Millions of consumers now use smart speakers, also known as personal digital assistants, at home to order pizza or a ride, to control their lighting, or to tune their radios. Recent research from Consumer Intelligence Research Partners estimates that, as of June of this year, 50 million U.S. households had such devices.
That is significant market penetration and yet it will probably pale in comparison to what we will see as AI-based digital assistants move into the workplace. And compared to earlier generations of automation (ranging in sophistication from annoying phone prompts to canned chats that are little better than year 2000-era FAQs), new chat agents powered by machine learning will improve things drastically for end users - and not just act as filters for beleaguered call center agents.
Think about the business applications that most companies run. They are powerful and keep operations humming. But they also tend to be complex and can be difficult to learn.
What if a digital assistant/smart chatbot parlays artificial intelligence to help new hires—or even seasoned veteran employees—navigate all those systems? Think of the wear-and-tear that would save time-strapped employees and the overworked help desk staff who probably answer the same question over and over (and over) each day/week/month?
"Given the current talent and resource constraints aligned with market realities - we need to help make our workforce more productive by 'hiring bots' to work alongside us, says Earl Newsome, CIO of Praxair, a Danbury, Conn.-based industrial gas company.
Meet your new digital colleague
Newsome adds that some companies even name these bots and give them employee IDs and training plans, a practice that he supports.
"Treat them as employees and you'll get the most out of them and have them work with your other employees to reduce fear, anxiety and reluctance."
The key to success in this arena is that the business bots need to learn about users as they interact with them.
If a given executive looks at hiring approvals only on a certain day of the week, the bot knows enough to hold all those requests till that day, so the exec isn't bombarded by requests during times he is busy on other things, for example.
Enterprise software goes bot crazy
Large companies tend to rely on many different applications of varying vintages and provenance. What if there were a bot that could walk new employees through this welter of features and functions and guide them as to their use?
Such a "superbot" would act as traffic cop for all bot interactions. As more enterprise applications come with their own bot, wouldn't it be nice to have a sort of uberbot to mediate between all of them?
This bot-above-all-bots could ensure that changes in employee status ripple throughout the organization. If, for instance, an employee moves to a new state, updating the corporate directory, the bot would know that the employee's tax forms and pay statements must also be updated and either remind the employee to do so or automate the process.
That's the future foreseen by Suhas Uliyar, vice president of product management for Oracle.
"Conversational interfaces are an ideal way to for employees to use natural language to interact with digital assistants that are powered by AI to provide immediate and consistent response to information without large time investments,' he says.
"They can reach across multiple software platforms and data stores to populate information, return information, and complete processes in an incredibly efficient manner. As businesses grow and scale, these bots, and their successors, digital assistants, will allow employees to quickly complete minor tasks and focus their time and energy on larger initiatives," he adds.
All of that is important to organizations that must look for ways to streamline processes, whether that be in human resources, benefits, hiring or approval applications. They can even aid in data analysis.
Anyone who has joined a large company knows how time-consuming and frustrating it can be to fill out paperwork on hiring, benefits and taxes, and check its accuracy. Bots would help in all that.
It's hard to gauge just how large this market will be, but as one indication, market research firm Gartner expects over half of medium-to-large companies will deploy chatbots by 2020.
Real AI smarts will be key to success
However, bots must be truly helpful to make a real impact in the workplace. That's why making good use of machine learning and other advanced AI tools is key. Machine learning, properly applied, enables bots to learn from individual users, and users in aggregate, how to respond best to queries—typed or spoken—and even anticipate user needs.
Getting specific answers to real questions, or useful guidance, is a lot better than being confronted with a list of web URLs, for example, and that's the difference between early bots and those that are now coming online.