Company

We build the AI workforce for regulated industries.

MarvelX builds agentic AI for financial services and highly regulated industries, starting with AI claims agents for insurance. Built in Amsterdam.

Amsterdam canal houses

Company

We build the AI workforce for regulated industries.

MarvelX builds agentic AI for financial services and highly regulated industries, starting with AI claims agents for insurance. Built in Amsterdam.

Amsterdam canal houses

Company

We build the AI workforce for regulated industries.

MarvelX builds agentic AI for financial services and highly regulated industries, starting with AI claims agents for insurance. Built in Amsterdam.

Amsterdam canal houses

Claims

0k+

claims processed per year

Policies

0M+

interactions with policies per year

Capacity

0x

increase in claims handling capacity

Why we exist

Insurance runs on documents that do not cooperate.

A handwritten intake form. Ten photographs taken in the rain. An invoice from a garage that uses its own template and always has. For years the only two answers were to hire more people to read them, or to buy software rigid enough to break the first time a form changed.

We took the third one. Agents that read a whole case the way an experienced handler reads it, score every field they extract, and hand back a decision with its reasoning attached. The work still belongs to your team. The reading, cross-checking and drafting no longer has to.

“AI has reached a tipping point. It can do much more than assist users in chat. It can take on full operational tasks. We are building the AI agents that can do that.”

Ali el Hassouni, Founder and CEO

How we work

Three commitments we do not trade away.

Your rules, not a model’s judgment

Agents are held to the policy wording and the claims manual your team already works from. The same case gives the same result, every time it runs.

Every conclusion cites its source

Coverage, fraud signals and settlement figures each point back to the document and the field they came from. An auditor can follow the trail without calling us.

Doubt goes to a person

Below your confidence threshold, a case routes to a handler with the reasoning attached. Human oversight is what this work requires, not a caveat.

Team

The people building it.

Eleven of us, between Amsterdam and New York. Engineers who sit in the client’s calls, and operators who know what a good claim file looks like.

Backed by Investors

Backed by security-first investors.

We have raised $6.7 Million in pre-seed funding to build the AI infrastructure for the insurance industry.

Led by

"MarvelX is building a next-generation vertical AI company in one of the most operationally complex industries. Their team combines deep AI expertise with a real understanding of the regulatory challenges insurers face."

Tom Mendoza, Partner at EQT Ventures

Joined by
Angels from

Mehdi Ghissassi

DeepMind & Google

Jobi George

Microsoft & Elastic

Keith Grose

Coinbase & A16Z

Gert Lanckriet

Amazon

Humphrey Valenbreder

Partior & bunq

Yana Abramova

Pretiosum Ventures

Lennard Zwart

Ascending AI

Backed by Investors

Backed by security-first investors.

We have raised $6.7 Million in pre-seed funding to build the AI infrastructure for the insurance industry.

Led by

"MarvelX is building a next-generation vertical AI company in one of the most operationally complex industries. Their team combines deep AI expertise with a real understanding of the regulatory challenges insurers face."

Tom Mendoza, Partner at EQT Ventures

Joined by
Angels from

Mehdi Ghissassi

DeepMind & Google

Jobi George

Microsoft & Elastic

Keith Grose

Coinbase & A16Z

Gert Lanckriet

Amazon

Humphrey Valenbreder

Partior & bunq

Yana Abramova

Pretiosum Ventures

Lennard Zwart

Ascending AI

Backed by Investors

Backed by security-first investors.

We have raised $6.7 Million in pre-seed funding to build the AI infrastructure for the insurance industry.

Led by

"MarvelX is building a next-generation vertical AI company in one of the most operationally complex industries. Their team combines deep AI expertise with a real understanding of the regulatory challenges insurers face."

Tom Mendoza, Partner at EQT Ventures

Joined by
Angels from

Mehdi Ghissassi

DeepMind & Google

Jobi George

Microsoft & Elastic

Keith Grose

Coinbase & A16Z

Gert Lanckriet

Amazon

Humphrey Valenbreder

Partior & bunq

Yana Abramova

Pretiosum Ventures

Lennard Zwart

Ascending AI

TRUST & SECURITY

Built for the strictest enterprise standards from day one.

Every time the Agent learns or makes a decision, it cites its sources. You can trace every output back to the specific document page, policy clause, or feedback instance that influenced it.

TRUST & SECURITY

Built for the strictest enterprise standards from day one.

Every time the Agent learns or makes a decision, it cites its sources. You can trace every output back to the specific document page, policy clause, or feedback instance that influenced it.

TRUST & SECURITY

Built for the strictest enterprise standards from day one.

Every time the Agent learns or makes a decision, it cites its sources. You can trace every output back to the specific document page, policy clause, or feedback instance that influenced it.

Where we are

Amsterdam and New York.

The team is in together at least three days a week. That is how a deployment gets from a client call on Monday to a working demo the same week.

Netherlands

Amsterdam

United States

New York

Careers

Open roles

See it run on your own documents.

Bring a real claim file. We will show you what the agent does with it, and what it refuses to decide on its own.

See it run on your own documents.

Bring a real claim file. We will show you what the agent does with it, and what it refuses to decide on its own.

See it run on your own documents.

Bring a real claim file. We will show you what the agent does with it, and what it refuses to decide on its own.

© Copyright 2026 MarvelX, Built in Amsterdam

© Copyright 2026 MarvelX, Built in Amsterdam

© Copyright 2026 MarvelX, Built in Amsterdam