Memory for Insurers

The current conditions of every product, for every channel that answers.

General conditions, coverage and claims procedures become a reviewed knowledge base, with an owner and a review deadline. Customer service, broker support and AI agents answer from the same version.

Product screen, with sample data

The problem

The conditions change. The answer does not always follow.

And a wrong answer becomes a complaint.

Several versions of the same conditions

Products are revised every year, and old conditions still apply to old policies. Finding the right version takes time.

Scattered claims procedures

Deadlines, documents and approval limits for each type of claim live in different manuals, emails and spreadsheets.

AI that invents cover

An AI without the insurer's knowledge base answers with what sounds right, without citing the clause.

Before and after

What changes at the insurer.

From the published conditions to the answer the policyholder gets.

Without the company's memoryThe right conditions are in one PDF among manyWith KlucoThe current conditions are in the knowledge base, with an owner and the date of their last verification
Without the company's memoryThe old version is still in useWith KlucoWhen the review deadline passes, the article is flagged as overdue
Without the company's memoryThe answer does not say where it came fromWith KlucoEvery answer shows the passages consulted and their verification date
Without the company's memoryA new question gets lost in customer serviceWith KlucoAn unanswered question goes into a queue and becomes an article
Without the company's memoryAI publishes what it wroteWith KlucoWhat AI writes can wait for the approval of whoever owns the product

What goes into memory

What an insurer keeps in memory.

Documents come in from Google Drive, OneDrive, SharePoint, file upload, API or webhook, and become articles that AI keeps up to date.

  • General and special conditions per product
  • Claims notification and adjustment procedures
  • Coverage, excesses and exclusions
  • Underwriting and acceptance rules
  • Policyholder service scripts
  • Endorsement and cancellation procedures
  • Broker support materials
  • Fraud prevention policy

Use cases

Where insurers start.

The most common use cases at insurers, with the part of the platform that does each one.

  • Procedures and rules in memory

    Internal rules and procedures live in one knowledge base, with an owner, a review deadline and a version history.

    Brain
  • A service agent that answers from the knowledge base

    The agent serves customers on WhatsApp and the other channels with what is in the reviewed knowledge base, and hands over to a person what it does not know.

    Agents
  • An unanswered question becomes an article

    What the knowledge base could not answer goes into a queue. Someone answers it once, and the answer becomes an article for everyone.

    Brain
  • AI proposes, a person approves

    When AI writes or updates an article, the change can wait for the approval of whoever owns the topic.

    Brain
  • One agent per area, with only its documents

    Each agent queries only the documents you choose, and the access profile sets who can use each agent.

    Agents
  • Personal data masked

    When the company turns it on, the agent masks personal data in its logs and answers.

    Agents

Brain · the memory

Every set of conditions with an owner, a deadline and a version.

The Brain turns conditions and procedures into articles with an owner, a review deadline and a version history. Every answer shows the passage and its verification date.

  • Human approval before publishing what AI writes
  • Contradictions between documents become review suggestions
  • A queue of unanswered questions
See the Brain
Product screen, with sample data

Agents · AI agents

Agents per product, with only that product's documents.

Create one agent for home insurance and another for motor, each limited to the documents you choose and with clear approval rules.

  • Knowledge limited to chosen documents
  • WhatsApp, Telegram, Slack, Microsoft Teams and Google Chat
  • Personal data masking, when turned on
See the Agents
Product screen, with sample data

LLM Management and Chat

AI for every team, with defined access.

Customer service, claims and underwriting use the models their profile allows, in the Chat, which draws on the insurer's memory.

  • Access profiles per group of people
  • A limit per person and a budget per team
  • Audit log for configuration and chat
See LLM Management
Product screen, with sample data

Control

What the insurer controls.

The company sets up each control below in the console.

See security and LGPD
  • An owner and a review deadline on every set of conditions, with a record of who verified it and when
  • A person's approval before publishing what AI writes
  • The documents each agent may query
  • Personal data masking in logs and answers, when turned on
  • A step-by-step history of every agent run

FAQ

What insurers ask.

How does AI know which version of the conditions to use?

The knowledge base keeps each document with an owner, a review deadline and a version history, and every answer shows the passage and its verification date. Versions that apply to older policies can stay in the knowledge base as separate articles, each with its own title.

Can I have one agent per product?

Yes. Each agent queries only the documents you choose, and the access profile sets who can use each agent.

Can the agent decide a claim?

That is not its role. Agents answer and organise based on the documents, and sensitive actions can require a person's approval. The decision stays with the insurer's team.

How does Kluco handle personal data?

With access profiles, data isolated per company, an audit log and the option to mask personal data in the logs and answers of each agent. The Security page covers each control.

Next step

See the current conditions reaching every channel.

In the demo, we show the conditions knowledge base, an agent per product and approval of what AI writes.