PAID POST FOR COMCAST LIFT LABS

Have a conversation with your data with Diana AI

This is one in a series of stories that highlights each company in the 2019 class of the Comcast NBCUniversal LIFT Labs Accelerator, powered by Techstars. Visit LIFT Labs to learn more about this world-class program for connectivity, media and entertainment startups, and get information on applications for the next class.

It all started when Geeta Banda and Teja Nanduri watched a rerun of "Star Trek: The Next Generation." Captain Jean Luc Picard (played brilliantly by Patrick Stewart) gave voice commands to a supercomputer in an effort to find a lost spaceship.

"Where is the last known place the spaceship was docked?"

"Show me the closest inhabitable planets."

"How far is that from here?"

Banda and Nanduri had a eureka moment. The voice control on the show featured something that intrigued the business partners for years — context.

In the show, the captain could not only interact with the computer in his natural language, but could also build one question off of another. No coding. No constant repetition of the original question. Could they build something like that for business data?

From those humble beginnings, Diana AI was born. The company offers a conversational business intelligence tool designed for non-technical users, allowing them to access data quickly and easily using natural language like voice or text. Since it's aimed at non-technical users (which are most employees in a typical business), it doesn't require complex coding. Instead, you just talk. Here's an example:

"Show me sales data for the past 10 years."

"Break it down by segment."

"Can I see that for just Texas?"

"Compare that to California."

The questions (and answers) build off of one another, making it far easier to drill down and get the most accurate data.

Reimagining Data Analysis

With the rise of big data, every organization wants to make decisions based on metrics and facts rather than gut feeling. But, doing so is still an arduous process. Let's say a human resources team wants to compare health benefit usage between their San Francisco and New York offices. Currently, they'd have to ask a data analyst to build those reports, rely on stale reports that analysts built a long time ago or do some complex coding themselves. It's an antiquated process to say the least.

"Data analysts like to do the more complex analysis and constant requests stop them from doing what's more fun and important to them," Nanduri said. "Our product takes much of the load off the data analyst, so they can focus on complex issues and leave the more simple stuff to Diana."

After working for 14 years at companies like McKinsey and KPMG, Banda knows that data analytics is a major pain point for large organizations. Sure, there are plenty of other tools in the market that help businesspeople analyze data, but nothing that works as easily as Diana AI.

"Over the years, I probably trained 5,000 to 6,000 people to use the existing tools in the market," Banda said. "These non-technical users were put through extensive training for two or three days, and within two weeks, they start forgetting what they learned. Within a month, they go back to their regular ways of working. So that whole training project is wasted. That made me realize that we need to build a tool that's as simple as speaking English."

Nanduri has spent years in data analytics, insights, and warehousing. His skills complement Banda's perfectly — so it's no surprise that they're working together on their third startup. First it was Jump The Line, which let you order and pay for your morning coffee via SMS message. Then it was an e-commerce platform for small to medium-sized businesses. Now, it's Diana AI.

"We complement each other's skills well," Nanduri said. "I'm very strong technically and she's very strong on the business side of things. But I also have an MBA and she's a very good coder. We're a perfect match."

This summer, they moved from San Francisco to Philadelphia for 13 weeks to take part in the Comcast NBCUniversal LIFT Labs Accelerator, powered by Techstars. They're meetingThey met with mentors, refining refined their business models, tried pilots with business units, and tinkering with their product.

"There has been such a support from all the women and men associated with this program," Banda said. "The support that comes through this network is really special."

During the LIFT Labs accelerator, they've already learned one key lesson: Focus on the product customers want, rather than the product they're eager to build.

"Before, we were more passionate about our products and wanted to build the best product based on what we thought was important," Banda said. "During the program, we've learned to focus on figuring out what our customer really wants and prioritizing our product towards that."

They plan on building Diana AI in the west coast after the LIFT Labs Accelerator. They see Philly Philadelphia as a place with great tech talent and plenty of potential clients. Plus, it doesn't hurt that they love the food.

"There's a lot of talent coming out of UPenn and Drexel that will be really helpful in setting up our startup and starting to grow in the next few years," they said. "We're very excited to work in Philly or close by when we finish the program. And, we love the restaurants here."

While they're focused on data analytics at the moment, the future of Diana AI just might include bringing conversational voice commands to the entire enterprise.

"We're very much used to voice at home with Siri, Alexa and Google Home, and the X1 voice remote," Banda said. "How do we bring that into an enterprise and make the experience as seamless as it is at home? For example, setting up a meeting takes a long time. Instead, simply ask, 'find a time when everyone is free on Tuesday.' We want to make it easier for end users at an enterprise [to use] voice."

Disclosure: Comcast and NBCUniversal are parent of CNBC.

This page was paid for by our sponsors. The editorial staff of CNBC had no role in the creation of this page.