FuKazee
ContactTalk to us
Man signing paperwork at a desk beside a laptop

Hands-on AI training for insurance companies in Kenya for claims, underwriting and agents.

Your teams practise on the letters, policy wordings and complaint files they handle every week. A person still makes every claim and underwriting decision.
Plan your team's trainingWhatsApp us

7 hands-on modules. More than 3,000 people trained since 2022.

Some of the organisations we've trained and worked with

What should insurance staff in Kenya learn about AI first?

AI training for insurance companies in Kenya should start with the documents staff already write: claim letters, policy summaries, broker emails and complaint replies. FuKazee's sessions show claims handlers, underwriters, agents and customer care teams how to draft these faster with AI, then check each line against the actual policy.

Clarity is under pressure in this sector. In October 2025 the Insurance Regulatory Authority published a draft claims management guideline aimed at standardising claims handling and protecting consumers. We could not confirm its final status when we checked in November 2026, so ask your compliance team what applies today.

The agent network makes this urgent too. Reports of the Authority's data showed 14,560 licensed agents at the end of 2024, each explaining cover to customers in their own words. Good AI habits help them explain products clearly, using wording your company has approved.

Choose your starting point.

FuKazee serves teams, postgraduates and working professionals. Pick the one that sounds like you.
A corporate team in a training session
Organisations

I Want to Train My Team

Practical AI training for teams in Kenya, built on the work they already do.

Plan your team's training
A postgraduate student
Masters and PhD

I'm a Postgraduate Student

Use AI for research, writing and review. Live online, so you can join from Kenya.

ScholarShift Accelerator
Two professionals in conversation
Working professionals

I Want to Learn for My Job

One-to-one AI training, shaped around your role and your working day.

Ask on WhatsApp

The insurance work your people bring into the room.

We shape every exercise around one department's documents, so a claims handler is never stuck on a marketing task.

Claim and decline letters

Turn a decision already made by the claims officer into plain language that quotes the clause and explains the customer's next step.

Underwriting summaries

Reduce a long risk survey to a one-page note listing hazards and open questions. The underwriter sets the terms.

Agent and broker support

Explain a product in simple English and Kiswahili from approved wording, and prepare for the objections customers really raise.

Complaints handling

Build a timeline from a long email thread and draft a reply that answers each point the customer made.

Investigation file notes

Summarise masked motor claim files into facts and gaps. AI never labels anybody a fraudster, and investigators reach their own findings.

Board and management packs

Excel and Copilot for loss-ratio commentary and charts, with every figure traced back to your core system.

Try it: explain a declined motor claim clearly

The claims officer has already declined a claim under a named exclusion. The customer needs a letter they can understand, and the customer's details have been replaced with placeholders.

Try this prompt

You are a claims writer at a Kenyan general insurer. The claims officer has already decided this claim cannot be paid. Write a letter to [Customer A] in under 200 words of plain English. State the decision in the first two sentences. Quote this clause exactly as written: [paste clause]. Explain what the clause means for this claim using only these facts: [masked facts]. Tell the customer what they can send us if they disagree, and that they can escalate to our complaints desk and then to the Insurance Regulatory Authority. Do not add facts. Do not use the word 'regrettably'.

Before you use the answer: The claims manager compares the quoted clause with the issued policy and confirms the facts against the file. Compliance approves the escalation wording before any template is used.

What an insurer needs to know before booking.

Regulated teams usually ask about data and sign-off before they ask about content, so here are those answers.

  1. 01

    Teams

    Run claims, underwriting, customer care and distribution apart, then bring leaders together to agree shared rules.

  2. 02

    Customer data

    ID numbers, KRA PINs, number plates and bank details stay out of public tools. Medical cover files need extra care, because health data is sensitive under the Data Protection Act, 2019.

  3. 03

    Regulator

    We found no AI-specific rule from the Insurance Regulatory Authority when we checked in November 2026. Existing duties still apply, and your compliance officer has the final word.

  4. 04

    Tools

    Copilot for firms on Microsoft 365, plus the free version of other tools, compared side by side. We sell no software and recommend no single product.

  5. 05

    Delivery

    On-site at your head office or branch by arrangement, or live online for agents around the country.

The seven modules your team can draw from.

We adapt the examples to your work, and you can run any module on its own.
Module 1

Understanding AI

The groundwork for adopting AI responsibly across your organisation.

Module 2

Prompt Engineering

Role prompts, few-shot examples and clearer instructions for better output.

Module 3

Best Practices and Verification

Token awareness, output checks and prompt libraries that keep quality high.

Module 4

Navigating the AI Landscape

Hands-on with ChatGPT, Gemini, Claude, Perplexity, Copilot and others.

Module 5

Tool Comparison and Evaluation

A way to choose tools without falling for vendor hype or lock-in.

Module 6

AI Agents and Workflow Optimisation

AI for reports, communication, research, document analysis and workflow redesign.

Module 7

AI Ethics, Risk, and Responsible Use

Data leakage, hallucinations, bias, prompt injection and policy design for real organisations.

3,000+People trained since 2022
20+Organisations and institutions trained
7Hands-on modules

Get a recommended training plan for insurers, brokers and agencies.

Six quick questions about your sector, team and goals. You'll see a recommended programme straight away, and the people who run the sessions follow up with you personally.

  • A programme matched to your sector and team size
  • An AI usage policy included if you need one
  • Takes about a minute. No obligation.
Step 1 of 7
What kind of organisation is this for?

Read more on this.

Guides from the FuKazee team, written to be useful even if you hire someone else.
Corporate AI17 min read

AI for SACCOs and Insurers in Kenya: How to Use It Safely

How Kenyan SACCOs and insurers can use AI safely: member and claims data, IRA, SASRA and ODPC duties, sample prompts and a decision table for your team.

Read article
Corporate AI15 min read

How Should Customer Service Teams Be Trained on AI?

How to train customer service staff on AI: WhatsApp replies, tone, Kiswahili and Sheng, when to escalate, sample prompts and a 30-day plan for Kenyan teams.

Read article
Corporate AI17 min read

Can Staff Use ChatGPT Under Kenya's Data Protection Act?

What Kenya's Data Protection Act means when staff at a private company use AI: vendor contracts, data leaving Kenya, HR risks, a staff rule and a checklist.

Read article

Questions about aI training for insurance companies in Kenya.

Short answers. Anything else, ask us on WhatsApp and a person replies.

Plan your team's trainingAsk on WhatsApp

Can AI decide whether to pay a claim?

Our training keeps that decision with a person. AI can organise the file and draft the letter once the outcome is settled. Someone accountable has to weigh the facts and the policy.

Can insurance agents use ChatGPT to explain products?

Yes, when they work from approved product wording and leave out client details. An agent remains responsible for what the customer is told, so a confident but wrong AI answer becomes the agent's problem.

Does a mistake in an AI-drafted letter matter legally?

It can. A wrong clause or promise sent to a customer may cause a dispute. Your legal and compliance teams should approve templates, and this page is not legal advice.

Is this suitable for brokers and small agencies?

Yes. A small agency can gain a lot from clearer renewal reminders, comparisons of cover written for customers and better follow-up, mostly using free tools.

Do the exercises use our real claims?

They do not need to. We use invented or fully masked files that look like yours, so staff can learn the method without any customer's information entering a tool.

Will AI replace claims handlers and underwriters?

Some drafting and sorting will shrink. Judgement, customer conversations and accountability remain with people, and nobody becomes an expert in a few days, so we plan for steady practice after the session.

What affects the cost for an insurance company?

The number of departments, staff, branches and days. You get a quote after we have talked through your departments. Our guide to what corporate AI training costs in Kenya sets out what to budget for.

Keep exploring.

FuKazee is based in Westlands, Nairobi. This page was last reviewed on 10 November 2026.

Related training

By industry

Ready to start with insurers, brokers and agencies?

Tell us who needs training and what you want them to be able to do. The people who run the sessions reply to you themselves.

Plan your team's trainingTalk on WhatsApp