Stack AI wants to make it easier to build AI-fueled workflows

Stack AI’s co-founders, Antoni Rosinol and Bernardo Aceituno, were PhD students at MIT wrapping up their degrees in 2022 just as large language models were becoming more mainstream. ChatGPT would be released to the world at the end of the year, but even before that, they recognized a problem inside companies putting data together with models without a lot of expertise and knowledge — and they wanted to change that.

After graduating, they moved to San Francisco and joined the Winter 23 cohort at Y Combinator, where they launched Stack and refined their idea. Today, the company has built a low-code workflow automation tool designed to help companies build AI-driven workflows including chatbots and AI assistants, for example. The company has raised $3 million so far.

“Our platform allows people to build workflows that require connecting different tools to work together. We focus on connecting data sources and LLMs, since doing so allows you to build powerful workflow automations. We also offer many other tools and functions to automate complex business processes,” Aceituno told TechCrunch. They’ve only had a working product for six months but already report over 200 customers using the product.

Essentially, that involves dragging components to a workflow canvas. That typically includes a data source such as Google Drive and an LLM along with other workflow components such as a trigger component or an action component to build the workflow, allowing the customer to create generative AI programs without a lot of coding. The coding itself is not AI-driven, but the tasks in the workflow often are, and could require some manual coding to make the workflow work smoothly.

Some of their earliest customers are in the healthcare industry, and Aceituno acknowledges they have to be careful with applications involving doctors and patients, especially when internal data sources aren’t always reliable or could contain contradictory or obsolete information.

In those cases, he says, it’s important to rely on the human expert, the doctor, to make the call on the quality of the answer. As another level of protection, they include source citations in every answer, so the healthcare professional can check the source before accepting the answer.

“That being said, it’s true that you can put garbage in and then the citations will also be garbage and that’s why it’s required that these assistants don’t take over the process completely,” he said.

Coming right from MIT and launching a startup, Rosinol says going to YC really helped them understand the business side of things and how to refine their startup idea by working with customers.

“We started with an initial version of this API, which was much more developer focused. And we started with a few clients with an idea that we wanted to use AI to automate RFP responses or automate sales. And by working with customers, it became very apparent that the true challenge was not in training a model, but rather in effectively querying and connecting data sources to these language models.”

The company currently has six employees, but it is hiring engineers and sales and marketing pros.

The $3 million investment closed about a year ago. Investors include Gradient Ventures, Beat Ventures and True Capital along with Lambda Labs, Y Combinator, Soma Capital and Epakon Capital.