Bringing AI to the Real Economy | Alexander Taubman

What It Covers

Alexander Taubman explains how Long Lake, a three-year-old firm he co-founded, acquires established service businesses across five verticals—HOA management, HR services, specialty tax, infrastructure, and travel—then deploys Nexus, its proprietary AI platform, to drive revenue growth. The firm just acquired American Express Global Business Travel for $6.3 billion.

Key Questions Answered

• •AI Diffusion Gap: The bottleneck in the AI economy is not model capability but real-world deployment. Legacy systems, messy data structures, and change management make diffusion extremely hard. Long Lake's thesis is that buying and operating businesses directly—rather than selling software—is the only way to drive genuine workflow transformation, because ownership enables the control needed to actually change how people work.
• •Nexus Platform Architecture: Long Lake's AI platform shares 70–80% of infrastructure across all five verticals, covering model routing, security, and agentic orchestration. The remaining 20–30% requires engineers embedded on-site in the field across 15–20 states, mapping workflows and integrating with legacy data systems. Each new vertical improves Nexus for all others, compounding the platform's capability with every acquisition.
• •AI Abundance Model: Contrary to the cost-cutting PE playbook, Long Lake's data from HOA management—its longest-held vertical with 20-plus companies—shows headcount growing, not shrinking. When team members become 50% more productive, the correct response is hiring more of them, not fewer. A salesperson hitting 3x quota with AI support makes adding more salespeople the rational move, not reducing headcount.
• •Prepared Mind Deal Sourcing: Long Lake maintains a ranked list of 15–20 target industries, continuously updated based on market dynamics and deal flow. The firm reviews 4,000–5,000 deals annually across sectors and enters only one or two new industries per year. This funnel produces a 3% hire rate for talent and a similarly tight filter for acquisitions, prioritizing businesses with high-90s customer logo retention and 100%-plus dollar retention.
• •Threshold Resistance Applied to AI Adoption: Taubman's grandfather, a retail pioneer, demonstrated that reducing physical friction at store entrances measurably increased foot traffic—even a half-step reduction changed behavior by roughly 20%. Long Lake applies this principle to Nexus: tools must feel so effortless that adoption spreads virally inside acquired companies within days, not months. Forced adoption signals the product is not good enough.
• •Permanent Capital Compounding: Long Lake targets businesses already generating 20–25% EBITDA margins and structures itself to never need external capital again. With $4 billion in projected revenue next year across its portfolio, internal cash flow can fund future acquisitions indefinitely. Taubman cites Henry Singleton's Teledyne—which compounded earnings per share above 40% annually for 20-plus years—as the model for long-term, flexible capital allocation without detailed strategic plans.