Kyber’s latest product update gives adjusters and reviewers deeper insight into how our policy reasoning engine works, while keeping the fast, accurate drafting experience they rely on.
With Explainable AI, adjusters can now hover over extracted policy language in the “Generate” view to see why it was included. It’s a simple way to understand the AI’s logic and get to a polished, compliant notice faster.
What Is Explainable AI?
Explainable AI makes Kyber’s logic transparent. For every section of policy language Kyber extracts into a notice, users can now view a short explanation of why it was chosen, based on the claim details and the policy at that point in time.
Previously, this process felt like magic. Kyber would pull in the exact right clause and place it into the draft instantly. Now it still feels like magic, but the adjuster has full visibility on how it works.
“You’re not just getting how the AI itself works or the actual AI output working, you’re also seeing exactly how it thinks.” — Arvind Sontha, Founder & CEO
How It Works
This feature appears automatically when you generate a letter in Kyber:
- Generate the letter as usual
- Kyber highlights the policy language it extracted
- Hover over any highlighted text to see an explanation of why that clause was included, based on the facts of the claim and the policy terms
For example, in a partial denial involving both water damage and wear and tear, Kyber may extract multiple sections from the policy. One might explain why payment for water damage was included, and another might justify the exclusion related to standard wear and tear.
Each explanation provides clear, concise reasoning so adjusters can immediately understand what the AI pulled and why, without needing to dig through the full policy.
Why We Built This
Kyber already ensures the right policy language is extracted and inserted with 0 hallucinations. Explainable AI takes it one step further by showing how that decision was made.
This matters because it allows adjusters and reviewers to:
- Save time reading through full policies for context
- Understand how the AI arrived at a decision before editing
- Quickly validate accuracy during the review process
- Confidently move through final approval without second-guessing
“It helps bridge that relationship between the adjuster generating the document and understanding Kyber’s AI." — Arvind Sontha
Whether you’re onboarding new team members, reviewing complex notices, or standardizing across distributed teams, Explainable AI helps you stay fast—without sacrificing clarity.
Built for Adjusters, Backed by Compliance
Explainable AI aligns with what Kyber already prioritizes: speed, consistency, and defensibility.
For claims teams, it adds an extra layer of defensibility. Every explanation is anchored in policy terms, pulled directly from the policy in force at the time of loss. There’s no guesswork, no generative language, and no hallucinated output.
This supports both auditability and adoption. Teams can trust what’s being generated, and adjusters can move faster because they understand the context from the start.
Available Now
Explainable AI is live in Kyber today.
To try it, generate a letter and hover over any highlighted policy language to see Kyber’s reasoning for the extraction. No additional setup is needed.
Want a quick walkthrough or best practices on how to use it with your team? Book a demo or reach out to your Kyber team.
%20(1).png)
.png)
