Can we talk rationally about Artificial Intelligence for a minute? The still-developing technology inspires some fairly intense fear-mongering among literally the brightest people on earth. These are often the same people deploying AI in cutting-edge technology like self-driving cars.

It’s hard to account for the disconnect. Are they so excited about the current leaps in AI cognition that they leap forward to an imagined future and then shrink back in horror? Perhaps. But it’s also quite clear that they’re hop, skipping and jumping over what’s happening between AI and business. My guess is it’s because the kind of AI being quietly deployed across all kinds of businesses is rarely called AI. Instead, the companies that use it and those that deploy and integrate it with existing systems call it Robotic Process Automation (RPA) and the bits of AI’s, or Virtual Robots, that handle repetitive and mundane tasks like pouring over 1000 resumes to pluck top 25 candidates.

I’ve covered AI for more than a decade, but until a recent, wide-ranging roundtable conversation with representatives from Avanade, a business solutions company, Blue Prism, which delivers digital workforce solutions to business, and London School of Economics Professor Leslie Willcocks, I’d never heard of Robotic Process Automation. However, much of what we discussed and the insights they shared about how businesses are preparing for and integrating these software-rebased robotics solutions into businesses aligned with much of what I already believe about the potential benefits of AI.

Hello Reality

As the media depicts it, “there are one of two stories. AI is either a triumph or a disaster,” Professor Willcocks, who has written extensively on automation, digital business, and the future of work, told me. In the disaster scenario, robots and AI take all of our jobs and replace humans. In the triumph scenario, robots and AI take over our tasks and we spend our lives on the beach. “The truth is somewhere between this.”

Even so, the very concept of automation leads to fear of human obsolescence.

“There’s still a concern when you mention automation,” said Avanade Product and Innovation Lead Pam Maynard. “And only when they start to see what that automation actually means in the organization. So, the first projects go in and they can see the implications for coworkers. They can build some trust on why automation is actually a good thing.”

These integrators often have to dispel the notion that automation somehow puts the robots in change.

“AI and robotics if done correctly can empower human workforce,” said Blue Prism’s Chief Evangelist Pat Geary. The robots are working for humans, “freeing up human capital to do things that humans are really good at.”

Where it works

Obviously, you can’t simply drop in a virtual robot or RPA and expect it to work. Geary told me that many companies have processes and system that live in stovepipes. The gaps between them, handing off data from one to another or processing the data out of a system so the other system can use it is a job for RPAs. Killing those gaps can also effectively smooth out operations by lessening downtime between disparate processes.

But implementing RPAs can mean taking a fresh look at your business and how it’s organized. Geary said that business leaders will sometimes want to take existing jobs apart to figure out which parts are good for humans to do and which parts are good for virtual machines.

“RPA allows humans more time doing the things that require more of an emotional connection,” said Maynard. AI’s, RPAs and robots still struggle with context. It’s why conversations with the AIs in our pockets and counters (Siri, Google Assistant, Alexa) are so short. They can’t (at least most can’t) carry on true conversations because they quickly lose the thread of context.

That’s why online customer services are often a mix of chatbots and humans. The RPA-driven chatbot might grab all the initial account information and even the first description of the problem, but if there isn’t a programmatic fix (Internet down = restart router), the chatbot is going to hand you off to a human customer service representative.

Inside business, it’s no different. For over-worked IT departments, RPA can, according to Maynard, vastly reduce incidence response rates through automation. In other words, a short-staffed IT department doesn’t get stuck responding to small, easily solvable IT issues like “I lost my password” and “I can’t install a printer.”

The self-service model is, as Geary sees it, not giving up control, but empowering business operations.

Even with some IT automation, the IT worker is unlikely to ever be out of a job. “There always has to be tech expertise in house and they are massively overburdened,” said Professor Willcocks. “Automation alleviates pressure on IT dept.”

In HR departments, the RPA can handle the first pass on a mass of resumes and entering the data for potential and new employees, but it won’t be the technology you deploy to handle performance evaluations. “Humans do prefer to be judged by other humans, even if the results are not optimal,” said Professor Willcocks.

On the other hand, RPA can spot patterns virtually impossible for the human eye to see. For instance, Maynard told me, it can spot trends in performance and hiring that “can be acted upon in a positive way.” And as the companies strive to be more diverse, a watchful RPA might be the first system to recognize when hiring practices are not diversifying a workforce.

Getting Ready

No one is recommending companies simply implement AI or RPAs and expect business improvement.

“What is our business problem and what is our business imperative? The technology should be a long way back,” said Professor Willcocks. “But you do want to listen to the technology.”

Knowing where RPA fits in is a crucial step, but not the only one.

One of the keys to getting ready for RPA or the introduction of a Virtual Robot is data, but perhaps not in the way you think.

“There’s too much focus on the data,” said Maynard, “We have the data. It’s just automatically created. It’s understanding what that data means in the context of a business process,” said Maynard.

More problematic is how often the data that companies do have is not very good. One company Professor Willcocks looked at had a raft of data, 15% of which was usable. “Most organizations have very poor data and need to gear up,” he said.

And that problem is compounded as more companies add smart devices that collect even more unrelated data streams.

After Blue Prism and Avanade walk clients through what RPA is and how it might help their companies, they need to have that hard conversation about the data. Maynard talks to them about improving the quality. For Geary, it’s figuring out how to connect the seemingly unrelated data streams.

Updating Skills

Business workers and their jobs will not go untouched by AI and RPA. There will be job loss, especially on so-called lower tasks and a reshuffling of duties for many knowledge workers. There will also be a growing need for new skills, especially for those building and managing RPAs.

Maynard said that along with the need for more data scientists to understand, program and learn from the algorithms, she sees a growing demand for “conversational analysts and conversational designers in order to be able to program conversational, cognitive robots that understand natural language.”

And as business workers shed mundane tasks for more intellectually and emotionally challenging ones, they may need, Professor Willcocks told me, more training in empathy, leadership and delegation.

These systems are not all about overhauls and rapid change. Geary said that there is a place for your trusty, old accounting system. “[Companies] want to get rid of the shackles of legacy. But if you can reimagine that legacy, if you can repurpose it by putting an intelligent robot layer around it, it actually becomes an incredibly useful asset.”

It is, in essence, another way of looking at artificial intelligence, not as a replacement for human thought and purpose, but as a solution for the jobs and tasks we never wanted to do in the first place. The history of new technology is not one of purpose decimation. Jobs disappear or change, but, more often than not, new ones appear to take their place.

Thanks to Avanade and Blue Prism for sponsoring this post.

Avanade is the leading provider of innovative digital and cloud services, business solutions and design-led experiences delivered through the power of people and the Microsoft ecosystem.

Blue Prism delivers the world’s most successful digital workforce which integrates with any system or application, freeing humans from repetitive work to drastically increase productivity.