Kalepa vs Sixfold (2026)
The same product idea with opposite answers on where control sits. Both take A on AI centrality and A on outcome evidence, and both read a commercial submission, score it against appetite and hand the underwriter a recommendation. Kalepa puts oversight in the execution path: independent judging agents validate critical fields, and Copilot reads and scores while the underwriter decides, which earns an A on autonomy and oversight. Sixfold puts it in governance: a staffed responsible AI function reviews new functionality quarterly and a NIST and EU AI Act anchored programme earns an A on model risk, but a configurable straight through path produces quote ready material with no described human checkpoint, so autonomy sits at B. The other split is data. Sixfold states in one sentence that each insurer's book is walled off and never trains another carrier's system. Kalepa says Copilot learns what the best underwriters do and never says whose.
- You want the control inside the workflow rather than beside it. Independent judging agents validate critical fields before an underwriter sees them, and the stated division of labour is that Copilot reads, cross references and scores while the underwriter makes the call.
- Time to value is the constraint. Copilot works on day one with no integrations, with inbox, portal and policy system connections added afterwards, which is the reverse of an implementation measured in quarters.
- You need to know the country of processing. Kalepa commits to United States servers, which is more than Sixfold states anywhere about where data sits.
- Your book is specialty and excess and surplus. Named users include Munich Re Specialty, Bowhead and SECURA across reinsurance specialty, excess and surplus and regional mutual business.
- Your data governance committee will ask whether your book trains a rival's model. Sixfold answers no in writing: each insurer's book is walled off and one carrier's data is never used to train another carrier's system.
- Model risk is a procurement gate. Sixfold runs a formal governance programme anchored to the NIST AI Risk Management Framework and cites the EU AI Act high risk provider articles 9 through 17 by number, the strongest A on that axis in this pocket.
- You write life and health as well as property and casualty. Sixfold serves both halves with dedicated product lines across more than 50 lines of business, and names bias testing tailored to each line as part of its programme.
- You want the scale evidence. Named carriers include Zurich North America, New York Life, Guardian, Generali Global Corporate and Commercial, AXIS and Skyward Specialty, with more than 1.5 million submissions processed since 2023.
This comparison is published by AI FinTech Index, an independent research platform that publishes independent ratings of AI vendors for financial services. Kalepa and Sixfold are each graded against the same capability taxonomy, from each vendor's own public materials and the regulatory record, under the AI FinTech Index verification standard. No vendor pays for placement, and no vendor has reviewed this page. How this evidence is graded
Plain facts
| Kalepa | Sixfold | |
|---|---|---|
| Primary category | Insurance AI | Insurance AI |
| Founded | 2018 | 2023 |
| Headquarters | New York, New York, United States | New York, New York, United States |
| Website | www.kalepa.com | www.sixfold.ai |
Side by Side
| Axis | K Kalepa |
S Sixfold |
|---|---|---|
| AI Centrality | ||
| Autonomy and Oversight Model | ||
| Model Risk Management and Transparency | ||
| Operational and Outcome Evidence | ||
| AI Safety and Data Stewardship | ||
| GLBA and Data Privacy Posture | ||
| Security Certifications and Trust Center | ||
| Regulatory Status and Licensure | ||
| AI Governance and Bias Disclosure | ||
| AI Liability and Recourse | ||
| Model Supply Chain Disclosure | ||
| Core Systems and Integration Depth | ||
| Deployment Model and Data Residency | ||
| Commercial Transparency | ||
| Institution and Segment Coverage |
The short version of each
Kalepa
Kalepa builds Copilot, an underwriting workbench for commercial and specialty property and casualty insurers that reads a submission as it arrives, cross references loss runs, web and third party data, surfaces exposures the submission left out and ranks risks by likelihood to bind and appetite fit. The AI FinTech Index records it at A on AI centrality, A on autonomy and oversight for independent judging agents and a stated rule that the underwriter decides, and A on outcome evidence for named users including Munich Re Specialty, Bowhead and SECURA. The index records the gaps: integration traded away for day one deployment, no model supplier named, an accuracy claim with no figure, and crime and legal action data used as exposure with no fairness testing published.
Source: AI FinTech Index, 2026
Sixfold
Sixfold builds underwriting AI for property and casualty and life and health carriers, and its AI Underwriter, launched June 2026, ingests a submission, flags what is missing, evaluates appetite and portfolio fit and returns a recommendation with its rationale. The AI FinTech Index records it at A on AI centrality, model risk management, data stewardship, segment coverage and outcome evidence, the last on named carriers including Zurich North America, New York Life, Guardian and AXIS. It is one of very few platforms in the index to state that one carrier's data never trains another carrier's system. The index records the gaps: no model supplier or data region disclosed, no error rate for recommendations, and a configurable straight through path with no described human checkpoint.
Source: AI FinTech Index, 2026
Common questions
Is Kalepa or Sixfold better for commercial underwriting?
They lead on different controls. Kalepa is stronger where the control sits inside the workflow, with independent judging agents validating critical fields and a stated rule that the underwriter decides, and it works on day one with no integrations. Sixfold is stronger on governance and data: a NIST and EU AI Act anchored model risk programme, bias testing by line of business, and a written commitment that one carrier's book never trains another carrier's system. Both hold A on AI centrality and on outcome evidence. Graded by AI FinTech Index against the same capability axes from each vendor's own published materials, verified September 21, 2026. No vendor pays for placement.
Does Sixfold or Kalepa use one insurer's data to train models for another?
Sixfold says no in writing: each insurer's book is walled off and one carrier's data is never used to train another carrier's system, which matters because its institutional memory feature feeds every underwriting decision back into the carrier's own deployment. Kalepa states that Copilot automatically learns what the best underwriters are doing and does not say whether that learning crosses carriers, so the question has to be settled in contract. Graded by AI FinTech Index against the same capability axes from each vendor's own published materials, verified September 21, 2026. No vendor pays for placement.
Where does the human underwriter sit in each product?
In Kalepa the human is in the execution path by design: Copilot reads, cross references and scores, judging agents validate critical fields, and the underwriter makes the decision. In Sixfold the human is protected mainly at the governance layer, through a staffed ethics and responsible AI function that reviews new functionality quarterly, while a configurable straight through path can produce quote ready and bind ready material without a described human checkpoint. Graded by AI FinTech Index against the same capability axes from each vendor's own published materials, verified September 21, 2026. No vendor pays for placement.
Do Kalepa and Sixfold cover life and health insurance?
Sixfold does, with a dedicated life and health product line alongside property and casualty and more than 50 lines of business in total, and it names algorithmic discrimination rules as bearing hardest on that side of its book. Kalepa is built for commercial and specialty property and casualty carriers and managing general agents. Graded by AI FinTech Index against the same capability axes from each vendor's own published materials, verified September 21, 2026. No vendor pays for placement.
How long does each take to deploy?
Kalepa is designed to work on day one with no integrations, adding inbox, portal and policy system connections afterwards. Sixfold also avoids replacing anything, running inside the underwriter's existing workbench, policy administration system and email, and Guidewire took a strategic position in its January 2026 Series B. Neither publishes pricing, so cost has to come from a sales conversation. Graded by AI FinTech Index against the same capability axes from each vendor's own published materials, verified September 21, 2026. No vendor pays for placement.
How does the AI FinTech Index grade Kalepa and Sixfold?
Both are graded on the same fifteen capability axes from public sources, each grade traceable to its artifact. The AI FinTech Index records Kalepa at A on autonomy and oversight and B on deployment residency for a United States server commitment, and records Sixfold at A on model risk management, data stewardship and segment coverage. Both hold A on AI centrality and outcome evidence and C on model supply chain and liability. The index publishes no composite score and declares no winner.
Related comparisons
Other published head to head assessments involving these vendors or their closest peers. The full set for this category is on the Insurance AI page.
Both are C on model supply chain. Sixfold reports more than 100 billion tokens processed, which makes large language models central, and names no provider, family, version or country of processing; Kalepa says its ensemble engine picks the right model per task and names none of them. Both are C on liability and recourse: neither publishes an accuracy or error rate for its recommendations, and Kalepa's market leading accuracy claim has no figure behind it. Each carries one sharper gap.
Kalepa surfaces criminal activity and pending legal actions from open web sources as underwriting exposures with no fairness testing published, so two identical risks in different postcodes may not screen alike. Sixfold places medical and health information into the pipeline through its life and health line while publishing no data processing agreement, subprocessor list or retention schedule, and it states no region where that data is held.