Insurers piloting agentic AI
- Multiple insurers and vendors are piloting agentic AI in underwriting, submission automation, and claims detection. - Examples include CFC Insurance underwriting-in-seconds pilots, Birlasoft automating broker submissions, and Novo AI tackling fraud and waste in claims. - These pilots show operational traction for automation in underwriting pipelines and claims triage, with vendor case studies emerging on execution (x.com) (x.com) (x.com).
Insurance carriers are moving agentic artificial intelligence from demos into live underwriting and claims work, with pilots now handling real submissions and claim documents. (cfc.com) In insurance, underwriting is the step where a carrier decides whether to quote a risk and at what price. CFC said on April 21, 2026 that its new “Lane Assist” pilot takes a submission from email to a quote recommendation “in seconds” and is already live inside its cyber underwriting team. (cfc.com) CFC said the pilot is reviewing a small number of real new-business submissions in live workflows and producing real quotes. The company said every recommendation is still reviewed, checked, and approved by a human underwriter before anything is issued. (cfc.com) A second bottleneck sits earlier in the pipeline: broker submissions arrive as emails, forms, and attachments that staff often rekey by hand. Birlasoft says its commercial-lines submission automation system pulls data from ACORD forms and other documents, applies rules for pre-clearance, and routes cases through a dashboard for underwriters. (birlasoft.com) Birlasoft’s published case study says one global insurer using that system doubled its quote-to-submission ratio from 6% to 12% and doubled its bind-to-submission ratio from 3% to 6%. The company says the same deployment cut manual work and sped broker submission clearances. (birlasoft.com) Claims is the other major target, because adjusters spend hours reading bills, reports, and correspondence before deciding what needs attention first. Novo AI says its platform extracts and analyzes claim documents, screens 100% of documents for potential fraud and abuse, and uses those reviews to reduce claim leakage and improve loss ratios. (heynovo.ai) Birlasoft is pitching the same document-intake logic beyond underwriting. Its insurance intake product says carriers can automate first notice of loss, the first report that starts a claim, and reduce settlement time and costs by processing structured and unstructured documents around the clock. (birlasoft.com) The common pattern is not full autonomy but supervised automation. CFC says its pilot follows established underwriting rules and keeps underwriters in the approval loop, while Birlasoft and Novo AI frame their tools as systems for intake, screening, and prioritization rather than final authority. (cfc.com) (birlasoft.com) (heynovo.ai) That leaves the next test in execution: whether insurers can turn faster intake and better triage into broader rollout without losing underwriting judgment or claims controls. For now, the clearest evidence is that these tools are already being tried on live insurance work, not just shown in slide decks. (cfc.com) (birlasoft.com) (heynovo.ai)
