Insurance / Insurtech · Deep dive

Federato

RiskOps for insurance underwriting — an AI-native workbench that sits on top of the policy-admin core and steers underwriters toward the risks a carrier actually wants to write.

emerging

The question that decides it: Federato's wedge is an underwriting workbench that scores every submission against appetite and portfolio and hands the underwriter the next best action, sitting on top of the Guidewire/Duck Creek/Sapiens core rather than replacing it. Does that decision-and-steering layer stay essential and defensible — or do the core vendors and carriers' own foundation-model agents absorb triage, winnability and 'next best action' into the systems that already own the policy data, collapsing Federato to a UI on top of someone else's book?

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

Federato sells “RiskOps” — an AI-native underwriting workbench that sits on top of a carrier’s policy-admin core and works to make the underwriter’s next action match the portfolio strategy management set. Every submission is scored on appetite fit, portfolio context and winnability; best-fit deals surface first, a strategy-aligned quote is drafted for review, and a “Control Tower” turns portfolio targets into live guardrails so a book does not drift between quarterly reviews. Founded in 2020 in Palo Alto by two Stanford collaborators who came to insurance through wildfire modeling, Federato has raised more than $180 million — capped by a $100 million Series D led by Growth Equity at Goldman Sachs Alternatives in November 2025 — and counts carriers, MGAs and mutuals including QBE North America, Nationwide, Kettle and Velocity Risk as named customers. Revenue is reportedly in the tens of millions, tripling yearly. Its bet: the underwriting decision layer is a durable category, not a feature the cores swallow.

Founding story

Federato did not begin as insurance software. Will Ross was an early member of IBM’s Watson group, where he worked with The Weather Company — acquired by IBM — turning weather data into environmental models. He went back to Stanford for graduate work in climate and atmospheric modeling and an MBA, and there met William Steenbergen, who was doing computational and mathematical engineering, reinforcement learning and dynamic optimization in Stanford’s HCI group. The two started collaborating on a wildfire-modeling project. Chasing where wildfire risk data actually gets used, they walked into the insurance industry and its underwriting function — and found a problem that fit their thesis exactly.

The thesis: AI is best applied to the optimization no human has time to do — weighing every submission against a live view of the whole portfolio — not automating clerical tasks, where most insurtech pointed. To ground it, the team ran what it describes as 1,200+ hours of interviews with underwriters before settling the product, and leans hard on “designed by underwriters, for underwriters.” The pairing is usefully lopsided — Ross on data, climate risk and go-to-market; Steenbergen on optimization and systems — and both learned insurance as outsiders, which is either the source of a fresh product or the reason they may underestimate how hard the domain is.

How it works

The detail that matters most is where Federato sits: not the policy-administration core (the Guidewire, Duck Creek or Sapiens system that holds the book of record), but the underwriter’s desktop above it — the layer where a human decides what to quote and on what terms — connected to the core through native APIs.

Follow a submission. A broker’s email or ACORD form arrives, and Federato scores it in real time on three axes: the carrier’s written appetite (classes, geographies, limits, hazards), the state of the portfolio (where the book is already over- or under-exposed), and “winnability” (how likely the account binds at an acceptable price). Good-fit, winnable deals rise to the top of the queue; bad-fit risk is flagged or deflected before an underwriter wastes time on it. For deals worth pursuing, the “Submission to Quote” workflow drafts a strategy-aligned quote to review rather than build from scratch — the “next best action.” Above the desk, the Control Tower converts portfolio strategy into live guardrails that catch drift, block concentration and nudge underwriters toward segments management wants to grow, without waiting for the next quarterly review. The company increasingly frames this as agentic — AI that pulls data, checks appetite, tracks policy changes and surfaces recommendations across the lifecycle. A pilot ships in 8-12 weeks; full rollout runs about two quarters.

Product and business overview

The product is a full-lifecycle underwriting platform: submission triage and scoring, the Submission-to-Quote workflow, portfolio steering via the Control Tower, renewal and policy-change management, native integrations to the cores (Guidewire, Duck Creek, Sapiens), and a broker/agent partner portal. Federato targets P&C and especially specialty lines — the judgment-heavy corners where appetite discipline matters most: aviation, commercial trucking, political violence, wildfire and hurricane-exposed property. Its named ICPs are carriers and MGAs, plus mutuals. The strategic claim is that RiskOps is a new category — underwriting operations as a discipline, the way DevOps or RevOps named a layer — and that owning the underwriter’s daily surface between strategy and the core is a defensible place to stand.

Business model and pricing

Federato is enterprise SaaS on annual contracts. Public pricing is not published — as with most core-adjacent insurance software, deals are quoted — but the model is reported to key off premium volume and seat count (number of underwriters), aligning revenue with both customer size and depth of deployment across the underwriting floor. That is a favorable land-and-expand structure: start with one line or a subset of underwriters, then add seats and lines inside the same carrier. Management attributes growth to net-new logos plus “significant client expansion,” which fits. The caveat: no unit economics are disclosed — no per-seat price, no net retention, no gross margin. What is public is direction, not level (tens of millions, reportedly tripling across 2024 and 2025), which shows the motion works without showing how durable or profitable each account is.

Traction over time

Reported product results are consistent across sources: ~90% faster time-to-quote, ~3x more “good” business bound, 50-90% fewer systems per underwriter, and a claimed ~15% hit-ratio lift. The named case studies convince more than the headline stats. QBE North America — $11.5B+ in global gross written premium — reportedly went pilot-to-production in weeks, scaled from 50 to 300+ underwriters across seven lines, and has been replacing its Majesco footprint. Velocity Risk, a specialty E&S property MGA/carrier hybrid, reports an 89% cut in time-to-quote and a 3.7x lift in high-appetite bound policies. Nationwide and Kettle also appear among referenced customers. Two independent signals — Goldman’s diligence verdict of “a step change in ROI” versus prior-generation cores (Nov 2025), and back-to-back tripling of revenue — suggest the traction is real, not marketing math.

Market analysis

The TAM depends on where you draw the line. The narrowest framing — underwriting and rating software — was worth about $7.15 billion in 2025, projected to reach roughly $12.9 billion by 2030 at ~12.5% CAGR (Mordor Intelligence, 2025). The broader P&C insurance software market was pegged near $13.85 billion in 2025, growing to about $32.6 billion by 2035 at ~8.9% CAGR (Market Research Future, 2025). Federato prefers to gesture at the underlying commercial insurance premium pool — roughly $1 trillion globally (Kalepa cites the same figure) — because a platform that lifts loss and hit ratios can argue for value tied to premium, not just a software line. The tailwind is real — social inflation, volatile catastrophe exposure and an aging underwriting workforce all push toward software that enforces discipline. So is the headwind: the incumbents are entrenched, well-capitalized and moving into AI themselves, and the biggest buyers can build in-house.

Competitive intel

The set splits into three layers. Core systems — Guidewire, Duck Creek, Sapiens — are integration partners and the gravest long-term threat; they own the policy data and the carrier relationship, and any of them extending appetite and next-best-action into the core turns the workbench above into an optional skin. Direct workbench rivals are led by Kalepa (NY, founded 2018, $14M Series A led by Inspired Capital; customers Munich Re Specialty NA, Bowhead), whose Copilot attacks the same seat with the same logic; Send Technology does so in the UK/Lloyd’s market; Sixfold brings a narrower GenAI angle. Adjacent point tools like Cytora (submission intake/routing) are the cautionary tale — Cytora was absorbed by Applied Systems in September 2025. The quiet competitor is the in-house build plus legacy suites like Majesco: a large carrier with its own data-science team can replicate appetite guardrails internally. Federato wins today on funding, the depth of its portfolio-steering story (the Control Tower is more than a copilot) and marquee references like QBE. It is exposed because its edge is a product advantage, not yet a data or distribution moat — and every layer above has more leverage over the customer.

History and evolution

• 2020 — Founded in Palo Alto by Will Ross and William Steenbergen after a wildfire-modeling collaboration at Stanford; 1,200+ hours of underwriter interviews shape the product.
• 2021 — Early seed backing; Pear VC is the earliest disclosed investor.
• Sep 2022 — $15M Series A led by Emergence Capital; positions RiskOps as the first platform to unify underwriting and portfolio management.
• 2023 — Named to Insider’s list of hot fintechs; June 2023 $25M Series B led by Caffeinated Capital at a reported ~$125M valuation.
• Nov 2024 — $40M Series C led by StepStone Group (Emergence, Caffeinated, Pear following); total raised reaches $80M; company reports 90% faster quote times and revenue tripling.
• 2025 — Ships agentic AI features and a submission-to-quote workflow; Sep 2025 rival Cytora is acquired by Applied Systems; Nov 18, 2025 $100M Series D led by Growth Equity at Goldman Sachs Alternatives pushes total past $180M, with revenue reported tripling again and global expansion the stated use of funds.

No public crises, layoffs or down rounds — but it is young enough that the hard tests (multi-year net retention, surviving a soft market, defending against core-vendor AI) are still ahead.

What people say

The case for. On Glassdoor, Federato holds ~4.2 of 5 across ~32 reviews (2026), with roughly 85% recommending it — high for a company this early — citing fast growth, customer impact and learning. Customers give the strongest testimony: QBE’s pilot-to-300-underwriters rollout, Velocity Risk’s 89% faster quoting and 3.7x high-appetite bind rate, and Goldman’s diligence verdict of a “step change in ROI versus prior-generation core systems” (Nov 2025) are validation most insurtech never gets. The “designed by underwriters, for underwriters” framing, grounded in 1,200+ hours of interviews, shows up in reviewers calling it built for the workflow rather than bolted on.

The complaints. The negatives are specific and they matter. Glassdoor carries pointed criticism of an “always-on” culture — reviewers describe 60-80+ hour weeks treated as a badge of honor, a CEO praising staff for working at 1 a.m. on a Sunday, and at least one scathing “dumpster fire” review citing burnout; the split-camp pattern (excellent for some, “sweatshop” for others) is a real signal for a company that needs to scale headcount. More strategically damaging: one review alleges the internal plan is to partner with other companies, learn from them, then copy their approach in-house — corrosive to trust in an ecosystem where Federato depends on integrations. And an engineer’s review dismisses the “AI Risk Operating System” branding as marketing over substance, calling the architecture an uninspired carbon copy of Azure App Services — the thin-moat critique from the inside. Layer on the structural critique of every workflow-layer insurtech: it sits on cores it does not control, its biggest customers can build in-house, and the “next best action” it sells is exactly what a foundation-model agent or a Guidewire feature could absorb.

Outlook: the open question

Federato is a strong emerging company whose outcome turns on one question: does the underwriting decision-and-steering layer stay a durable, defensible category, or do the cores and carriers’ own AI agents absorb it? The bull case is well-evidenced: founders who learned the domain the hard way, measurable ROI, marquee expanding customers (QBE from 50 to 300+ seats), revenue reportedly tripling two years running, and capital — Emergence, then StepStone, then Goldman’s growth arm at $100M — that has underwritten the thesis repeatedly. RiskOps as a category, owning the underwriter’s daily surface between strategy and the core, is a good place to stand if it holds.

It works if Federato converts workflow into a moat — proprietary appetite and portfolio data that compounds, switching costs from being the daily surface across every line, and expansion economics (net retention well above 100%, seats and lines stacking inside each carrier) that make ripping it out unthinkable. It stalls if the layer proves absorbable — if Guidewire or Duck Creek push appetite, triage and next-best-action into the core the carrier already licenses, if foundation-model agents let a carrier’s own team wire governed underwriting flows in-house cheaply, or if the value collapses to a UI on someone else’s book. Cytora-into-Applied (Sep 2025) is the shape of the downside. Watch three things: whether net retention and multi-line expansion hold through a softening market; whether the cores ship competing appetite features; and whether Federato builds a data advantage the in-house team cannot cheaply replicate. At $180M+ raised, the runway is real — but so is the Series D pressure to prove this is a category, not a feature.

How a challenger would attack it

Sell the outcome, not the workbench. Federato’s price keys off premium volume and seat count, which means every new underwriter seat costs the carrier more — a pricing structure that invites an attacker to charge flat or outcome-based fees and frame Federato as a tax on headcount. The sharper attack exploits what Federato hasn’t built: a moat beyond product. Its own engineer-reviewer calls the architecture “an uninspired carbon copy of Azure App Services,” and the company discloses no net retention, no data advantage that compounds across carriers. A challenger built agent-first on frontier models — skipping the workbench UI entirely and delivering appetite scoring, triage, and drafted quotes as governed agents inside the carrier’s existing surfaces (email, the Guidewire screen, Teams) — turns Federato’s proudest asset, the daily desktop it took 1,200 interview hours to design, into legacy UI. The trust seam is exploitable too: a Glassdoor review alleges Federato partners with companies to learn and then copies their approach in-house — a challenger courting the same core vendors and MGA ecosystem can position itself as the safe integration partner. And the 60-80-hour “always-on” culture reviews describe is a recruiting pitch: hire Federato’s burned-out insurance-fluent engineers and take the domain knowledge with them.

Same playbook, new buyer

Take appetite-and-portfolio steering to books Federato ignores. Federato is tuned to US P&C specialty — aviation, trucking, cat-exposed property — sold to carriers and MGAs with enterprise SaaS motion and two-quarter rollouts. The same decision layer fits buyers that motion can’t reach: reinsurance treaty underwriting, where portfolio-context optimization is the entire job and the buyer count is small but premium is enormous; life and health underwriting, a different data regime the P&C-trained product doesn’t touch; and the London market beyond what Send covers, where syndicate capacity steering is native portfolio logic. The most interesting shift is down-market: small MGAs and program administrators writing $10-50M of premium can’t absorb a quoted enterprise contract or a two-quarter implementation, but a self-serve, priced-per-policy version of appetite guardrails would land in weeks. Federato won’t follow soon — its Series D economics, Goldman-underwritten growth story, and premium-plus-seats pricing all depend on large-account expansion (QBE going 50 to 300 seats), and rebuilding for self-serve or a new insurance domain means abandoning the ICP focus that got it here.

Sources and further reading

• Federato raises $100M Series D led by Goldman Sachs (Federato, November 2025)
• Federato Raises $100 Million Series D Led by Growth Equity at Goldman Sachs Alternatives (BusinessWire, November 18 2025)
• Federato fixes insurance risk analysis with AI, raises $40M (TechCrunch, November 2024)
• Federato announces $80 million raised to bring RiskOps to insurance (PR Newswire, November 2024)
• Federato, First Solution To Unify Insurance Underwriting And Portfolio Management, Announces $15M Series A (PR Newswire, September 2022)
• Meet the insurtech: Federato (Digital Insurance / Dig-In, founder background)
• Federato: Bringing the Insurance Underwriting Function Into The Future (Emergence Capital, investor thesis)
• Velocity Risk Reduces Time to Quote by 89% (Case Study) (Federato, customer case study)
• Federato Reviews on Glassdoor (Glassdoor, accessed July 2026)
• Kalepa raises $14M Series A led by Inspired Capital (Kalepa, 2024)
• Underwriting Software Market Size, Share, 2025-2030 Outlook (Mordor Intelligence, 2025)