May 2026
AI in PE-backed software: the sequence is the strategy
The capital is there. The mandate is there. The unsexy foundation work is what separates the programs that ship from the ones that stall.
All articles
The capital is there. The mandate is there. The unsexy foundation work is what separates the programs that ship from the ones that stall.
Sell-side bankers run carve-out processes. Operators architect them. The difference shows up where the headline price doesn't.
Most portcos run corp dev as a transaction function. The platforms that compound run it as an operator discipline.
Multi-entity, multi-ERP platforms run on consolidated reporting that hides where the margin actually lives. The work to surface it is unglamorous, and it's where the operating-model thesis starts.
Capital allocation governance is built into PE firms by definition. Inside their portcos, it's often built into nothing. The case for an IC at the platform level -- and what it actually changes.
Why a senior operator at two-to-four days per week solves a problem a full-time hire often makes worse -- and the engagement structures that hold up across the LBO-to-exit lifecycle.
The 100-day plan is fine for the diligence deck. The integration that compounds takes 18–24 months, and the synergy schedule sponsors price often forces decisions that destroy the synergy itself.
Notes from architecting OrchestrAI8 -- what changes about the engineering org, the data layer, and the build-vs-buy frame when the product depends on agents working across systems.