How Nairo Works

Nairo is organised around a simple principle: bring relevant insurance material into one workspace, use AI to support expert review, and keep humans responsible for important business decisions.

This page describes a typical way teams use Nairo across its main surfaces. It is not a rigid process. Each team can adapt the approach to its own governance, authority limits and review process.

The four-step workflow

1. Bring relevant material together

Before a case or document set can be reviewed, the relevant material needs to be available in the workspace.

In Nairo, material can come from:

• Library - uploaded documents and folders, such as guidelines, policy wordings and reference materials
• Projects - materials uploaded to a persistent document workspace
• Experts - reference materials and instructions configured for a specialist assistant
• Actions - repeatable tasks designed to produce a defined output

The objective is to make relevant material easier to access and use during review, not only to store files.

2. Review the material with AI assistance

Once the relevant material is available, users can work through the task using the Nairo surface that fits the job.

The main surfaces are:

• Assistant - open questions, document Q&A, comparisons and exploratory analysis
• Projects - persistent workspaces for document sets such as submission packages, policy wordings or claims files
• Experts - specialist review using configured instructions and selected reference materials
• Actions - repeatable tasks such as summaries, comparisons or checks
• Insight Table - structured review across multiple documents using columns and prompts
• Library - shared reference materials used across other surfaces

Where citations or references are available, outputs can help users understand which documents or materials informed the answer.

3. Apply configured context and review criteria

Insurance review depends on more than document text. Teams often need to apply appetite, policy language, internal guidance, review criteria and expert judgment.

Nairo supports this through:

• Library materials selected for a workspace or task
• Materials uploaded to a Project
• Expert instructions, personas, guardrails and team rules
• Column prompts and extraction fields in Insight Table
• Actions that encode repeatable review steps

The goal is to help users apply relevant workspace materials and review logic more consistently, rather than relying on generic AI responses.

4. Review and use the output

Nairo provides AI-generated assistance. Users remain responsible for reviewing outputs before using them in business workflows.

Depending on the surface, users may:

• Review answers from the Assistant
• Inspect cited material where citations are available
• Work with documents inside a Project
• Edit extracted values where the product supports editing
• Retry or re-run generation where the product allows it
• Use outputs to prepare follow-up work, reports, comparisons or internal analysis

The output can support tasks such as:

• Underwriting review
• Claims file review
• Policy comparison
• Broker communication
• Audit preparation
• Internal analysis

Nairo supports expert work. It does not replace expert judgment.

How the main surfaces fit together

Most teams do not use every surface on day one. A focused pilot usually starts with one use case and expands from there.

What the product supports today

Today, Nairo supports review through:

• Uploaded documents and workspace materials
• Assistant conversations over selected context
• Projects for persistent document-based work
• Library folders and documents attached to supported surfaces
• Configured Experts with personas, instructions and selected reference materials
• Actions that run defined multi-step tasks and produce deliverables for user review
• Insight Table reviews using column prompts across multiple documents
• Citations in supported Assistant and review surfaces
• Manual correction of certain extracted values where the product supports editing

The exact experience depends on the workspace configuration and the use case.

Product direction: decision memory

Nairo’s longer-term direction is to help insurance organisations preserve useful reasoning from their work.

Many insurance teams lose valuable context because reasoning lives in emails, documents, conversations or individual experience. Over time, Nairo aims to help teams capture useful review logic, corrections, decisions and rationale so that materials can be reused more systematically.

This is product direction, not a claim that all decision-memory capabilities are available today.

What this workflow is not

This workflow does not assume Nairo makes autonomous decisions.

It does not assume every output is automatically stored as institutional memory.

It does not replace internal approval processes, underwriting authority, claims authority or compliance sign-off.

It does not create a native precedent database unless those materials have been uploaded or configured in the workspace.

Nairo supports expert review. Governance remains with the organisation.

Getting started

A practical first setup usually involves:

1. Confirming the workspace and users are set up
2. Uploading key reference documents to the Library
3. Creating a first Project or choosing another starting surface, such as Insight Table or an Expert-led review
4. Defining who reviews outputs before they are used in business decisions

For setup and pilot planning, see Workspace Setup. For a practical first use case, see Underwriting Review.