Three tech leaders this morning took LSA19 attendees on a brief but deep dive into how small business can leverage artificial intelligence to drive revenue.

Representing a range of brands and companies that tap big data to help small business owners, the three speakers focused on the “democratization of data” and attempted to demystify AI as they detailed steps that can be taken by entrepreneurs and smaller companies.

Veronica Buitron, CTO for Tangocode with 20 years in tech, said AI is not something in the future, as digital advertising is already being impacted and small business owners should start paying attention. “This is not just for big brands,” she said.

It’s also a misconception that AI can only work if the customer has a lot of data. What’s more important than a large set of data is that the data is clean and that the user is clear on their goals and how the data might correlate to intended outcomes, she said.

Sana Khan, AI lead for Microsoft, said that in the last five years, she has seen a trend where large media buys are coming from increasingly smaller companies. She’s also seeing companies get more creative in how they gather and leverage data. “The old audience clusters aren’t going to work,” she said.

Kushagra Shrivastava, managing director for Yahoo Small Business, a Verizon company, said that for small businesses the issue isn’t necessarily accessing data, it’s understanding what to pay attention to and not getting overwhelmed. The former Googler said SMBs have broad access to data and analytics that are often so dense that business owners can become lost.

He said that has led Yahoo Small Business to start working on a voice-response based system that will allow SMBs to quickly get top level daily analytics, and then have the system send additional details to their phone or laptop. He suggested this is an obvious progression for technology that often starts first with enterprise companies and then filters to small firms.

“Sometimes it may take a little longer for the technology to be digestible for small businesses,” Kushagra said.

How Can Small Businesses Use AI? 

Machine learning at its very root does two things, according to the panelists; either you can classify or you can predict. Classification is all about segmenting customers, while prediction determines how a group of customers will behave when presented with new conditions or programs.

“Classification starts with completely understanding what customer data you have and what you can acquire,” said Kushagra.

The next step is determining your customer’s “journey”, targeting those cohorts you want to influence and then finding experts in those customer life stages and creating a plan.

“You don’t need to go very far and still be able to do some very sophisticated marketing processes,” said Kushagra.

Will AI Eliminate Human Marketers? 

Sana was joined by the other panelists in arguing that human marketing and business leads will continue to play an essential role in directing AI and providing context to both the data and to resulting insights.

“Human cognitive processes are going to help provide context to classification,” Sana said.

“When you talk about the art and science of AI, the art side is equally as important as the science side,” said Kushagra.

In fact, the human side is so important that it’s already driving hiring decisions for small businesses, said Sana.

For companies wanting to get onboard, Veronica recommends looking for marketers who are on the technical side and can quickly understand the data.

Kushagra suggested finding someone who is a master of Excel and understands data. “There is a lot you can do just with the right pivot tables,” he said, getting nods of agreement from the rest of the panel.

Final Words of Advice on AI 

Sana — Learn everything you can now and get your data into the cloud so it can be used without complex and costly transitions.  That way “running AI on a small set of data is really not expensive,” she said

Veronica — Take care of your data and be sure it’s clean.

Kush — Be sure your data is not siloed.

Ultimately, the good-data-in-good-data-out dictum was at the center of the discussion of AI for SMBs. All the panels also seemed to agree that future uses of AI will be led by professionals who know how to ensure the systems are asking the right questions and that answers accurately address business and marketing needs.