The Acquisition Layer

Every major acquisition eventually becomes a systems problem.

The press release goes out. The deal closes. The narrative is about synergies, market positioning, combined revenue. What nobody announces is the part that follows: two separate financial systems, two chart-of-accounts structures, two sets of approval workflows, two compliance environments, and hundreds of people who need to keep working while all of it gets reconciled.

The IPG acquisition at Omnicom was $13 billion. At that scale, the integration isn't a project, it's an ongoing operational state. You're running two systems simultaneously while migrating to one, and the cost of downtime isn't measured in lost productivity. It's measured in reporting failures, regulatory exposure, and cash flow blind spots across a global operation.

What I learned: the systems that look most stable are often the most brittle. Years of undocumented customizations, workarounds that became load-bearing, integrations nobody remembers building. The only way through is documentation-first, then automation, then governance. You can't automate what you can't describe.

The AI angle matters here more than anywhere. Acquisitions produce massive amounts of unmapped data, entities, hierarchies, GL codes, cost centers, that need to be reconciled between two environments. An agentic system that can interpret, map, and validate at scale does in days what used to take months of manual analyst work.

Most companies treat integration as a one-time event. The companies that treat it as an ongoing capability, who can absorb, rationalize, and operate new entities without a two-year transformation program, have a real structural edge.

That's what I've been building toward.