Insurance SaaS

From a product idea to a complete SaaS for the insurance industry

Area
Digital products
Customer
ClaimMentor

The client came with a product idea, UX designs and part of the technical solution. Our original task was to build the product.

During delivery, however, we took responsibility for the whole — from architecture and security through the user flow to running costs.

The result is a multi-tenant SaaS that connects meeting scheduling, communication with participants, video calls, recordings, transcripts and AI summaries into one process.

Share this case study
Result~USD 150 / monthestimated running costs of the resulting cloud solution
Scope1 end-to-end workflowfrom scheduling a meeting to the recording, transcript and summary
ResultDelivered on timethe product was built as one whole, not as a set of isolated features

Starting point

The client had a vision. It needed to become a product.

A start-up in the insurance segment had a clear idea of the problem it wanted to solve, first UX designs, its own multi-tenant sign-in built on AWS Cognito and an idea of the integrations it would need.

The individual parts, however, did not yet form one consistent product ready to be offered to financial institutions.

Idealine originally joined the project as a development team.

As the work went on, we began to address things that were formally outside the development brief — because they could fundamentally affect the security, operation or usability of the final product.

Our responsibility

Not implementing a brief. Delivering a working whole.

When we see a decision during delivery that could put the product at risk, saying “that was not in the brief” is not enough for us.

We look at the product end-to-end.

For ClaimMentor, that meant bringing product analysis, architecture, development, integrations, security, UX and running costs together into one solution.

01

Security

We addressed a risk even though it was not part of the brief

The initial solution included custom multi-tenant sign-in built on AWS Cognito.

During analysis we identified risks in how individual customer organisations were separated. We described the problem, proposed changes and, together with the client, included them in further development.

It was not a feature visible on screen.

It was, however, something that could decide whether the product could be operated safely for multiple financial institutions.

Responsibility for the product does not end at the boundary of the original backlog.

02

Running costs

Instead of an expensive service, we looked for a simpler way

One of the key parts of the product was scheduling meetings between the claims adjuster and other participants.

The first option relied on external services which, together with the rest of the cloud infrastructure, would have meant running costs of roughly USD 2,000–4,000 a month.

For a start-up, that difference is fundamental.

So we looked at the problem again and designed our own scheduling component built specifically for the needs of the product.

The resulting cloud architecture has estimated running costs of roughly USD 150 a month.

We did not optimise the infrastructure just to make it technically elegant.

We optimised it so that it made sense for the business that is meant to grow on it.

03

Product

Individual features became one process

A calendar, e-mail, SMS, video calls, recordings or an AI transcript do not make a product on their own.

What mattered was connecting them into one process the user understands.

The claims adjuster therefore works with a single insurance claim, in which they can:

  • schedule a meeting
  • coordinate multiple participants
  • send invitations and notifications
  • hold a video call
  • work with its recording
  • view the full transcript
  • quickly go through the AI summary

The individual technologies stay in the background.

The user sees one product.

What we delivered

One team from design to deployment

Product and analysis

Domain model, user scenarios and the scope of the individual parts of the solution.

Architecture

Multi-tenant SaaS architecture and the design of operations in AWS.

Development

Backend, frontend and the business logic of the product.

Integrations

Calendars, e-mail and SMS services, video calls and other external systems.

AI

Converting recordings to text and automatically summarising the content of meetings.

Security

Analysis of identity and the separation of customer organisations.

Testing and delivery

Continuous verification of functional parts and delivery of the final product.

Impact

The client got a product, not just an implemented brief

The result is not a handful of connected integrations.

It is one consistent SaaS product in which the claims adjuster has the entire course of a remote meeting in the context of a specific insurance claim.

The client also gained:

  • a solution ready for multiple customer organisations
  • a safer multi-tenant architecture
  • significantly lower expected running costs
  • a unified user process
  • an architecture ready for further development

And above all, a team that took responsibility not only for the code, but for the resulting solution.

What this case study shows

The outcome is the goal. Technology is only the means.

The client does not need to know every technical decision before development starts.

They need to know what outcome they want to achieve.

Our role is to take responsibility for the path between those two points — to point out risks, look for simpler solutions and make decisions in the context of the whole product.

Technologies and approaches

  • TypeScript
  • Node.js
  • AWS Lambda
  • API Gateway
  • DynamoDB
  • S3
  • CloudFront
  • AWS Cognito
  • Zoom
  • Google Calendar
  • Microsoft Graph
  • Speech-to-text
  • AI summarisation
  • CI/CD
Share this case study

Do you have a product idea, but not yet a complete technical brief?

You do not need to resolve the architecture, integrations or every screen in advance.

Tell us who the product is meant to help and what change it should bring. Together we will define the first meaningful whole and propose a path to building it.

Discuss your product
Discuss your product