$100T is managed by "human duct tape" | E2308

Episode Summary

Executive Summary: The episode centers on Hanover Park, an AI-native fund operations platform targeting the massive but outdated fund admin market. Chris Holonzik argues that legacy fund servicing relies on humans, fragmented tools, and poor data access, and that AI agents plus a system-of-record ledger can automate migration, reporting, and analytics while preserving CPA oversight. The latter half revisits a 2020 Figma interview to show how SaaS-era pricing, trust, and enterprise buying behavior foreshadow today’s AI/software shifts.

Main Topics: Hanover Park’s mission: modernizing fund operations (Priority: 5/5): Chris Holonzik frames fund administration as a huge, legacy-heavy market with low software penetration, where humans act as duct tape across reporting, accounting, and data access. Building a fund ERP and ledger system (Priority: 5/5): The company is not just selling software but rebuilding the core system of record for funds, including complex entity structures, profit-and-loss allocations, and legal terms. AI agents and long-horizon data migration (Priority: 5/5): Hanover Park uses AI to ingest large document sets, map ontology, and automate migration and reporting, with the goal of moving customers from multi-month implementations to one-click onboarding. Human-in-the-loop CPA model and advisory layer (Priority: 4/5): Rather than eliminating accountants, Hanover Park positions CPAs as review and advisory specialists for edge cases while AI handles repetitive bookkeeping and reporting tasks. Pricing, trust, and B2B SaaS evolution (Priority: 4/5): The discussion covers AUM-based pricing, the decline of classic SaaS, and the move toward bundled, transparent, outcome-oriented pricing in financial infrastructure. Revisiting Figma, SaaS, and enterprise adoption (Priority: 3/5): The episode also replays a 2020 Figma interview to highlight bottom-up adoption, SOC 2/security concerns, and how SaaS pricing and distribution patterns evolved into the AI era.

Key Arguments: Legacy fund administration is dominated by human middlemen using tools like QuickBooks, Bill.com, and Excel, creating slow workflows and poor data access. Funds increasingly want their own data back because it is now strategic, not just operational; data is the backbone of investing decisions. The hardest problem is not just reporting, but modeling complex legal entities, LP agreements, and allocation logic inside a ledger. AI agents are now good enough to extract and map fund ontology, enabling much faster migration and reducing onboarding from months to days. The company’s value proposition is a bundled operating system for funds: system of record, automated accounting/reporting, and intelligence layers on top. CPAs remain important, but mainly as reviewers and advisors for edge cases rather than manual processors of routine journal entries. The pricing model mirrors the legacy market but aims to be simpler and more transparent than opaque hourly or per-service fees. Figma’s old bottom-up adoption and usage-based pricing discussion illustrates how SaaS trust, pricing, and enterprise procurement patterns have shifted toward today’s AI product dynamics.

Data Points: Global assets under management: $100 trillion - Used to describe the scale of the fund management market and why outdated operations matter. Company AUM today: $20 billion - Chris says Hanover Park currently serves roughly this amount of assets. AUM at last raise: $15 billion - Referenced as the company’s AUM in March when it raised its prior round. Prior AUM baseline: $1 billion - Chris says the company was at this level 12 months before the current point in the interview. Last funding round: $27 million - Amount raised in the most recent round mentioned by the speaker. Team size: 50 people - Hanover Park says it has about 50 employees in New York City. Growth in team size: 2x in the past quarter - Chris says the company doubled headcount in the previous quarter. Migration time for a VC fund: 6 days - Example of a completed migration from data intake to LP launch. Migration volume: 300,000 documents - Used to illustrate the scale of data Hanover Park ingests for a fund migration. Historical onboarding target: 24 months - Chris contrasts the old expected migration timeline with the faster one now possible. Model capability threshold: 3 to 6 months ago - He says one-click migration became feasible only recently with newer model capabilities. Model reference: Opus 4.6 - Chris cites this as sufficient for the workflow, implying frontier models were recently good enough. Figma enterprise pricing example: Editors only - In the Figma replay, only editors were charged, while viewers were free. Figma user behavior metric: 28 users - Jason references the classic bottom-up adoption pattern where a company discovers many users already on a tool. Savings claimed by Vanta customers: 82% less time on audits - Advertisement read during the episode. Vanta offer: $1,000 off - Promotional discount mentioned for compliance tooling. Sentry offer: $240 in free credits - Promotional credit for new users. Northwest formation promise: 10 clicks and 10 minutes - Advertisement claims for company formation speed.

Pivotal Quotes: "How did we end up in a world with $100 trillion in assets and a terrible fund management stack? It's human duct tape." — Chris Holonzik: Opening framing of the fund-ops problem and why legacy infrastructure persists. "You're stuck with a bunch of human middlemen that are holding your own data hostage." — Chris Holonzik: Explaining the core frustration with traditional fund administration and data access. "B2B SaaS was dead, but everyone's like, you're insane. You want to go build financial info for the most complex investment firms in the entire world?" — Chris Holonzik: Describing the contrarian thesis behind building an AI-native services company.

Implications: The episode suggests AI will reshape high-trust, high-complexity financial operations first, where legacy workflows are human-heavy and data-rich. It also shows pricing, trust, and security remain central as AI replaces SaaS-era tools.