About BenchBot
BenchBot is built by xyndata to make conversational AI — chatbots and voicebots — correct, safe, and compliant, from first deployment through every release.
Our Story
BenchBot was created in 2024 by xyndata, a software company based in Prague. It started with a real-world problem: one of Germany's largest retailers couldn't reliably test the conversational AI behind its customer service. Their chatbots and voicebots — built on Cognigy.ai — looked great in demos, but every release risked breaking a conversation flow, leaking data, or mishandling an edge case nobody had thought to check.
We watched their team try to keep pace with manual testing — replaying conversations by hand, maintaining brittle scripts, and still missing the prompt-injection attempts, hallucinations, and compliance gaps that only surfaced in production. Traditional QA was never designed for systems that can be unpredictable, manipulated, or confidently wrong. Testing conversational AI had to become as rigorous and automated as testing any other critical software.
So we built the platform we wished they'd had. BenchBot grew from that experience into a complete testing suite that automatically explores a bot's capabilities, grounds it against a knowledge base, and runs functional, security, and compliance checks across more than 60 connectors — for chatbots and voicebots on any platform, not just Cognigy. Today, xyndata helps teams across retail, finance, and healthcare ship conversational AI they can actually trust.
Our Values
These principles guide everything we do, from product development to customer relationships.
Precision
We believe in getting it right the first time. Every test, every insight, every recommendation is crafted with meticulous attention to detail.
Trust
Building AI that people can rely on is our mission. We help organizations create conversational AI that users trust and regulators approve.
Innovation
We're constantly pushing the boundaries of what's possible in AI testing, staying ahead of emerging risks and opportunities.
Leadership Team
Meet the team building BenchBot.
Patrik Tesar
Founder & CEO
Founder of xyndata, leading the company's mission to make conversational AI testable, safe and trustworthy.
Louis Peterscheck
Head of AI
Leads xyndata's AI and evaluation work — the red-teaming, model evaluation and methods behind BenchBot's testing.
Ahmed Hameed
Head of Product Management
Shapes BenchBot's product and roadmap, turning real customer needs into testing workflows teams can rely on.