Stack AI co-founders Antoni Rosinol and Bernardo Aceituno were graduate students at MIT and graduated in 2022, just as large language models were becoming more mainstream. ChatGPT was scheduled to be released globally at the top of the yr, but even before then they recognized an issue with firms putting data along with models without much expertise and knowledge – they usually wanted to vary 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 corporate has developed a low-code workflow automation tool designed to assist firms construct AI-driven workflows, including chatbots and AI assistants. The company has raised $3 million to date.
“Our platform enables people to create workflows that require connecting different tools to work together. We deal with connecting data sources and LLMs since it means that you can create powerful workflow automations. We also offer many other tools and features to automate complex business processes,” Aceituno told TechCrunch. They've only had a working product for six months, but they already report over 200 customers using the product.
Essentially it involves dragging components onto a workflow canvas. This typically includes a knowledge source comparable to Google Drive and an LLM, in addition to other workflow components comparable to a trigger component or an motion component to create the workflow. This allows the client to create generative AI programs without loads of code. The code itself isn’t AI-driven, however the tasks within the workflow often are and should require manual coding to maintain the workflow running easily.
Some of their earliest customers are within the healthcare industry, and Aceituno admits that they should be careful with applications involving doctors and patients, especially when internal data sources will not be all the time reliable or could contain conflicting or outdated information.
In such cases, he says, it is crucial to depend on the human expert, the doctor, to make a decision the standard of the response. As an additional safeguard, they include source citations in each answer in order that the healthcare skilled can confirm the source before accepting the reply.
“That being said, it’s true which you could throw garbage in after which the citations will even be garbage and that’s the reason it’s mandatory that these assistants don’t completely take over the method,” he said.
Rosinol comes straight from MIT and founded a startup. Rosinol says attending YC really helped them understand the business side of things and refine their startup idea by working with customers.
“We began with an initial version of this API that was way more developer-focused. And we began with some customers with the concept that we wanted to make use of AI to automate RFP responses or sales. And through working with customers, it became very clear that the true challenge was not training a model, but fairly effectively querying data sources and connecting them to those language models.”
The company currently employs six people but is hiring engineers and sales and marketing professionals.
The $3 million investment was accomplished a few yr ago. Investors include Gradient Ventures, Beat Ventures and True Capital in addition to LambdaLabs, Y Combinator, Soma Capital and Epakon Capital.
