The Reference That Works Is a Peer, Not a Case Study

Harvey

L2

Harvey's reference architecture was not a case study library. It was a structured trust cascade through a clearly ranked prestige hierarchy. The mechanism worked because legal has one of the most legible prestige hierarchies in professional services: Magic Circle and AmLaw 10 at the apex, Vault 50 and AmLaw 100 below, then regional and specialty firms downstream. A Magic Circle firm's adoption — under the most demanding professional liability standards in the world — was a trust signal that required no further elaboration for any firm at or below that tier. Weinberg articulated the dynamic precisely: "If you earn the trust of a few of those firms, the rest of them will trust you, and the rest of the firms downstream will definitely trust you." (Weinberg, Sequoia podcast.) Allen & Overy was not just a first customer — it was the credibility anchor for the entire subsequent sales motion. The first 50 enterprise customers were all referrals. Harvey ran no case study program, no marketing-produced testimonials for early-stage enterprise sales. The reference was structural: "Allen & Overy uses Harvey. We are in the same peer set as Allen & Overy. Therefore we should evaluate Harvey on its merits." That is a categorically different conversation from reading a vendor-produced case study about an unfamiliar company. By end of 2024, Harvey had reached 28 of the AmLaw 100 and the majority of the AmLaw 10 and Vault 10. Each tier's adoption cascaded trust to the tier below. The reference was not a sales tactic — it was the distribution mechanism itself. Harvey also deployed practitioner-level peer references when needed: its "Legal Engineers" are former Big Law attorneys from White & Case, Latham & Watkins, Skadden, and Paul Weiss. When a managing partner's skepticism required a peer conversation, Harvey's sales team could provide it from within their own team — not by connecting two customers, but by having a former colleague of the prospect present the product. This is the peer reference internalized into the GTM organization.

Hebbia

L3

Hebbia operated in a market where formal references were nearly impossible — PE megafunds do not give public endorsements and NDAs prevented named case studies at the Series A stage. The reference architecture Hebbia built was informal, peer-to-peer, and structurally suited to how the finance community actually shares information. The mechanism was concentration. Private equity at the megafund level is an extraordinarily small community: the ten largest US PE megafunds by AUM include KKR, Blackstone, Apollo, Carlyle, and their peers. Every LP knows every GP. Every managing director knows their counterpart at the competing fund. Winning nine of ten created a market signal that spread through informal peer conversation faster than any marketing program could reach. George Sivulka's observation captures the dynamic: "Finance is the slowest moving, most lethargic Leviathan. It's the worst possible customer base to go after unless you're providing outsized alpha or real value, in which case, the minute that there's something real, finance moves faster than any other industry." (Sivulka, 20VC, January 2025.) The peer network that makes finance slow to adopt is the same network that, once the product proved itself, spread the reference at speed. The SVB crisis in March 2023 was the reference-accelerating event. When PE clients used Hebbia to map their entire portfolio's banking exposure across thousands of documents within hours during a live crisis, the result was the highest-trust reference possible: not a vendor-produced case study, but a war story that managing directors told each other informally. "We used Hebbia during SVB and it worked" carries more weight in the PE community than any KLAS rating or Gartner recognition. Named customer outcomes that function as public reference: Oak Hill Advisors ($108B AUM, T. Rowe Price umbrella) published a 6x ROI figure with 75% reduction in review times. Permira's CTO stated publicly: "This is by far the most advanced tool on the market that we've seen." These are peer-level endorsements from executives whose judgment the target community respects — not vendor testimonials but practitioner-to-practitioner signals.

Abridge

L2

Abridge solved the healthcare enterprise reference problem with a structural move that most companies cannot replicate: it turned its reference customers into investors. The Series B (October 2023, $30M) included Mayo Clinic, Kaiser Permanente Ventures, CVS Health Ventures, and Lifepoint Health as investors. These entities did not just write checks — they deployed the product. The signal to every other health system: "If Mayo Clinic trusts this technology enough to invest in it and use it at scale, your legal and compliance team's objections are weaker." The investor-as-reference-customer pattern resolved the standard enterprise AI objection loop in one move. The typical objection chain for AI in healthcare — legal concern about liability, compliance concern about data, clinical leadership concern about quality — all had to run uphill against the fact that Mayo Clinic's legal team had already signed off. Mayo Clinic is the most trusted brand in US medicine. Its institutional endorsement, expressed through both equity investment and clinical deployment, functioned as the most credible possible peer reference for every subsequent health system buyer. Shiv Rao described healthcare's trust economics: "Trust is the most important currency in healthcare and it's the ultimate network effect. When you think about what trust is — it's some combination of transparency, reliability, and credibility." (Rao, NerdMDs GenAI Series, August 2024.) The Mayo/Kaiser investor-customer combination supplied all three dimensions simultaneously. KLAS recognition compounded the reference effect. Best in KLAS Ambient AI 2025 and 2026 (94.1/100, among the highest scores ever recorded for a first-year category evaluation) gave CIOs a third-party credibility signal that did not require direct peer conversation. Brian Wilson (CCO): "Healthcare moves at the speed of trust, and this affirmation of our technology from healthcare experts verifies the credibility and reliability of the Abridge platform." The KLAS score was the peer-equivalent reference for procurement teams that could not easily arrange a CIO-to-CIO call. The result of this architecture: Abridge grew from ~$6M ARR in 2023 to $100M+ ARR by May 2025 — roughly 17x in under 30 months — in a buyer community that Rao himself called "the slowest moving, most regulated" in enterprise software.