206,000 COVID Tests in a Day & $5BN in Revenue | Curative Co-founder & CEO

Why does it matter? Because the AI transition at Curative already happened — most companies are still running pilots.

Curative built an AI agent that finds doctors, emails them nine times, negotiates rates, redlines contracts, and clicks sign on DocuSign with the CEO's own signature. It outperformed the entire human contracting team in 8 weeks. Fred Turner didn't run a controlled experiment — he replaced whole departments at a 650-person regulated healthcare company and walked away with a structural cost advantage that incumbents will spend a decade trying to replicate.

• One agent did 3,500 contracts in 8 weeks vs. 2,300 for the full prior-year team — at $70 per contract vs. $1,500–$2,000 with humans • 80% of Curative's SaaS spend is being cut this year; Salesforce replaced by a vibe-coded CRM in two months, $600K gone • Current AI can handle every back-office task at Curative today — deployment, not capability, is the only gap • Obamacare's 85% medical-loss ratio cap accidentally rewards insurers for spending more, making AI-driven efficiency a moat incumbents structurally can't close

One AI agent did more contracts in 8 weeks than Curative's full team did all year

3,500 contracts in 8 weeks. The entire 45-person team did 2,300 in the full prior year.

Gwen — Curative's AI contracting agent — autonomously finds primary care offices, researches their practice, pulls payer transparency data to learn what competitors are paying them, sources emails from ZoomInfo, then pings providers repeatedly with customized outreach. She sends 15,000 emails a day. "A lot of providers will get them on the ninth email," Turner says. "There's no way that a human is going to email them nine times."

When a provider responds, Gwen negotiates rates across multiple rounds, redlines Word documents by writing Python to edit them (the models can't directly edit Word well; they're excellent at generating code that does), and closes the deal. She opens DocuSign, clicks sign, and it's Turner's signature on the agreement — legally binding. Getting the internal team comfortable with that last step required several rounds of convincing.

Cost per contract with people: $1,500–$2,000. With Gwen: $70. Turner says the economics hold even if Anthropic quintupled its prices. The human team didn't go to zero — it got redirected to hospital groups and broker relationships that require in-person meetings. But Curative isn't asking the agent to replicate the old weekly output. The target is 10 times as many contracts this year.

Curative is cutting 80% of its SaaS spend this year — and there's a whole internal slide for it

One internal meeting at Curative features a dedicated slide: every SaaS renewal date, and the person responsible for telling that vendor they're not renewing.

Salesforce is already gone. A vibe-coded internal CRM, built in two months, replaced it and works better. "$600,000 a year. Gone to zero." The full-time Salesforce administrator's role dissolved with it. The off-the-shelf claims processing platform Curative migrated to a few years ago is being phased out by an in-house build going live in July. Looker was replaced with Snowflake dashboards, migrated by an agentic workflow that converted each dashboard automatically — a project that would have taken a year with traditional headcount became a two-person, two-month effort.

The line between what survives and what doesn't: infrastructure with deep integrations (Sentry, Slack) is sticky. Anything that is fundamentally a database with a UI and a workflow on top is a replacement candidate. That's the category that built the SaaS industry. Turner's frame for every renewal: not "is this tool useful?" but "can we build it in two months and own it?"

Every back-office task at Curative can already be done by AI today — only deployment stands in the way

Every back-office task at Curative can already be handled by AI today. Turner was uncertain about that a year ago. He isn't anymore.

"The current gen models can do every back office task we have at curative — it's just a matter of deploying them, getting them set up, getting them configured, having the right policies."

The credentialing department made it concrete. Verifying that network doctors hold valid licenses and aren't facing malpractice suits — historically two to three months per doctor, roughly $50 each — required a person to visit medical board websites, pull transcripts, and search liability databases. An agent built on Claude now does the full workflow end to end. Turnaround: 12 hours. Cost: 20 cents. The credentialing team went to zero people.

The changed belief matters more than the specific case. Turner isn't forecasting what AI will eventually do — he's describing what it already did, across multiple functions, in the last 18 months, at a company operating under healthcare regulation. Organizations still running AI readiness assessments are solving a problem that has already been solved.

AI doesn't just reduce headcount by 1x — it makes 10x the volume possible

$70 per contract. That's Gwen's cost. The human team ran $1,500 to $2,000. And Turner's first move wasn't to cut the team — it was to ask what a 10x volume target looks like.

"You don't just displace the labor. We've not said okay we're doing 100 a week so we'll get the agent to do 100 a week. What we've done is said, 'Well, now that we have the agent, we can do 10 times as many contracts this year as we could do last year.'"

Jobs that survive are ones attached to technical depth or genuine relationship capital: engineers deploying and supervising agents, account managers closing broker deals that require dinner and a finalist presentation. Everything in between is being reassessed. Curative is heading from 650 employees toward roughly 400 near-term, before growing again as membership scales. The clinical side — care navigators who stay with each member throughout their journey — grows linearly. The back office gets rebuilt differently.

Model AI ROI as a volume multiplier and the growth picture is unrecognizable from a cost-reduction frame. What market share becomes reachable when you do 10 times the contract work at a fraction of the cost?

Obamacare's profit cap accidentally rewards insurers for spending more — that's the moat Curative is exploiting

Obamacare designed a rule to protect consumers. It built in a financial incentive to spend as much as possible.

The medical-loss ratio requirement: 85% of premiums must go to paying for care. Fall short and you give money back to employers. "From an insurance company standpoint, if your profit's capped at 15%, the only way to increase profits is to increase total spending — which is not what you want your insurance company incentivized to do."

Why invest in preventive care? Why build wellness programs? Why cut admin costs? Efficiency in the 15% directly compresses the profit base. The legacy incumbents built 50-year businesses around the opposite incentive.

Curative's AI-driven admin reductions change this math because they're still in the growth phase — expanding into the gap between what they save on admin and what care actually costs. The incumbents can't follow quickly: cutting admin workforces by 50,000 people to gain margin improvement is a political and operational problem that will take a decade. Turner's runway to build share while competitors move slowly isn't just a technology gap. It's a structural one written into law.

A failed Series B is why Curative exists — and why it made $5 billion

Curative exists because a funding round died three weeks into second-round due diligence. A strategic investor's CEO pulled out — the product was too competitive with their own. Three weeks of cash remained.

Turner sold the company's CLIA lab license — two years in the making — for $150,000 to pay creditors. Five months later, he acquired a different lab license for $27 million to run COVID tests.

"Would you be where you are today if that round had come together?" the host asked. "No," Turner said. "I don't think we would have pivoted as hard into COVID when COVID hit."

The license that went for $150K at the end of 2019 was effectively worth $27M by mid-2020. The failure that looked like the end was the only path to the beginning. The question at the moment of collapse wasn't how to rescue the deal — it was what the forced constraint made possible that success had blocked.

Nuclear's safety problem was solved in the 1980s — the obstacle since then has been regulatory, not scientific

Safe nuclear reactors have existed since the 1960s. The technology hasn't meaningfully changed since then. Turner's read: that's not an engineering failure — it's a regulatory one.

The energy amplifier, developed in the late 1980s by Carlo Rubbia (former CERN director, Nobel laureate in physics), operates at 0.97 criticality — permanently below the 1.0 threshold where a reaction sustains itself. A particle accelerator supplies the missing neutrons. Shut it off: the reaction stops immediately. No knife-edge control problem. "There's nothing you can do to it to cause it to go critical or to have a criticality accident."

The only active construction of such a system is in China, using a US design the US abandoned after Fukushima.

Subcritical — the company Turner co-founded with his wife — is betting that the engineering is solved and regulatory navigation is the actual work. For a founder who scaled a $5B testing business by moving faster than established labs inside a crisis, a decades-old licensing framework blocking a proven design looks like exactly his kind of problem.

The deployment gap is the only remaining advantage — and it closes fast

What this episode actually reveals is that AI capability and AI deployment are now separated by organizational inertia, not technology. Turner went from uncertain to certain about what current models can do in a single year — after watching his own back-office functions go to zero. The companies still modeling AI as a cost-savings program are already thinking about it wrong. The ones that ask what market share becomes reachable at 10x throughput are building something incumbents can't reverse-engineer by laying off half their staff. The deployment gap closes. The lead it builds doesn't.

---

Topics: AI agents, health insurance, COVID testing, SaaS disruption, startup building, nuclear energy, agentic workflows, enterprise AI adoption, labor displacement, Silicon Valley vs UK