Moats matter more than ever - cost of building is less, product needs to be sticky

The cost-value equation that governs all hiring

• You would never hire a human producing less value than their cost—that math doesn't work.

• AI changes the equation: cost goes down, value stays the same.

• I'd hire someone to answer phones at 4 p.m. but never at 2 a.m.—the value-to-cost equation is inverted at 2 a.m.

• AI makes the 2 a.m. shift viable.

• Infinite opportunities where everyone would want something at $5, but it's only sold for $10—so nobody buys it, so it's not a business.

• AI lets you sell it for $5. Now it's a business.

Why "collect more" beats "save money" as a pitch

• Salient's CEO kept pitching "I'm going to save you money"—people like saving money, but it's not the killer value prop.

• The key thing: Salient collects 50% more revenue.

• Go to a customer and say "I will collect 50% more revenue for you every single month"—that's why the company is growing explosively.

• It's much more about value generation than cost savings.

The full pitch combines both

• We're going to make you more money AND it's going to cost you less—that's very hard to move away from.

• Plus compliance: AI won't say something it's not supposed to say, so you don't go to jail.

• First customer had a $50M/year call center with 40-70% annual employee churn—not from firing, nobody wants the job.

• Yes the cost is much lower, but lead with revenue generation.

"98% of Americans were farmers in 1789. The tractor made some unemployed"

Why differentiation is not the same as defensibility

• AI is an incredible tool for differentiation—a voice agent speaking 50 languages is highly differentiated versus a human.

• But that capability alone is not a source of defensibility.

• The source of defensibility is owning the end-to-end workflow and becoming the system of record.

• It's not the AI in the voice or the ability to summarize documents—it's becoming contextual to all the work the customer has to do.

Why moats matter more than ever in the AI era

• Word Perfect, VisiCalc, Lotus 123—all became toast when someone with more distribution copied them.

• In the old days it took 5 years for the bread to become toast.

• Now you can vibe code a software product very quickly.

• Your margin is my opportunity—and I can vibe code against your opportunity.

• This increases the peril for anyone with a software product that has an enormous margin pool.

• Products have to be very sticky with some unique competitive advantage.

• Data is often that advantage.

How Salient answered the defensibility question

• The key diligence question: imagine there's a company called Taliant and a company called Salient—why does Salient win?
• Answer 1: Data moat—millions of phone calls means they know exactly what script to say.
• Answer 2: They ingest every law as it's even proposed as a statute across all 50 states, sometimes at the county level.
• They're doing things that make it much harder to compete so they won't lose a deal.

Why vertical operating systems are hard to displace

• Toast runs a restaurant's entire business: DoorDash integration, paying wait staff, everything around operating the business.

• A vertical operating system is very hard to displace.

• Same applies to software eating labor—you can't just do the labor and have someone undercut you by a penny.

• You need to build a system of record so customers can't switch to the cheaper player.

What makes consumption-based AI apps graduate to essential

• Many consumption-based AI apps are easy to switch on or off—they struggle to become mission critical.
• The pattern that works: own the end-to-end workflow, build contextually around all the work the customer does.
• Proprietary data that isn't public and creates compounding competitive advantage.
• The more usage, the smarter the product—a reinforcing loop that becomes very difficult to displace.

Eve Legal AI: What Eve actually does:

• Own the whole workflow from intake to outcome.
• Voice agent collects evidence from potential clients. AI sifts through medical records and employment documents.
• Predicts case value based on characteristics. Tells attorneys which cases are worth $50k vs $5 million.
• Then it helps through every phase: drafts medical chronologies, writes demand letters, files complaints.

Why plaintiff legal AI is a better market than corporate legal AI:

• Plaintiff attorneys work on contingency. They only get paid if they win. So if AI makes them more productive, they make more money.

• Corporate attorneys bill hourly. If AI makes their juniors more productive, they actually lose revenue because there are fewer hours to bill.

• The incentives are completely aligned on the plaintiff side.

• Plaintiff attorneys are extremely selective. Out of every 100 leads, they might take 1 case.

• Every case is an investment of their own time and money.

• This creates a natural use case for AI that can help them pick better cases.

Why Eve is defensible:

• The voice agent speaking 50 languages is differentiation, not defensibility.
• The moat is owning the end-to-end workflow and becoming the system of record.
• 100% of cases flow through the product. Attorneys live in it all day.
• As Eve handles more cases, it collects outcome data that isn't public. Labs can't train on this.
• That data makes intake smarter over time.
• They can say "cases with these three characteristics are worth more money." This creates a compounding advantage. More cases means smarter product means more cases.

How Eve expands the market:

Before, attorneys would only take cases worth at least $50k to them. With lower operating costs, they can now take $5k cases. This unlocks a huge supply-demand imbalance on the plaintiff side. More people can get representation.

Growth and go-to-market:

• Growth from 2 to 30 ( million ARR) happened fast. No outbound sales motion needed. Pure inbound demand.
• Market pull was stronger than expected.
• This is the "lazier and richer" thesis in action: AI that drives revenue and saves money at the same time.