Textio announced today that it has raised a $20 million series B round of funding to expand its AI-powered augmented writing platform into fields beyond improving job descriptions.
The company’s product offers automated guidance so that people writing job descriptions are able to maximize their response from interested and qualified applicants. As people write in Textio, the system will score their work on a 100 point scale. If their writing scores above a 90, they’re promised better outcomes when it comes to the qualification level of applicants and the speed at which they can hire someone.
It’s a fusion of machine learning analysis and human creativity that Textio CEO Kieran Snyder calls “augmented writing.” The company ingests 10 million new job postings every month and uses that data to improve its model for how to create a good listing.
Scale Venture Partners led the round, with participation from Bloomberg Beta, Cowboy Ventures, Emergence Capital, and Upside Partnership. Stacey Bishop, a partner at Scale, will join Textio’s board as part of the deal.
Right now, the product is focused on helping companies like Expedia write job descriptions that are more attractive for applicants, so they can build a more diverse applicant pool more quickly. Textio is looking to expand into other areas of writing, like sales collateral.
But even with the expansion, don’t expect Textio to take on Microsoft Word or Google Docs. Snyder said that the company will stay focused on narrower applications with defined outcomes.
“If you’re trying to provide one-size-fits-all guidance, you can’t get to these quantitative promises about how writing will actually work,” she said.
Textio has to do a lot of work to help its users write better based on the AI recommendations that its system can generate.
“The hardest part of the technology stack is turning the pattern in the data into guidance that a real human being can follow,” Snyder said.
While scoring a post is important, it doesn’t matter quite as much as why a score changed, according to Snyder.
To determine that, Textio actually tests different suggestions for users to try to figure out which tips correlate with score improvement, since the system underpinning the scoring doesn’t naturally produce those insights in a human-intelligible format.