsheets, were scattered across deal team members’ inboxes and shared drives. When someone needed a specific document, they would ask a colleague, who would dig through their folders and send whatever they had, which may or may not have been the latest version. Files named “Final Final V3” were a regular occurrence.

Solving this meant working across three areas: understanding and categorising files by type and subject; deduplicating related versions and surfacing the correct, latest one; and capturing metadata so each file could be associated with the companies, people, and time periods it related to. Curation was a separate concern: the ability to feature or flag specific files so that important material could be surfaced deliberately, independent of search.

With that metadata in place, files could be connected to the companies and people they related to. That opened up the possibility of cross-linking across the platform.

I had been pushing for a universal search that would surface results from every product in a single view. Buy-in for a shared search service across the platform needed more groundwork. I built the cross-linking logic within Knowledge Center first.

When you searched in Knowledge Center, it could return documents, companies, and people. From a company page, you could see related documents. From a person, you could see their documents. The connections worked both ways.

I aimed for something reasonable and functional that demonstrated the value of connected data. My longer-term vision included operator-based search, where you could type something like “company:” followed by a name to narrow results, similar to Jira’s filtering syntax. But that could wait. The immediate goal was to prove that cross-linking made everything more discoverable.

Once people saw the working designs and implementation, the argument for a universal search became self-evident. The question shifted from whether to build it to when. I gained buy-in to roll it out as the shared search across the platform.

Product thinking: To prove a concept, build it where you already have control. A working demonstration is more persuasive than an argument.

4. The Outcome

Once launched, the unified search meant users naturally encountered all three products through a single entry point. Products that had been invisible became discoverable, and each became more valuable because they were now connected.

Knowledge Center reached 50% user adoption within weeks. The unified search increased monthly active users across the suite by roughly 30%.