Alex Rampell

1. Key Themes

Legal AI Is a Revenue-Expansion Play, Not a Cost-Cutting Play

The dominant narrative about AI is headcount reduction, but in plaintiff law the constraint is labor supply, so AI expands the number of clients a firm can serve. Alex Rampell framed it directly: "Most people think of AI as like, 'Oh, you're going to fire everybody and cut costs.' And it's like, no, actually, like if I use Eve, I make a lot more money for my firm." 00:00:10 Jay Madheswaran's evidence is a concrete before-and-after: "One firm in LA went from managing about 100 clients per attorney to 500." 00:00:19

The Market Failure: Demand for Justice Is Cut Off by Labor Economics

Rampell uses supply and demand to explain why small cases go unserved: lawyers cannot afford to take low-value cases because the labor needed to evaluate and work them exceeds the likely fee. Madheswaran quantifies the cutoff: "today, if a lawyer doesn't make at least 5,000 in fees off a case, they do not take it on." 00:06:47 The result is that people with real grievances can't find representation: "that is why 96% of cases on the labor and employment side, for example, gets rejected." 00:08:46 When AI amortizes the 10 hours of intake and research, the economics of whole categories of cases change.

Software Spend Understates the Market Because Software Is Becoming a Shadow Labor Market

Rampell argues that VCs would have wrongly dismissed plaintiff-law software as tiny because incumbents spent only on Microsoft Office, ignoring that the real constraint was labor: "it turns out the software market is kind of like this shadow labor market, and it's like massively bigger than anybody would have expected." 00:16:09 Madheswaran confirms the historical stack was minimal: "historically, plaintiff law really didn't need more than a cell phone, email address." 00:00:00 The addressable market is the labor and outcomes, not the license fees.

Legal AI Has Been Built for Defense and Institutions, Leaving Plaintiffs Underserved

Budgets sit with big firms and in-house teams, so most legal software companies target them. Madheswaran said: "vast majority of them are on the defense side, because that's where the budgets are." 00:18:36 Plaintiff firms are unforgiving small businesses: "If you're not driving outcomes for them from day one, you're out. You're not getting renewed." 00:19:22 That combination, a hard customer and a neglected side of the market, created the opening for Eve.

Proprietary, Non-Scrapable Data Is the Moat Against Foundation Models

Madheswaran's answer to "why not just use raw intelligence?" is that the relevant knowledge isn't on the internet: "in plaintiff law, the vast majority data is actually non-scrapable. About less than a percent of cases actually settle and get turned into case opinions." 00:21:28 Insurers hold the collective view across 100,000 firms, while each plaintiff firm sees only its own shard. His summary of the gap: "If you drop Albert Einstein into a law firm today, he will not be productive for quite some time." 00:00:41

Long-Horizon Agents: Thousands of Steps per Case

A single case involves roughly 10,000 interactions, so end-to-end automation requires agents with very high accuracy across long sequences. Madheswaran: "to put things in perspective, a single case has probably 10,000 back and forths." 00:48:40 He ties feasibility to accuracy: "as accuracy of Eve is getting higher and higher, this long horizon becomes possible." 00:09:05 Outbound voice agents already collect medical records, a step that takes two weeks manually, and he notes "agents talking to other agents and learning how to get past each other, which is hilarious." 00:48:52

The Lawyer's Role Shifts to Advocacy and "Data Scientist" Work

As routine work moves to agents, attorneys focus on what is distinctively human. Madheswaran: "you as a lawyer are really going to look more like almost a data scientist and engineer, engineering something to be better and better." 00:15:42 Trials become less catastrophic for a firm: "many customers on Eve are actually choosing to spend more time on their trial work... Previously, going to trial took down your entire firm." 00:30:14

Law Firms Become Software-Like Businesses: Specialization and Consolidation

Madheswaran predicts firms will accrue proprietary datasets and specialize at a finer grain: "someone could get really good at soft tissue injury cases that involve X and Y fact patterns, and it advertising for those cases might be a lot more cost-effective than some other larger firm like Morgan & Morgan." 00:38:01 At the same time, large firms are using Eve to acquire others: "some of them actually are coming together... they realized with Eve, they can very quickly consolidate what used to be a very painful reintegration process." 00:38:40 Rampell's thesis is that a small firm could grow into "a billion-dollar revenue business." 00:36:05

Outcome-Based Pricing Is Native to Plaintiff Law and Spreading

Plaintiff law was outcome-priced from the start through contingency fees, and Eve's own pricing evolved the same way. Madheswaran: "from day one, we started probably with the Tall, Grande, Venti model... then... we realized like what they care about is outcomes." 00:44:19 Rampell notes that SaaS giants like Zendesk face the same shift away from per-seat pricing. Madheswaran expects billable hours to follow seats: "the billable hour model, I don't think it's going to die completely, but I think it's going to be on the way of the seat-based pricing model." 00:44:53

2. Contrarian Perspectives

The Legal Software Market Is Not Small; It's a Labor Market in Disguise

Conventional venture logic says a segment whose software spend is "tens of millions of dollars a year" is too small to matter. Rampell inverts this: "the outcomes that they deliver is in the hundreds of billions of dollars a year. They can't deliver more outcomes, because they're really constrained on labor." 00:16:09 Madheswaran extends the point, arguing that unlocking just one case type is enormous: "the amount of value that gets unlocked for the world from that is in the hundreds of billions of dollars. And that's just one subtype." 00:46:56

The Best AI Opportunity in Law Is the Unglamorous Side, Not Big Law

Silicon Valley fixates on elite corporate firms billing $2,000 an hour. Rampell points out the overlooked reality: "Morgan & Morgan is probably a lot more profitable than Wilson Sonsini." 00:01:38 Madheswaran explains why the defense and enterprise side attracted builders: "that's where the budgets are... easy to build for, easy to get money for." 00:18:36 The harder, scrappier plaintiff market was left open.

AI Will Increase the Number of Trials, Not Replace Courtroom Advocacy

The common expectation is that AI automates lawyers away or pushes everything to settlement. Madheswaran sees the opposite: customers "actually choosing to spend more time on their trial work," because automation frees the firm from being consumed by trial prep. 00:30:14 He adds that "a lot more cases are going to have teeth," and that the human lawyer's local, client-specific insight "I don't think... ever goes away." 00:30:14

Small Local Firms Can Become National Specialists, and Mega-Firms Will Roll Up the Rest

Rather than commoditizing, AI lets a local firm convert its data and competence into a national niche business. Madheswaran: "some law firms are going to look more like software companies. I think they're going to finally participate in the software revolution that's been happening for the last 30 years they've been left out of." 00:36:43 He also reports that Morgan & Morgan-scale firms are acquiring others because Eve removes integration pain. 00:38:35

Billable Hours Fade Because AI Makes Effort Predictable

The hourly model persists because lawyers can't know in advance whether a matter is 10 or 100 hours. Madheswaran argues AI changes that: "from a set of facts, you need to know how much compute resources you need to arrive at a outcome." 00:41:48 Once compute cost per outcome is knowable, pricing by outcome becomes feasible even outside plaintiff law.

3. Companies Identified

Eve

Legal AI platform for plaintiff law firms, handling intake through settlement. Co-founded and run by Jay Madheswaran, backed by a16z. Mentioned as the focus of the episode. Madheswaran: "What Eve has aimed to do is really handle that whole thing end to end of how do you represent a client all the way from the very first call to getting them a settlement." 00:00:19 On traction: "We've probably helped a couple hundred thousand clients today, and we have hundreds of millions to touch." 00:51:55 Also: "we're largely limited by engineers at this point." 00:46:56

Morgan & Morgan

The largest U.S. plaintiff firm, with billboards across the country. Cited as proof the plaintiff market is large and profitable. Rampell: "Morgan & Morgan is probably a lot more profitable than Wilson Sonsini." 00:01:38 Madheswaran says Eve works with firms at that scale, some of which are acquiring other firms. 00:38:35

Wilson Sonsini

Elite Silicon Valley corporate law firm. Used as the contrast to plaintiff law and its hourly-billing model. Madheswaran: "you talked about how Wilson Sonsini might charge $2,000 an hour, and that's really the business model." 00:02:42

Harvey

Legal AI company oriented toward large firms and institutions. Madheswaran references it as an example of defense and institution-focused legal software: "For Harvey, you know, really any company, honestly, legal software companies today, vast majority of them are on the defense side, because that's where the budgets are." 00:18:36

Anthropic and OpenAI

Frontier model labs. Madheswaran says Eve works closely with them and is betting on model improvement rather than competing with it: "we're betting on these models getting better and better." 00:21:28 He adds that legal benchmarks they work on emphasize single tasks while "the world is now moving into agentic work." 00:46:56

Colossus

Insurance-industry claims valuation software. Cited as evidence of the information asymmetry plaintiffs face: "they even have ominous-sounding software like Colossus, which... try to basically say a broken thumb, this is how much we think it's worth... also saying, 'This is how much we can get away with giving them away for.'" 00:23:14

GEICO and State Farm

Large insurers used as the archetypal repeat adversaries of plaintiff firms. Madheswaran notes insurers see patterns across thousands of firms: "insurance is the one that's dealing with 100,000 different law firms." 00:23:01 Eve helps attorneys learn how to counter their repeated defenses: "State Farm repeats that defense across the board." 00:26:19

Zendesk

SaaS company with $2 billion of ARR when last public, taken private by Permira and Hellman & Friedman. Used by Rampell to illustrate the shift from seat-based to outcome-based pricing: "How many seats do you need? Like maybe zero. We need to deliver outcomes." 00:39:56

Permira and Hellman & Friedman

Private equity firms that took Zendesk private; mentioned only in the context of that pricing-shift example. 00:39:15

McDermott Will & Emery

Law firm of Rampell's trust and estate attorney, mentioned as an example of an expensive hourly-billed practice where outcome pricing is awkward. 00:39:15

Muse

Mentioned by Rampell as an example of consumer agents that can call businesses (e.g., for restaurant reservations), foreshadowing agent-vs-agent dynamics. 00:26:18

Wells Fargo

Not an excellent company, but the target of Rampell's anecdote about using ChatGPT/Codex to sue it, which he used to illustrate how AI is collapsing the effort of legal filings. 00:33:16

Blue Bottle

Coffee shop in Los Altos where Rampell explained Eve to a trial attorney who then asked to try it. 00:10:17

EEOC and USPTO

Government bodies mentioned as bottlenecks and alternative dispute channels: the EEOC for employment cases 00:34:13 and the USPTO as the remaining bottleneck in patent filing. 00:33:16

4. People Identified

Jay Madheswaran

Founder and CEO of Eve; formerly a venture investor, per Rampell ("You were a VC once upon a time" 00:36:05). Mentioned for building a platform that reframed plaintiff-law software as a revenue-expansion product. His core claim: "what Eve has aimed to do is really handle that whole thing end to end." 00:00:19 He also stated: "think about giving David a slingshot to take down Goliath." 00:02:42

General Partner at a16z and Eve investor, host of the conversation. His framing of AI as revenue expansion and law as a "shadow labor market" shaped the discussion. Quote: "software market is kind of like this shadow labor market, and it's like massively bigger than anybody would have expected." 00:16:09

Albert Einstein (hypothetical)

Used by Madheswaran as a thought experiment for why raw intelligence isn't enough in plaintiff law: "If you drop Albert Einstein into a law firm today, he will not be productive for quite some time." 00:00:41

5. Operating Insights

Meet Customers at Their Real Bottleneck, Which Shifts as They Grow

Madheswaran described how a firm's constraint moves with scale: a solo attorney is limited by call handling, a systematized firm by its intake team, and a larger one by case-management handoffs. "the bottleneck keeps shifting... when you increase your marketing spend... Suddenly, now that team is overwhelmed." 00:12:03 The product lesson is to cover the full workflow so growth at one stage doesn't just move the pain downstream, which is why Eve consolidates "all of that into the Eve platform such that they are getting pulled in only for what they need to."

Make Tacit Local Knowledge Explicit and Encode It as Process

Eve's value comes from converting scattered tactical knowledge into reusable workflows. Madheswaran: "if you go and Google, how do you write an amazing demand letter for against State Farm for soft tissue injury cases? Like, that today exists nowhere." 00:23:45 His insight about the new software paradigm: "once you do a task manually, even tasks that include crunching new piece of data and transforming it... you can now encode it as a process." 00:19:22 Operators in any services-heavy field can look for knowledge that lives only in senior people's heads.

Build Quality Assurance Into the Platform, Not Just the Workflow

Eve hires its own attorneys to track outcomes and runs an overnight review product: "we have this product called Auditor, which actually analyzes every single case through every single page overnight and looks for situations that a perfect attorney would catch every time." 00:26:35 His finding is that "almost every single case has missed information." Pairing automation with an always-on auditing layer both improves outcomes and builds trust with a skeptical customer base.

Respect Industry Norms on Data to Earn Sharing

Rather than forcing a new data regime, Eve mirrors existing professional norms: "for their own cases, they want it to be very private... attorney-client privilege. And we don't train on that data ever," while replicating how attorneys already share tactics at CLE events. 00:26:18 Operators entering regulated or trust-sensitive industries can gain adoption by amplifying existing norms with software instead of overriding them.

Let Customers Turn Repeated Adversary Tactics into Agents

A customer firm countered defense counsel's habit of timing discovery requests before holidays by building an agent on Eve to respond the same day: "they created an agent on Eve to respond the same day... and the defense were like, 'What is going on here?'" 00:50:59 The tactic is to expose a flexible agent-building layer so customers can neutralize their own recurring pain points.

6. Overlooked Insights

Agent-vs-Agent Negotiation Is Already Happening, and Defense Pricing Hasn't Adapted

Buried in the discussion of medical-records retrieval is Madheswaran's offhand remark that Eve's outbound voice agents are already "talking to other agents and learning how to get past each other." 00:48:52 Later he adds that "all the routine back and forth, I think is going to happen through agents talking to agents." 00:30:14 Combined with his observation that the defense side "very much still works on an hourly business model," 00:50:35 there is a structural asymmetry: plaintiff firms are paid on outcomes and gain from automation, while defense firms and insurers who bill by the hour lose revenue when work gets cheaper. That incentive mismatch could make defense-side adoption of AI slower than expected, an advantage for plaintiff-side tools.

Contingency Plus AI Creates a Natural Funnel for Litigation Finance Data

In passing, Rampell notes that litigation funders already behave like data scientists: "They're using data... Here's the judge, and here's this." Yet "this data just like never really lived anywhere." 00:32:21 Madheswaran's description of Eve tracking success rates across cases, and of Auditor reviewing every page, implies Eve is building the first centralized outcome dataset for small cases. Whoever owns structured, outcome-labeled data across thousands of unlisted settlements holds an asset valuable not only to lawyers but potentially to funders, underwriters, and pricing models, a second-order business hiding inside the first.