top of page

CHAT-KI – AI for Inclusive Employment

Helping employers find funding for inclusion and workers with disabilities discover the support they deserve.

Context

The founders of entender.IA, Jens and Natalie Beyer, have been delivering human-centered AI solutions since 2016 through their German company LAVRIO.solutions GmbH. This flagship project demonstrates the expertise they bring to entender.IA—and is directly transferable to our mission of building accessible AI for vulnerable populations.

The Challenge

In Germany, people with disabilities and employers face a complex landscape of support programs, subsidies, and legal entitlements. Information is scattered across multiple agencies, written in bureaucratic language, and difficult to navigate—especially for those who need it most.

The Solution

CHAT-KI is an AI-powered chatbot that makes disability employment support accessible to everyone. Using natural language processing, the system understands user questions, searches through official sources, and provides clear answers with full source citations. Key accessibility features include voice input/output and easy-to-read language options.

Our Founders' Role

Through LAVRIO.solutions, our founders led the technical development of the AI system, implementing human-centered design principles throughout. They developed the retrieval-augmented generation (RAG) architecture ensuring all answers are grounded in official sources, and created the explainability layer that shows users exactly where information comes from.

Impact

  • Dual-target system serving both job seekers with disabilities and employers

  • Barrier-free access through voice interface and simplified language

  • Transparent AI: every answer includes verifiable source citations

  • Model for inclusive technology development through participatory design

Funded by

German Federal Ministry of Labour and Social Affairs (BMAS)

Partners

Karlsruhe University of Applied Sciences (HKA/ILIN), Hagsfelder Werkstätten und Wohngemeinschaften (HWK)

bottom of page