Dutch startup exits stealth with backing from Google’s Gradient Ventures
In a manner akin to how companies adapt their software to function across various desktop, mobile, and cloud operating systems, businesses also face the need to configure their software for the rapidly evolving AI revolution. This revolution is characterized by the emergence of large language models (LLMs), which facilitate potent new AI applications capable of comprehending and generating human-language text.
While a company can already develop an “LLM-instance” of their software based on their existing API documentation, the challenge lies in ensuring compatibility with the broader LLM ecosystem. It is crucial to gain insight into how well this particular instance of their product performs in real-world scenarios.
This is precisely the challenge that Tidalflow aims to address with its comprehensive platform, which enables developers to seamlessly integrate their existing software into the LLM ecosystem. This startup has recently emerged from stealth mode, securing $1.7 million in funding from investors such as Google’s Gradient Ventures and Dig Ventures, a VC firm founded by MuleSoft’s Ross Mason, with participation from Antler.
Consider a hypothetical scenario to illustrate the importance of Tidalflow’s solution: An online travel platform desires to embrace LLM-enabled chatbots like ChatGPT and Google’s Bard. They want customers to be able to request airfares and book tickets using natural language prompts within a search engine. The company creates separate LLM-instances for each chatbot. However, they have no way of knowing if, for example, 2 percent of ChatGPT results suggest destinations that customers did not actually request, and the error rate might be even higher for Bard. This lack of certainty can lead to hesitation, especially for companies with a fail tolerance of less than one percent.
This is where Tidalflow steps in, offering modules that not only facilitate the creation of LLM-instances but also enable testing, deployment, monitoring, security, and eventual monetization. Companies can fine-tune their LLM-instances within a simulated sandboxed environment, tailored to each ecosystem, until they meet their specified fail-tolerance threshold.
Tidalflow functions as an application lifecycle management (ALM) platform, where companies input their OpenAPI specification/documentation. The platform then generates a battle-tested LLM-instance of the product, offering monitoring and observability to assess its real-world performance.
Tidalflow allows companies to rigorously test their generative AI’s capabilities, exploring numerous edge cases that may or may not challenge its sophisticated AI capabilities. This is particularly critical for larger enterprises where the risks associated with compromised software reliability are substantial.
Tidalflow was founded just three months ago, with its founders meeting through Antler’s entrepreneur-in-residence program in Amsterdam. The company quickly became the fastest to secure funding in the history of Antler Netherlands once the official program commenced in the summer.
As of now, Tidalflow boasts a team of three, including its co-founders and Chief Product Officer (CPO) Henry Wynaendts. With a fresh injection of $1.7 million in funding, the company is actively seeking to expand its team, recruiting for various front-end and back-end engineering roles as it progresses toward a full commercial launch.
The rapid journey from inception to funding underscores the current frenzy surrounding generative AI. With ChatGPT offering an API and support for third-party plugins, Google following suit with the Bard ecosystem, and Microsoft integrating its Copilot AI assistant into Microsoft 365, businesses and developers have a significant opportunity to harness generative AI for their products and reach a vast user base.
Much like how the iPhone transformed the landscape of mobile-friendly software in 2007, we are now at a similar juncture where software is becoming LLM-compatible, as highlighted by Tidalflow’s CEO, Sebastian Jorna.
While Tidalflow will remain in closed beta for the time being, the company plans to launch its platform to the public commercially by the end of 2023.