Six months after emerging from stealth mode, Karlsruhe-based Octomind has raised €4.5 million to transform web testing with generative AI. The round is led by European early-stage venture capital firm Cherry Ventures and features a complementary set of angels such as Sean Mullaney of Algolia, Charlie Songhurst (ex-Microsoft) and Lutz Finger (ex-Linkedin, ex-Snapchat). The fresh capital will go towards baking stability and trust into AI-based testing.
Founded in 2023, Octomind created the first working AI agent that knows what to test, writes the tests, and keeps them up to date – all without human input, solely from looking at a website. It enables businesses to find bugs before customers do.
In 2024, the majority of software development is shifting to no-code tools and AI-generated code. Latest research points to the low quality of AI-generated code. Poor code leads to app failures, revenue loss, and skyrocketing costs to clean up the mess. AI will churn out even more apps in the future. It will require tools that can handle much bigger volumes much faster – and that’s where Octomind comes in.
“Thanks to no-code tools and AI, everyone can generate web apps. But that’s not the full story. They also need an expert who guarantees that the generated code doesn’t contain bugs. We’ve built Octomind to be that testing expert,” said Marc Mengler, co-founder and CEO of Octomind.
Its founders, Marc Mengler and Daniel Roedler, brought Octomind to life out of a shared frustration with the status quo in software testing. Marc has been developing web apps to bring his AI research to users since 2015. Daniel was previously responsible for GoToMeeting and AI data labeling products at Understand.ai. These featured millions of users per day. Testing was a critical part of the development routine in these companies, but no good solution existed to automate the process. They had to invest heavily in manual quality assurance.
“We all come from building applications that had to run flawlessly 24/7 for customers all over the world. We wanted to address something that bugged us the whole time there. Testing every release to ensure we didn’t break the app was a major pain and slowed us down. Luckily, we knew AI and how to bend it to working applications. It was a no-brainer,” commented Daniel Roedler, co-founder and CTO of Octomind.
The biggest challenge of AI-based testing is combining the deterministic nature of tests with non-deterministic AI. Developers need to ensure that the test code is 100% valid. That’s why Octomind uses AI for its commonsense knowledge and leaves the code generation to more predictable systems. Its AI agent navigates through the web app, understands it, and identifies relevant user flows that need to work. It mimics the user (e.g., clicks on the input field and enters email, clicks on sign-up to newsletter, etc.), records and stores the interaction chain. The corresponding test code is created deterministically without any hallucinations.
“Octomind addresses one of the world’s largest ($38B) and least automated markets that can be augmented by AI. Today, software testing is the single most hated task for developers. Yet, more and more CTOs expect devs to do the testing themselves, which grows the total market to $184 billion worldwide. Thanks to their impressive experience and well-advanced AI agent product, I believe Octomind will make developers love software testing soon,” added Jasper Masemann, partner at Cherry Ventures.
