HomeCompaniesVeltha
Veltha

AI native TPA starting with workers comp

Veltha is the AI native TPA starting with workers comp. We handle claim for insurance providers end to end. Learn more at https://veltha.ai
Active Founders
Edin Citaku
Edin Citaku
Founder/CEO
Edin is CEO and co-founder of Veltha, building the AI adjuster for regulated insurance. Prior to Veltha he led Palantir's AI deployment at the german FBI. He graduated with BSc and MSc in computer science at the TU Munich
Roland Jaloszynski
Roland Jaloszynski
Founder/CTO
Roland is a co-founder of Veltha, an AI-native TPA. He was previously a technical lead on mortgage underwriting automation at Palantir for a $642B servicer, compiling agency guidebooks into executable rules. He earned his computing degree from Poznan University of Technology while working as a software engineer at Amazon, LocalBini and DAC.digital.
Company Launches
Veltha: AI claims adjuster for regulated insurance
See original launch post

TL;DR: Veltha reads an entire claim file, flags what's missing, and applies the exact federal regulation that governs it, producing a citation-backed determination a human adjuster just signs off on. Work that took an adjuster ~6 hours now takes ~10 minutes. Starting with workers' comp and crop insurance.

https://www.loom.com/share/9e9701f61bf4400890b4799fdda3d3ff

The problem

1 in 4 claims adjusters expect to retire within five years. Experienced adjusters are retiring faster than carriers can replace them, and processing capacity is now the binding constraint on growth.

What we built

Veltha is an AI claims adjuster that runs the whole file end to end:

  • Triage - reads incoming claim emails and documents, classifies them, and routes each claim to the right workflow.
  • Gap detection - checks the file for missing documents and records, then requests them by email, phone, or fax and runs follow-up cadences until they land.
  • Compliance review - applies the relevant federal rulebook, and produces a determination where every output element is tied to a source span and a specific regulatory provision.

The adjuster's job shifts from doing the claim to reviewing Veltha's output. Less chasing. Faster decisions. Output that survives an audit.

What we understand that others don't: regulation is the moat, not the obstacle. Horizontal platforms (Guidewire, Duck Creek) and generic AI-claims startups can summarize a claim file. They can't tell you which provision triggered which adjustment, with citations. We treat the federal rulebook as an anchor for the workflow we embed within the organisation. That's the difference between a summary a human re-does from scratch and an output an adjuster will actually sign.

Our ask

We'd love intros to:

  • Anyone in claims at a carrier, MGA, or TPA - especially workers' comp
  • People at large self-insured employers with claims-heavy, high-risk workforces - Companies exactly like these are who we're built for: Disney, Marriott, Ford, PG&E, Caterpillar. If you know someone at that kind of company, an intro is gold.

Reach us at edin@veltha.ai or ping us here. Thanks YC 🙏

veltha.ai

Veltha
Founded:2026
Batch:Fall 2026
Team Size:2
Status:
Active
Location:San Francisco
Primary Partner:Gustaf Alstromer