Insurance · Deep dive

Akur8

Paris-headquartered actuarial AI platform building transparent GLM/GAM pricing and reserving software for P&C and now life insurers — 330+ carriers in 40+ countries, $180M raised through a $120M Series C led by One Peak in September 2024, expanded via the Arius (Milliman) reserving buyout in September 2024 and the Slope Software life-actuarial acquisition in March 2026.

emerging

The question that decides it: Akur8 depends on core-system distribution — Guidewire is both a strategic investor and, via its own pricing engine and marketplace, its most credible bundler — while vertical AI-native attackers (Sixfold on underwriting triage, Federato on portfolio steering, hyperexponential on specialty pricing) each carve off a slice of the same actuarial workflow. Falsifiable test: does Akur8's net customer count keep growing through 2027 — clearing 400 logos on organic wins not tied to Arius or Slope customer bases it just bought — or does it plateau in the 330-360 range as Guidewire, Duck Creek and WTW ship or bundle AI-native pricing inside systems the insurer already owns?

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

Snapshot

Akur8 is a Paris-headquartered actuarial-AI platform selling ML-powered pricing and reserving software to P&C insurers and — since March 2026 — life and annuity carriers. It claims 330+ carriers in 40+ countries (AXA, Generali, Munich Re, MAPFRE, HDI, Tokio Marine, MS&AD). Its “transparent” ML engine automates GLM/GAM models actuaries otherwise hand-build in WTW Emblem, preserving coefficient-level explainability regulators require. On Sep 16, 2024 it closed a $120M Series C led by One Peak, with Partners Group and Guidewire following on — total ~$180M. Two acquisitions bracket the raise: Arius reserving (Milliman, Sep 2024) and Slope Software (life actuarial modelling, Mar 2026). ARR: $10.2M (2023) → $20.5M (2024) → 50%+ growth in 2025 — respectable but light for $180M cap stack.

Founding story

Akur8 was incubated inside a French insurance carrier where Guillaume Beraud-Sudreau led actuaries and data scientists trying to break WTW Emblem’s grip on GLM modelling. The bottleneck was throughput: moving a rate change from actuarial notebook to Radar Live production could burn quarters. The group built an ML routine that automated variable-selection and coefficient-grouping while keeping the output expressible as a GLM/GAM regulators and the Chief Actuary could still audit line by line.

Samuel Falmagne came in as CEO in 2018 — 15 years at IBM international sales, then sales lead at Shift Technology, which gave him the chief-actuary/CDO buying network at global carriers (Finance Innovation, 2020). Anne-Laure Klein joined as COO from Carrefour’s digital-and-data partnerships. That triad is why Akur8’s early customer list skewed AXA/Generali/Munich Re rather than the mid-market MGA world most European insurtechs settle for.

How it works

The Akur8 engine is a supervised-learning routine that produces GLMs and GAMs — the model classes actuaries have priced insurance on for forty years — but automates variable selection and the grouping of continuous variables into piecewise coefficients. Output is a fully-decomposed additive model: every prediction traces to the effect of each input, with visible coefficients. That is the point. Regulators expect actuaries to defend rating factors line by line; a GBM that is 2% more accurate but not decomposable fails a rate filing at the Colorado DOI or a Solvency II review. Akur8’s whitepaper argues Transparent ML gets most GBM accuracy inside a model shape regulators will approve.

Around it sit five surfaces: Risk (pure-premium), Demand (elasticity/conversion), Rate (candidate rate structures), Optim (rate optimization, only where legal), and Deploy (rating engine pushing models into production, incl. a Guidewire PolicyCenter Marketplace app that auto-exports into PolicyCenter’s Rating Engine — Guidewire IR, Nov 2023). Post-acquisition, Reserving (Arius) covers loss triangles/IBNR and Life (Slope) covers cash-flow modelling for life and annuity carriers.

Product and business overview

Akur8 sells three SKUs: Pricing (Risk, Demand, Rate, Optim, Deploy), Reserving (Arius), Life (Slope, March 2026) — all cloud SaaS. Customer press releases (Alpha Insure, 2024) describe weeks-long deployment, not the multi-month Radar/Earnix rollouts. Guidewire Marketplace is the flagship channel for Guidewire Cloud carriers; Milliman — reinforced by its Arius-sale equity — is the reserving channel.

Business model and pricing

Akur8 does not publish list pricing. RFP.wiki (2026) notes buyers must engage sales; opacity complicates RFP cost scoring. Deals are annual SaaS, priced per module, per line and per user tier — Radar-style pricebook in a SaaS wrapper — with a services attach on enterprise deals. Deploy is the closest thing to a usage lever; Optim is gated by regulatory geography. 2025 disclosures of 50%+ ARR growth on 50+ new logos (25+ North America) imply average ACV in the low six figures — consistent with actuarial-software norms, well below Guidewire InsuranceSuite ACVs. That ceiling is the strategic problem: pricing is a tool line item inside a policy-admin budget incumbents already own.

Traction over time

Two things stand out. Arius brought 150+ reserving clients and Slope brought tier-one life/reinsurer relationships, so a meaningful share of the 330-logo count is inorganic. And ~$20M ARR on ~$180M raised is capital-efficiency territory investors will price harder as growth-equity normalizes.

Market analysis

The narrow addressable market is small. datainsightsmarket sizes global Actuarial Software for Insurance Pricing at ~$860M in 2025, 6.6% CAGR; broader actuarial modelling software is low-single-digit billions (MarketsandMarkets). The wider P&C insurance software market is $24.85B in 2026 → $47.25B by 2035 (Market Research Future), but almost all of that is core policy/claims administration, not pricing. Akur8’s realistic serviceable market is chief-actuary and pricing budgets at 2,000-3,000 mid-and-large P&C carriers globally, plus life via Slope and reserving via Arius. Tailwinds: regulatory push on explainability (NAIC, EIOPA), personal-lines loss-ratio pressure post-2022 inflation, cloud migration of core systems finally making rate-deployment automation viable. Against it: consolidation into Guidewire, Duck Creek, Sapiens and Majesco, all bundling rating and analytics.

Competitive intel

WTW Emblem/Radar is the incumbent every RFP starts with; switching cost is two decades of actuary training. Earnix ($150M+, Insight Partners) is the direct SaaS peer — broader in real-time personalization, ranked by RFP.wiki as top same-category alternative. hyperexponential (~$91M) is partner-shaped for now (hx on Lloyd’s/specialty underwriting, Akur8 on actuarial modelling) — converges if either pushes onto the other’s ground. Guidewire is investor-and-frenemy: its PolicyCenter Rating Engine directly overlaps Deploy and Rate; every native feature shrinks the wedge. Duck Creek (Vista, since 2023) is the same threat from the other core system. Verisk/ISO is the pre-built rate-plan substitute for smaller US carriers. Milliman/WTW/EY/Deloitte control adjacent reserving and outsourced-actuarial spend. Vertical AI attackers — Sixfold ($30M Series B Jan 2026, Guidewire strategic), Federato ($182M total, $100M Series D Nov 2025), Novee — each carve off a slice of underwriting/portfolio/rating.

History and evolution

Mar 2018: Founded in Paris. Feb 2020: €8M Series A. Jun 2021: $30M Series B; US expansion. Nov 2023: Guidewire strategic investment plus Marketplace app auto-exporting Akur8 models into PolicyCenter. Sep 9, 2024: Acquires Arius from Milliman; Milliman takes equity. Sep 16, 2024: $120M Series C led by One Peak; ~$180M total. 2025: Madrid office; 50%+ ARR growth; 50+ new logos incl. 25+ North America. Mar 17, 2026: Acquires Slope Software (Atlanta); enters life and annuity as “Akur8 Life.” No public restructurings or executive exits through September 2026, though Glassdoor cites constant attrition.

What people say

The case for. Enterprise carrier press releases (AXA, Generali, Munich Re, MAPFRE) cite three benefits: model build time cut ~10x vs manual GLM workflows; regulator-defensible transparency because output stays an additive GLM/GAM; and Guidewire Marketplace deployment collapsing the IT bottleneck between actuarial and production rating. Celent (2023-2024) calls Akur8 the reference automated-GLM/GAM vendor. Guidewire investing twice, hosting a Marketplace app, and following on at Series C is the strongest third-party validation. Glassdoor: 4.7/5, 91% recommend (2026).

The complaints. Independent review coverage is thin — a red flag itself. G2 has zero reviews; Capterra, Gartner Peer Insights, TrustRadius and Software Advice have no meaningful review base (RFP.wiki, 2026), so procurement gets no peer-verified signal. On Glassdoor, alongside the 4.7, individual reviews cite extractive high-pressure culture, constant background attrition (people disappearing without explanation), product leaders with domain depth but limited SaaS scale experience, and interview-process ghosting. Actuarial-community discussions cluster around: (1) transparent-ML is still less predictive on complex non-linear risks than modern GBMs, forcing sophisticated shops to keep a shadow model; (2) Optim’s regulatory gating means the module’s value depends on where the customer writes business; (3) opaque commercial terms and per-module bundling create RFP friction versus WTW’s established pricebook. The strategic complaint an investor should not ignore: ~$20M ARR on ~$180M raised is capital-intensive, and a chunk of the 330-logo count is inherited from Arius.

Outlook: the open question

For Akur8 to justify the September 2024 Series C, three things must be true through 2027. First, organic logo count keeps expanding past 400 excluding Arius and Slope customers — inorganic disclosure is the easiest way a SaaS story hides a slowing new-logo motion. Second, Guidewire and Duck Creek native pricing modules do not become “good enough” for the mid-market wedge Akur8 depends on; if either ships an in-suite engine clearing the transparent-model bar, Akur8 becomes optional at exactly the accounts it needs to convert. Third, the life-and-annuity bet via Slope produces a genuine second growth vector — Slope is much smaller than Arius, and life actuarial modelling is already crowded (Milliman MG-ALFA, Moody’s AXIS, FIS Prophet, RGA Meridian). If any fails, Akur8 becomes a strategic asset Guidewire, Duck Creek, WTW or Verisk buys at a multiple that clears investors but does not build the standalone franchise the pitch requires. Downside is not a wipeout — an OK acquisition at 4-6x forward ARR.

How to attack it

The most attackable seam is commercial structure. Akur8 sells opaque, per-module, per-line SaaS to chief actuaries. An attacker should sell the same category on loss-ratio-improvement bounty pricing: managed pricing-and-reserving-as-a-service, monthly fee tied to measured combined-ratio improvement, vendor data scientists embedded in the carrier. Milliman/WTW cannot ship that without cannibalizing consulting hours, Guidewire cannot without unwinding license economics, and Akur8 cannot counter without abandoning its per-module pricebook.

The second seam is regulatory-geography gating. Optim, the highest-margin module, only sells where price optimization is legal. An attacker in the 20+ US jurisdictions where it is restricted could ship a fairness-first pricing engine satisfying NAIC Model Bulletin on AI plus state disparate-impact rules by construction — turning a compliance blocker into a wedge.

The third seam is the transparency-vs-accuracy tradeoff. Akur8’s own whitepaper concedes GBMs are more predictive on complex non-linear risks; sophisticated actuaries already run a shadow GBM. A vendor shipping regulator-defensible wrapping around GBMs — SHAP + monotonicity + model-card tooling that gets a black-box model through a rate filing — attacks the technical premise directly.

Enumerated weaknesses: no peer reviews on G2/Capterra/Gartner (RFP.wiki, 2026) leaves enterprise buyers without external validation; Guidewire dependency is asymmetric (hosts and invests but competes); Optim’s regulatory gating caps the highest-value module; Glassdoor attrition (2026) suggests roadmap execution risk; ~$20M ARR on ~$180M raised limits price flexibility in downturn RFPs; and much of the 330-logo count came via Arius and Slope, not organic sales.

Adjacent-segment play

The most natural adjacency is reinsurance pricing and portfolio structuring. Reinsurers use the same machinery (severity/frequency GLMs, treaty-layer optimization) but with cash-flow horizons and layered contract terms neither Akur8 Pricing nor Arius natively covers. A vendor adapting transparent-ML for treaty pricing and cession optimization would sell to the ~150 global reinsurers Munich Re, Swiss Re, SCOR and Hannover Re anchor — smaller logo count than P&C direct, ACVs 5-10x higher. Munich Re is already an Akur8 customer, validating demand but capping greenfield.

The second adjacency is captive insurance and self-insured risk pools — corporates and public entities self-insuring via captives (Marsh, Aon, WTW captive) that today buy actuarial services rather than software. A cloud-native pricing/reserving product priced monthly, without a six-figure services attach, could open a segment traditional vendors deemed too small.

A third is US health insurance actuarial modelling (Milliman MedInsight, OptumIQ) — mirrors P&C in shape under different regulation; Slope’s cash-flow chassis is a natural base. Where the wedge does not generalise: usage-based consumer pricing (Root, Metromile), broker quoting (Bold Penguin, Vertafore) and claims analytics (Tractable, CCC) run on different data foundations dominated by incumbents.

Sources and further reading

• Akur8 Secures $120 Million in Series C Round — Akur8, September 16, 2024
• Parisian Akur8 secures $120M Series C round for actuarial platform — Tech.eu, September 16, 2024
• One Peak Leads $120 Million Series C Round in Akur8 — One Peak, September 2024
• Akur8 Expands Core Capabilities with Strategic Acquisition of Arius — Akur8, September 9, 2024
• Akur8 snaps up Arius reserving software tool from Milliman — InsuranceERM, September 2024
• Akur8 Enters Life and Annuity Insurance Market with Strategic Acquisition of Slope Software — GlobeNewswire, March 17, 2026
• Akur8 Raises $30M in Series B Funding — Finsmes, June 2021
• Automate Pricing Processes with Akur8’s New Guidewire Marketplace App — Guidewire, November 13, 2023
• The truth about GAMs, GLMs and GBMs in insurance pricing — Akur8 whitepaper (via Celent), 2024
• Akur8 Cost Drivers: Pricing & Contracts (2026) — RFP.wiki, 2026
• AKUR8 Revenue 2024: $20.5M ARR — Latka, 2024
• InsurTech firm Sixfold secures $30m to advance AI underwriting — FinTech Global, January 30, 2026