a16zWhy AI Could Turn Small Law Firms Into Billion-Dollar Businesses
CHAPTERS
- 0:00 – 2:32
Eve’s mission: end-to-end legal AI for plaintiff law and “real people”
Alex Rampell frames the conversation around the less-discussed world of plaintiff law, contrasting it with corporate “big law.” Jay Madheswaran introduces Eve as legal AI built to help people who can’t afford traditional legal services by supporting plaintiff firms from first contact through settlement.
- •Plaintiff law is positioned as a broader, more distributed side of the legal market than corporate law.
- •Eve is described as “legal AI for real people,” not institutions or in-house legal teams.
- •Contingency-fee plaintiff work exists because many people can’t pay hourly legal rates.
- •Eve’s aim is to support representation from the first call through settlement, strengthening individuals against well-resourced opponents.
- 2:32 – 8:51
Why small claims go unrepresented: contingency economics and the legal labor bottleneck
Alex and Jay unpack why many legitimate matters never find representation even under contingency arrangements. Jay explains that legal work is labor-intensive and that firms often can’t take cases that don’t clear a minimum economic threshold, leaving many people without practical recourse.
- •People can’t pay hourly rates “on a blank check,” motivating contingency models.
- •Jay says many lawyers won’t take cases that don’t generate at least a certain minimum in fees.
- •Small-dollar harms (like under-deductible vehicle incidents) often get no lawyer attention because the work isn’t economical.
- •Employment matters can require substantial upfront research and intake time before a firm even knows if the case is viable.
- •Jay cites a high rejection rate for labor and employment inquiries due to the time cost of evaluation.
- 8:51 – 10:01
What Eve automates first: intake, qualification, and early case momentum
Jay describes how Eve targets the earliest and most repetitive parts of plaintiff workflows: answering calls, qualifying matters, and initiating preliminary work. The promise is to compress hours of upfront labor into near-immediate progress and change which cases are economically feasible.
- •Eve can answer calls and qualify whether a firm should take a matter based on the firm’s own insights.
- •Jay describes reducing what used to take many hours into immediate workflow steps.
- •Plaintiff cases can stretch over long timelines, with many actions and back-and-forth steps across months.
- •As Eve’s accuracy improves, Jay expects it to handle longer-horizon work and enable representation for more people.
- •Alex frames the impact as expanding the addressable market by making smaller cases viable.
- 10:01 – 14:36
Revenue expansion over cost cutting: law firm growth constraints and shifting bottlenecks
Alex challenges the common AI narrative that value comes from layoffs and cost reduction, arguing Eve enables firms to take more cases and earn more. Jay explains plaintiff firms are often constrained by internal capacity and operational bottlenecks that move as firms scale marketing and intake.
- •Alex characterizes Eve as a “revenue expansion” tool rather than a cost-savings tool.
- •Jay emphasizes demand isn’t the main constraint; the supply of labor and capacity is.
- •Bottlenecks can shift from the lawyer to intake teams or call-center operations as firms grow.
- •Many plaintiff firms are small businesses focused on outcomes rather than operational optimization.
- •Eve is positioned as consolidating intake and workflow so attorneys are pulled in only when needed.
- 14:36 – 15:53
Case throughput transformation: from 100 to 500 clients per attorney
Jay provides a concrete example of operational leverage from using Eve across the case lifecycle. He describes a firm increasing the number of managed clients per attorney dramatically by using Eve for intake, handoffs, and settlement preparation.
- •Jay cites a firm that went from managing about 100 clients per attorney to 500.
- •The firm used Eve across intake, qualification, and workflow handoffs that previously required relays between roles.
- •Jay describes agents in Eve handling tasks from intake through settlement-related preparation.
- •Settlement support includes briefs for mediation and guidance on focus areas and tactics.
- •Jay frames the future as continuous case-forward motion with AI assistance and attorney oversight.
- 15:53 – 20:25
The “software stack” before Eve: a thin toolset and a hidden labor market
Alex and Jay discuss why plaintiff law historically bought little software and why that misled observers about market size. Jay argues the real constraint is labor, and Alex reframes legal software as a proxy labor market that can expand outcomes far beyond prior software spend.
- •Alex contrasts big-firm tools like time tracking with plaintiff firms’ different needs.
- •Jay says many plaintiff practices historically needed little beyond phone and email; even word processing wasn’t always essential.
- •Jay notes the work is less about collaborative redlining and more about adversarial representation.
- •Alex argues software becomes a “shadow labor market” when it substitutes for scarce human labor.
- •Jay observes software investment historically skewed to defense/institutional buyers because budgets were there.
- 20:25 – 25:44
Why Eve beats “raw intelligence”: non-scrapable data, settlement opacity, and local context
Alex asks why a general AI model isn’t enough. Jay explains that plaintiff outcomes rely on information that isn’t readily available online, because most matters settle and never become published opinions, and because strategy depends on jurisdictional and counterparty-specific behavior.
- •Jay says many relevant outcomes don’t appear in public case opinions because most matters settle earlier.
- •Much of the work involves negotiation and tactics not captured in searchable sources.
- •Local court environments and opposing counsel behavior vary, making context essential.
- •Jay describes Eve as organizing multi-attribute information to answer: what’s the best next action right now?
- •Intake is highlighted as the first area where better information and automation immediately improve results.
- 25:44 – 26:58
Privacy, knowledge sharing, and “Auditor”: finding what a perfect attorney would catch
Alex raises concerns about data sharing between small firms. Jay describes following existing industry norms: maintaining case privacy while capturing broader tactical knowledge, and he introduces a product that reviews case materials to flag missing or overlooked information.
- •Jay emphasizes attorney-client privilege and says they don’t train on firms’ private case data.
- •He notes attorneys already share tactics through professional gatherings and CLE-style knowledge exchange.
- •Jay describes tracking success rates and using in-house attorneys to evaluate platform performance.
- •“Auditor” is described as analyzing case documents overnight to identify missed information.
- •Jay claims missed information is common across cases and that the tooling helps reduce that gap.
- 26:58 – 33:01
When both sides get AI: agent-to-agent workflow and the enduring “human touch”
Alex asks what happens as institutions also gain AI “superpowers.” Jay predicts routine exchanges will increasingly be handled by agents on both sides, while human lawyers remain important for nuanced advocacy rooted in client facts and local dynamics.
- •Jay expects routine back-and-forth to shift toward agents communicating with agents.
- •He argues some portion of representation will still require “human touch” to affect outcomes.
- •Jay highlights defense advantages: resources, money, and time, which can pressure individuals financially.
- •He says attorneys’ differentiator becomes insight into the client, facts, and local adversaries.
- •Jay notes some firms use Eve to free time for trial work that previously overwhelmed operations.
- 33:01 – 36:28
Suing with AI and the system-level future: more ADR and AI-assisted mediation
Alex shares an anecdote about using an AI tool to sue Wells Fargo and asks how the legal system evolves as AI becomes common. Jay predicts increased routing of disputes away from courts and more use of AI to accelerate fact gathering and decision support in alternative processes.
- •Alex describes using an AI tool to generate and initiate a lawsuit workflow against Wells Fargo.
- •Jay predicts growth in processes that divert matters from courts, building on existing structures in some domains.
- •He references frameworks like agency/committee pathways and pre-court processes in certain areas.
- •Jay anticipates more mediation/arbitration-style paths that don’t rely on full court proceedings.
- •He suggests AI could help neutral decision-makers by summarizing historical patterns and relevant facts.
- 36:28 – 39:16
Can small firms become billion-dollar businesses? Specialization and consolidation dynamics
Alex explores whether higher throughput means small firms can scale dramatically. Jay argues firms will look more like software companies, building unique datasets and deeper specialization, while larger firms also accelerate acquisitions because platforms like Eve reduce integration friction.
- •Jay says law firms may begin to “participate in the software revolution” and operate more like software companies.
- •He describes the difficulty of reinvesting in better service given typical settlement economics and cost structure.
- •He expects deeper specialization: narrower fact-pattern expertise within already-specialized practice areas.
- •Alex and Jay discuss scaling local strengths into broader geographic reach by building repeatable competence.
- •Jay notes large firms are acquiring others and using Eve to consolidate operations faster and reduce change-management burdens.
- 39:16 – 46:08
The death (or decline) of the billable hour: outcome-based pricing and long-horizon work
Alex connects legal pricing to broader shifts in software pricing from per-seat to outcomes, noting plaintiff law has long been outcome-oriented. Jay explains Eve began with per-seat style pricing but moved toward outcome alignment, arguing more legal work may shift away from billable-hour models as AI makes effort more predictable.
- •Alex argues many clients would prefer outcomes-based pricing over paying for hours and inputs.
- •Jay says Eve initially used a tiered per-seat style model, then shifted as they learned what customers value.
- •Jay predicts the billable hour won’t disappear entirely but may decline in relevance for many areas.
- •He frames Eve’s focus as end-to-end delivery: hand a client to Eve and have the lawyer pulled in only when needed.
- •Jay ties pricing evolution to better predictability of compute/work needed to reach outcomes in some categories.
- 46:08 – 49:37
What it takes to automate a case end-to-end: 10,000 touches and the engineering bottleneck
Jay describes the operational reality of plaintiff cases as thousands of micro-interactions, including gathering medical records and negotiating with counterparties. He says full autonomy isn’t complete yet, outlines early progress with outbound voice agents, and identifies engineering capacity as the main limiter to expanding end-to-end automation across case types.
- •Jay estimates a single case can involve around 10,000 back-and-forth actions.
- •He describes record collection as a time-consuming step that can require outbound calls to insurers and healthcare entities.
- •Jay says Eve has outbound voice agents and has seen “agents talking to other agents” in the field.
- •Full end-to-end autonomy hasn’t happened yet, but he says they are close and seeing early signs with some customers.
- •Jay identifies engineering as the key constraint, and he emphasizes the value unlocked when a single case type becomes fully end-to-end automated.
- 49:37 – 52:10
Defense tactics vs AI responsiveness: the holiday discovery play and closing reflections
Jay gives a specific example of how automation changes adversarial tactics, describing a pattern where defense counsel timed discovery around holidays to create pressure. He explains how a firm used an Eve agent to respond immediately, blunting the tactic, and the conversation closes with both reflecting on the broader impact and ambition.
- •Jay describes a defense tactic of sending discovery requests right before major holidays to exploit workload and timing pressure.
- •A firm used an Eve agent to generate and send responses the same day, surprising the defense side.
- •Jay says these moments reveal what changes when the plaintiff side has comparable resources.
- •Alex reacts to the scale of case “touches” and reiterates interest in expanding Eve’s reach.
- •Jay closes by describing the scale of impact he wants Eve to have and invites people to join the company.
