Re-Founding Incumbents for the AI Era with Sequence Holdings Co-Founder and CEO Michael Lee

Michael Lee founded Sequence Holdings in early 2023 after recognizing that AI would have uneven impacts across the economy, with certain incumbent-led industries positioned to benefit most from technological transformation. Rather than competing as a startup, Sequence's thesis involves acquiring majority or significant stakes in high-quality incumbent businesses with exceptional management teams, bringing world-class engineering talent to bear, and taking a permanent, long-term ownership approach aligned with sustained value creation.

The company's business model operates as a permanent holding company rather than a fund structure, which Lee argues is critical for several reasons. First, it allows Sequence to attract and celebrate engineers as the core organizational persona rather than investors—a departure from traditional PE firms. Second, it enables long-term capital commitment without fund cycle pressures; Lee notes they aim for one deal per year, not a deployment cadence. Third, it permits retained earnings reinvestment across portfolio companies and creates genuine alignment around building market leaders rather than packaging for exit within three to five years.

Sequence's first investment was BankSouth, initially engaged as a services provider starting August 2023, then becoming a minority investor in March 2024 after Fed and OCC approval. The bank proved an ideal pilot: regulated environment provided well-defined operations and excellent data hygiene (features rather than bugs for AI agents), centralized operations allowed amortization of technology across branches, and the partnership demonstrated the night-and-day difference between service vendor and ownership mentalities. Since March, Sequence built systems reducing average consumer loan underwriting time by 94%, reduced overall loan processing from 30 days to 11 days, and enabled the bank to double loan volumes in Q2 relative to Q1 without increasing underwriting headcount or relaxing standards. One underwriter retired, one moved to sales, yet the organization handled significantly more volume.

The Baldwin transaction—a $7.7 billion take-private of a leading insurance broker—represents validation of the model at scale. Insurance brokerage fits Sequence's thesis: it's a $2+ trillion annual premium market where incumbents possess structural advantages (relationships generate 90%+ gross retention rates, customers don't pay brokers directly so price competition is limited, regulatory barriers protect against startups), the industry has minted many large successful companies, and it's centralized operationally. Lee emphasizes they look for companies that can scale to $100+ billion valuations, where incumbents have defensible advantages, and where AI capability overlap with business processes creates something special.

Sequence built Atlas, a reusable platform across portfolio companies comprising four layers: data ontology (making businesses legible to models by defining organizational structures in code), agent builder (constructing high-performance grounded agents), lattice orchestration engine (instrumenting workflows using agents), and artifacts application builder (user-facing interfaces). All core infrastructure built at BankSouth is leverageable at Baldwin and future companies.

Regarding management team selection, Lee emphasizes two critical criteria: operational excellence in their existing industry and demonstrated commitment to technological change before Sequence's involvement. He cites Baldwin CEO Trevor Baldwin's early adoption of Anthropic across the organization and consolidation onto a single AMS platform instance as signals of readiness for transformation. Lee stresses that change management and human engineering are harder problems than technology itself, requiring sensitive, ongoing attention to employee anxiety and ensuring AI elevates rather than displaces roles.

Lee contrasts Sequence's model with traditional PE by noting PE firms are structurally incentivized toward capital deployment and three-year exit timelines, making radical transformation difficult. They celebrate investors rather than engineers, limiting talent recruitment. They operate under LP pressure and fund structure constraints. Sequence instead operates under permanent capital, celebrates engineers, takes multi-decade horizons, and maintains extreme selectivity—potentially doing zero deals in a year if the right asset doesn't materialize. He positions this as fundamentally different genetics than incremental value creation through consolidation and multiple expansion.

Drawing on his career spanning Goldman, Apollo, Apollo's deal structuring, and Lone Pine's study of best-in-class companies, Lee articulates how his evolution informs Sequence's approach: combining PE's capital structure discipline, understanding of leverage and incentive design; public markets' appreciation for great business compounding and long-term value creation; and venture's recognition of exceptional people's ability to will something from nothing. He emphasizes that whether early or late stage, backing exceptional people in huge markets working on hard problems consistently outperforms other signals.