Cytora harnesses the power of LLMs for next generation risk and underwriting intelligence 

Cytora, the AI-powered digital risk processing platform, today unveiled its latest enhancement, leveraging Large Language Models (LLMs) to bring a new level of accuracy and efficiency to risk assessment and underwriting processes.

The availability of LLMs alongside Cytora’s proprietary AI transforms digitisation performance across end-to-end underwriting workflows. The Cytora platform is deployed with industry-leading insurers like Allianz, Arch, Beazley and Markel International.

The use of LLMs in the Cytora platform uplifts performance through a host of new features that enable commercial insurers to accelerate digitisation by reducing the time to onboard new lines of business and lifting automation levels by enabling fields that have traditionally required interpretation by risk professionals to be easily standardised through an AI-powered approval workflow. These features enable insurers to digitise the full prebind workflow, provide more consistency on decisioning across the workflow and further uplift GWP / FTE.

New capabilities include:

1. Field builder enabling risk professionals to add new fields, building upon pretrained LoB templates across all lines of business in Commercial P&C and Specialty, directly from their configuration console.
2. Field management enables risk professionals to standardise heterodox fields through approval workflow enabling insurers to easily identify, manage and onboard new declarations
3. Chain of (risk) thought executes consistent multi-step reasoning turning hard to fathom submission data into standardised customer data

Field builder accelerates line of business onboarding, business users can quickly and easily add new fields on top of pretrained line of business templates directly from the Cytora console. This accelerates line of business configuration and reduces the time and effort to scale digitisation across diverse lines of business.

Field management standardises heterogeneous fields that historically have required lengthy interpretation by risk professionals. Cytora auto-matches heterodox fields to standardised lists based on similarity to unlock streamlined risk appetite, selection and pricing decisions. Customers receive automation lift in lines of business where heterogeneous fields – including declarations, business activity and property usage descriptions – are common.

Chain of thought interprets risk data enabling fields to be pieced together from heterogeneous data. Commercial insurers receive risk submissions using many different questions and are not in control of the data they receive. The heterogeneity of submissions requires multiple reasoning steps to be carried out before the right field is chosen to fulfil the insurers’ target schema. Chain of Thought thinks like a risk professional and consistently executes multi-step reasoning chains to power complex interpretation decisions, uplifting digitisation accuracy.

Richard Hartley, CEO of Cytora, said: “We are excited to launch this latest enhancement to our platform. By providing risk specific LLMs that build upon our proprietary AI, Cytora enables underwriters to cut through the noise and reach decisions more efficiently and effectively.

“In an industry where data is key to driving decision making, these cutting-edge features are an important step in our mission to empower insurers with the tools they need to better manage risk in an ever-evolving landscape.

“Our aim is to allow insurers to optimise their operations and stay ahead of the curve, and by setting a new standard in the industry for data extraction, I am delighted that Cytora continues to do this.”