Introducing AI Blocks: The Building Blocks of Smarter Templates

Kyber's advantage when it comes to templates is built on a combination of full static template support, with Kyber-managed "AI Templates" for typical complex correspondence use cases like settlements and payment letters.

Static templates offer total control. Every word is locked, reviewed, and governed. They ensure consistency but can become brittle. 

AI templates changed that equation. They were the first of their kind in a claims-specific CCM, using Kyber’s AI to draft correspondence dynamically with relevant policy language and claim information. Adjusters could generate appropriate, compliant letters across hundreds of use cases without maintaining a massive library.

With that breakthrough, teams were freed from the operational weight that had always come with brittle templates.

Both approaches are essential. But between them sat a growing question from claims leaders:

What if we could manage our own AI, inside our own templates, with the same precision we already have for everything else?

That question became AI Blocks.

“AI Blocks live in the middle,” says Arvind Sontha, Kyber’s CEO. “You keep your template, add AI where it helps, and stay prescriptive about what it should do.”

Why We Built AI Blocks

Every carrier’s document library looks a little different. Policy wording evolves. States add new requirements. Teams refine workflows as regulations change.

Kyber’s AI Templates already handle that variation with advanced reasoning, but many operations leaders wanted finer control: the ability to direct how AI should behave inside a single letter. They wanted to build, test, and adjust the logic themselves.

AI Blocks were built for that middle ground.

They let claims teams automate the sections that need reasoning while keeping the rest completely static. It is a hybrid approach that feels familiar but moves faster.

What AI Blocks Are

An AI Block is a modular component inside Kyber’s template editor. Think of it as a “smart field” that understands instructions and performs one specific AI-driven task.

A few examples:

• Policy Language Block – Extracts and formats the exact clause you specify from the policy file.
• Estimate Table Block – Builds a clean payment table directly from the estimate file.
• Narrative Block – Summarizes claim notes into clear, human-readable language.

Each block operates inside your governance framework. You decide where AI belongs, what it should reference, and how the output is reviewed. When combined with the static fields teams already use, like jurisdictional clauses or data pulled automatically from the claim system, a new type of template emerges.

Kyber users have control where they need it and flexibility where they need it, not just across template types but within a single document.

How AI Blocks Work

When you open a template in Kyber, AI Blocks appear alongside the fields you already use.

1. Insert a block where dynamic content belongs.
2. Add a simple AI prompt, such as “Extract the Agreement clause."
3. Test the generated letter, to confirm the AI creates the expected output.
4. Refine and publish. Adjust the instruction, confirm the output, and make the block part of your standard template.

When the template is used, the document will use Explainable AI to describe why Kyber selected that language. Further, compliance teams can trace where text came from and why it was inserted with Kyber's built in audit-trail.

In practice, it feels like adding intelligence directly into the blueprint of your document.

Why It Matters

AI Blocks change who drives automation.

Before, only the Kyber team managed the logic behind AI templates for complex letters like Reservations of Rights, Payment Letters, Denials, Partial Denials, Status Letters, and DOI Complaint Responses. Carriers managed the library of static templates.

Now those capabilities converge. Admins can design automation, test it, and push it live without outside help.

For adjusters, this means fewer templates and faster drafting.

For managers, it means every output is explainable and logged.

For carriers, it means scaling automation safely across business lines.

“This marks a milestone in enabling claims teams with AI,” Arvind adds. “Business users can now prompt AI in the correspondence process while still maintaining organization guardrails for compliance and governance.”

What Comes Next

This first release introduces Policy Language Extraction as the initial block type. Next, the team is rolling out:

• Payment Table Blocks for estimate information.
• Narrative Blocks for dynamic claim summaries.

Each new type will get its own deep-dive article on askkyber.com/resources so you can see how it fits into the broader workflow.

The long-term goal is simple.

Kyber will give carriers a single system where every template can combine static content, structured parameters, and AI reasoning – all governed, all explainable, and all under your control.

AI Blocks are available now inside Kyber.

Log in to explore them in your template editor or book a demo to see how they can streamline your correspondence library.

“AI Blocks are the next step in Kyber’s mission,” says Arvind. “They let teams bring intelligence into their templates on their own terms, safely and confidently.”

‍