Yann Magnan says 73 Strings applies AI only to deterministic data collection, leaving valuation judgment to clients for accountability.
73 Strings keeps an audit trail from each KPI to source documents so clients can prove numbers are sourceable.
Magnan predicts daily pricing of alternatives spreads fast as 401(k) plans enter, requiring daily NAVs even for $20m raises.
AI takes over the spreadsheet work behind private asset valuations
Valuing private assets has always been manual work wrapped around human judgment, and the judgment still cannot be automated. What changed is the cost of software and the source of the money. Cloud made a purpose-built platform affordable, while retail capital through evergreen funds and possibly 401(k) plans turned valuation from a quarterly report into a transaction-grade price investors buy and sell on. Magnan expects daily NAVs and more transparency, and 73 Strings works only on the deterministic half, reading portfolio documents and tracing every number to its source.
Takeaways
AI stops where valuation judgment begins. 73 Strings applies technology only to deterministic work such as DCF models, multiples, waterfalls, OPM and CVM, plus the workflow from portfolio data to LP reporting. Choosing 13 times EBITDA inside a peer range of 12 to 14 stays with an accountable human.
Evergreen funds make valuation transaction grade. Retail investors transact on the valuation rather than read it. A GP must offer entry at any point and exit mostly at any point, so quarterly reporting for LP financials and regulatory filings gives way to higher-frequency valuation. 73 Strings has clients pricing assets daily.
401(k) money will force daily NAVs. Magnan predicts a significant chunk of the market moves to daily pricing, and fast, because 401(k) plans will require it. A manager raising even $20m from 401(k) plans in a year would need daily NAVs across its portfolio.
Private capital is becoming less private. Magnan argues more transparency in valuation and in the underlying data is the only way to create the trust that pulls significant capital into the industry. Ben Haber agrees that liquidity is very hard to have without a price.
AI's first job is collecting portfolio data. The platform reads PDFs, PowerPoints, Excel files, emails, compliance certificates and board decks to identify, structure and standardise KPIs, most self-reported by portfolio companies. Investment teams may still adjust reported EBITDA for pro forma items such as acquisitions.
Every number traces back to source. 73 Strings keeps full lineage and an audit trail from each KPI to the original document, because source documents are the client's truth and clients have fiduciary duties to prove every number is sourceable. GPs choose how they collect documents.
How 73 Strings came to be
Magnan spent 25 years valuing alternative and non-listed assets, building Duff & Phelps' Paris office and leading its EMEA business.
By 2020 he saw that all of it, his work, his teams' work and his clients' processes, ran on Excel, Google search and email.
Tap a name to go there · tap a line to see why they are linked
On this episode 2
Show 1
The work was structuring data by hand and reconciling EBITDA definitions that differed between companies, with models rerun under different assumptions.
Errors hid in the spreadsheets. An intern could hard-code a formula and nobody would find it.
Two trends converged between 2015 and 2020. Cloud made the technology cheap enough to build on, and a flood of capital into private markets made funds grow fast and demand more structure in LP reporting.
The founding moment was realising a platform could be built at reasonable cost to bring structure to a disjointed process.