Small Businesses Are Using AI—Sometimes

Recent data shows that around a quarter of current US businesses are using artificial intelligence (AI) in some aspect of their business. Sure, that includes the behemoths, including automobile manufacturers, financial institutions, and Trump's legal team. But these days, the adopter list also includes small to midsized businesses (SMBs).

Unsupervised, an AI and data analysis software company, recently completed a study with data compiled from two primary sources, namely Stanford University's 2021 AI Index Report and the Deloitte study, State of AI in the Enterprise. It also factored in relevant data from CB Insights and International Data Corporation (IDC). Unsupervised backed all that up with its own survey of 520 small to midsized business (SMB) owners querying their attitudes towards implementing AI in their operations.

It's an important question: IDC is saying that the use of AI has evolved from the "early adopter" stage into the "mainstream use" phase, so it's predicting that business and government spending on AI will bump up to $97.9 billion in 2023—about two and a half times what it was in 2019. The big driver is that fully 20% of survey respondents say that using AI has helped increase revenue.

Until now big industrial companies have been the main AI adopters, using it primarily to power robotics for manufacturing. They're also using it to build bleeding-edge machine-learning (ML) algorithms for advanced data-processing techniques, such as predictive analytics. Aside from those primary uses, the Unsupervised study showed that 21% were using it for product and service development, and another 21% focused their operational strategies with AI. Fifteen percent were adding it to their marketing and sales tool kits to get a more accurate handle on important trends and hard-to-pinpoint metrics such as customer sentiment.

SMBs are the new adopters on the block, but they still aren't using AI as pervasively as enterprises. Unsupervised's data did show that many are deploying it for "softer" tasks, though. For example, the most popular AI chore for smaller businesses is creating content for marketing emails and advertising, with 47% of business owners saying they engaged in this practice. Here, they're using AI to match sales content against much more granular customer demographic slices and even automatically tailoring the content to fit. That means a reduced need for those pesky writers and editors. Awesome.

Another 44% cited chatbots as their primary use case. Creepy or not, AI's are making significant strides in the chatbot arena. Many companies use them as the first line of contact for help desk and sales efforts since most people can't tell the difference between humans and machines in those conversations. Meanwhile, 35% said inventory management. If you're wondering how AI applies to inventory management, you probably haven't tried to order Christmas presents this year.

Cybersecurity is also a big use for AI, even among SMBs, since any firm can be the target of ever-more-sophisticated hacker attacks. Around 29% of Unsupervised respondents fell into this category using AI for advanced endpoint protection, threat analysis, and ultra-fast automatic response to attacks. However, SMBs also point AI at all that data they've been collecting and never know how to use. Up to 26% of respondents said they were using it for fast and very detailed business analytics on large data sets, while 20% were using it to get a handle on what kinds of data they have.

The Blockers

Respondents who said they weren't yet using AI cited three key reasons. First, almost half found the technology too expensive, and 28% indicated they were still looking for more use cases to justify that cost. And similar to the current problem with cybersecurity, AI is experiencing a skills gap: About 40% of Unsupervised's respondents indicated they would implement the technology if only they could find adequate on-staff expertise. Finally, AI risk mitigation was cited as another key blocker. Here, the Deloitte study went into more detail on business attitudes towards AI risk than either the Unsupervised report or Stanford's study.

Deloitte found that cybersecurity is the most-cited risk among medium and large businesses, with 62% of its respondents saying it was a significant adoption concern. And since so many companies said they were worried about a qualified staff shortage, it's no surprise that 58% of Deloitte's surveyed CEOs were also worried about the impact an AI failure might have on their immediate business process.

Recommended by Our Editors

10 Steps to Adopting Artificial Intelligence in Your Business

What Is Deep Learning?

Artificial Intelligence Has a Bias Problem, and It's Our Fault

Expounding a little more on the AI creepiness factor, businesses of all sizes were concerned with customer blowback once they found out the company was using AI. Apparently, 57% of CEOs are worried about how their customers might react to discovering that a machine intelligence was picking over their data. That's not only because it feels invasive but also because people are concerned AIs might use that data in new ways without permission. This attitude seems to be in response to various government surveillance projects that use AI, notably widespread cameras capable of facial recognition, which many people think is about as attractive as an open autopsy.

Finally, 57% were also concerned with the current state of AI legislation. AI legalities in different jurisdictions and a longstanding and increasingly nasty debate about proper AI ethics have those CEOs worried about whether any current implementation might suddenly become illegal in the future. Roughly half of Deloitte's CEO respondents believe the government should heavily regulate the use of AI. In contrast, the other half believes heavy regulation will cripple ongoing research, and with it, the benefits of the technology.

Unsupervised reports that these and other concerns are the reason why almost three-quarters of its respondents have yet to implement AI. But if the industry can conquer cost, customer attitude, risk, and staffing hurdles, and then squish the benefits into more easily used tools, we might see a marked adoption uptick in 2022.