Despite having spent billions of dollars on boosting employee engagement and improving the employee experience,...
enterprises have made little headway, according to a new report from Forrester Research. Indeed, between the labyrinth of legacy systems, poorly integrated apps and a slew of regulations and channels typical for most enterprises today, workplaces are more complicated and chaotic than ever -- and it's only going to get worse.
One possible solution, the report states, is to bring more AI into the fold. The same conversational AI technology that enterprises are deploying to assist customers, for example, can also be used to relay business data, knowledge and protocols to employees -- advising them, answering their questions and taking over some of their mundane tasks.
"We shouldn't be afraid of machine intelligence -- we depend on it more each day," said Craig Le Clair, author of the report, which is aimed at infrastructure and operations professionals. Humans couldn't possibly navigate as well as a GPS or memorize the content of millions of websites as ably as the lowliest search engine, he pointed out. "For these tasks, machines are a lot smarter than we are and are advancing rapidly to support the workforce."
As AI technologies are applied to business process and internal operations, Le Clair said the onus is on enterprises -- and, in particular, on IT, technology's in-house ambassador -- to help employees make the shift toward relying on machine intelligence to help them do their jobs better.
Embrace cyborg thinking
The first step in effectively utilizing AI in the enterprise is figuring out which tasks are better suited to humans and which are better given to machines. Machines excel at searching for data, correlation and pattern matching, repetitive data entry, computational reasoning and understanding multichannel context, and risk management. Humans, on the other hand, excel at conversational intelligence, collaboration and brainstorming, sentiment analysis, connecting data and ideas, and learning.
Craig Le Clairanalyst, Forrester Research
This division of talent opens up a lot of interesting opportunities for human-machine collaboration, Le Clair said, particularly in areas like human resources, IT service management, call center support, financial advisor support and insurance intermediaries support.
Use cases of human-machine collaboration in the workplace that have paid dividends include:
- An insurance company that built an AI HR advisor to help employees resolve issues by combining business rules, machine learning and a natural language processing model.
- A global bank that built a chatbot that's available behind the scenes to advise financial advisors on complex tasks like retirement rollovers and new account setups.
- An insurance company that built a virtual assistant that unlicensed representatives could text to ask questions and determine whether compliance would allow them to handle a call.
- Financial group SEB, which built a cognitive application for internal IT services support. SEB combined a chatbot and robotic process automation (RPA) to build a machine that can reset passwords, unlock Microsoft Active Directory accounts and point employees to the right IT service solution.
"In science fiction, a cyborg has a human brain and a mechanized body," Le Clair said. "Enterprises can embrace the metaphorical cyborg form and split human and machine tasks in a similar way. In general, humans should possess the 'head' but give up a good part of the 'body' to AI and RPA."
That will be a fairly dramatic shift for employees, Le Clair said. The goal right now is to get employees to "become more comfortable turning to a machine and asking for help."
What can CIOs and IT leaders do to help facilitate this change?
"Show [employees] that the work the bot is doing is low-value and allows them to progress in their careers and leverage human skills," Le Clair said. "Communicate often and involve [employees] in any AI design -- make them a part of the training and feedback process."
Restructuring jobs rather than replacing them is central to Le Clair's enterprise AI philosophy. He tells IT leaders to first understand the specific tasks that make up a job, then they can prioritize the lowest-value, most annoying and most tedious tasks and give those to the machines. This also promotes incremental machine support rather than wholesale full-time equivalent replacement.
"Many jobs are mundane," Le Clair wrote in the report. "Routine calls, data entry, report preparation, logs, and documentation often dominate. They tend to be task-oriented and downplay human interaction. Yet the most fulfilling part of the job may be exactly that element: conferring with others, which AI can enhance."
Bringing in the machines won't be a walk in the park, Le Clair warned. He listed a few challenging questions that arise with more human-machine collaboration in the workplace:
- AI can shift control -- decision-making -- from employees to machines. How is this governed?
- What new, potentially adverse events need to be managed? Brand harm or litigation due to training bias or less transparent decision-making are examples.
- Change management: How are employees' attitudes and anxieties managed through advancing automation?
Le Clair's No. 1 piece of advice for CIOs bringing AI into the fold at their organizations: "Take a broader view of automation."
He suggested CIOs create an automation operating model that is based on a framework that allows different types of automation to be categorized -- from simple workflow to RPA to AI and machine learning. Le Clair also tasked CIOs with providing advice to the business on aligning the appropriate level of automation to the right use case.