This company needed more review than its own team could give, so it sent the work to super.AI through an API and kept building its product.
This company was building a meeting bot that would summarize meetings and send notes to the people who attended. Its own team was checking that output by hand, and that wasn't going to scale.
They didn't want to hand everything to AI and lose quality. Adding people to their own review team wasn't going to keep up either.
They sent meeting data to super.AI through our API. AI did the first pass, people reviewed the parts that needed a person, and structured notes came back into their product.
Because it ran through the API, their team didn't have to babysit it, and they could get back to building features.
Later we worked with them on their customer support chatbot. The bot had to work out what someone was asking and send them the most useful help article. We rated how helpful each article was for each kind of question, so they could tune how the bot picked its answers.
This was one of our earlier projects, from when a lot of our work was building training data for machine learning teams. The idea behind it is the same one super.AI runs on today: AI does the work, and people check what it isn't sure about.
If your team is checking AI output by hand and can't keep up, that's the problem we worked on here. Talk to us about what you're building.
