It seeks to close the AI skills gap.

August 22, 2017

Databricks, a software startup that wants to make artificial intelligence applicable to many businesses, just raised $140 million in a new funding round led by Andreessen Horowitz.

This Series D round brings total funding to $247 million, and includes contributions from new investor Battery Ventures as well as New Enterprise Associates. The company does not disclose valuation. This news comes just six months after the four-year-old company announced a $60 million funding round.

Databricks co-founder and CEO Ali Ghosdi—a well-known AI expert and adjunct professor at the University of California, Berkeley—says the demand for AI is exploding. But he argues there’s a huge gap between that demand and actual, useable AI systems—as opposed to sets of technologies that must be assembled to meet customer needs.

Related: Why Cloudera Is Saying Goodbye to MapReduce and Hello to Spark

“The media talks about AI and how it’s taking over the world and how people are afraid it’s replacing humans,” Ghosdi says but a small fraction of actual businesses—he puts it at just 1% of companies—are actually using the technology successfully. “There’s a massive need here. CIOs and CEOs know that to stay competitive in the next 10 to 20 years, they need AI and analytics.”

Most companies, unlike Google or Facebook, for example, do not have thousands of engineers and data scientists at their disposal. Rather, these tech giants have scouted and hired most of the top talent in the industry, leaving everyone else scrambling.

Databricks says it attacks that gap by providing a collaborative set of software, which lets a company’s business execs and IT professionals share information and requirements about AI projects. Ghosdi explains one portion of the company’s offering, a collaborative workspace, operates like a “Google Docs for AI” because it lets a company’s data scientists, engineers, and managers to log in, chat, look at results, predictions and graphs, and work together to fine-tune results.

Get Data Sheet, Fortune’s daily tech newsletter.

What typically happens in order for AI to work is that lots of data has to be collected and then cleaned up—to get rid of redundant or irrelevant data— and then run through models created for a given task. What then nearly invariably happens is the data has to be cleansed again, and the models tweaked and re-tweaked to get useable results.

Databricks software—which runs on Amazon Web Services, Google Cloud Platform, and Microsoft msft Azure—offers a platform for all of that painstaking work, but lets all authorized parties into the process.

Related: IBM Watson Hurt by Hype?

The new funding will be used to craft industry-specific versions of the software—for health care, finance, mass media, and government tasks—and to expand Databricks’ international presence.

The overall goal here, Ghosdi says, is to democratize AI. But that also happens to be the self-proclaimed mission of Microsoft and other companies in the burgeoning AI market.

SPONSORED FINANCIAL CONTENT

You May Like