The transformative potential of artificial intelligence for smarter, faster operations is no secret. As more implementations roll out across businesses, AI will present a paradigm shift to what constitutes “work,” which has long been defined by the manual processing of exceptions, or the last-mile problems, that automation could not address.
But AI and automation will not eliminate people from the picture. Far from it, since technology only solves half of the equation. We will need people with domain knowledge of their industry and business processes to add context to digital applications and guide desired business outcomes. Such human-and-machine collaboration is what will define the next generation of work, with intelligent systems certainly taking over some processes, but also freeing up employees to take on more strategic jobs focused on growth.
In these new environments, people will work alongside digital tools in augmented tasks and new roles, creating an unseen hybrid workforce. Genpact research found nearly 80 percent of global companies that are AI leaders believe their employees will be comfortable working with robots by 2020. Moreover, in a related Genpact study of workers, 40 percent said they would be comfortable with robots in the same timeframe.
However, managing this hybrid workforce is very different from managing people alone. We will need to implement new structures and processes for effective management and ensuring successful change.
Plan with change management in mind
We sometimes overlook change management in a rush to get projects off the ground, which creates problems down the line – especially as some workers only think of the negative consequences of adding robots to the workforce. We need to develop a clear and actionable change management strategy to mitigate possible business disruptions – from people and technology.
On the employee side, we need to communicate the expected changes well in advance so people are aware, and can prepare for the change. For the digital workforce, we should remember that processes do not exist in a vacuum, but instead involve numerous exchanges upstream and downstream.
For example, a minor change to a webform on the front end can disrupt robotic process automation working in the back end. Mapping out how the various bots and systems will work together can prevent roadblocks.
Visualize the operational workbench
In the past, we could easily see when employees clocked in, clocked out, or did not show up for work. With the integration of robots, it becomes more difficult to track if there is a glitch that stops them from working.
For instance, if there if there is a coding glitch in a CRM system, it can take time to isolate, which then manifests in a problem with the profit & loss numbers weeks later.
A potential solution is to develop a “visualization dashboard” to oversee automated operations. A dashboard provides a single view across locations, environments, and systems. We then have visibility over all robots, intelligent automation, and workflow orchestration. This makes it easier to identify any issues and deploy quick fixes.
Establish a strong governance protocol
By tracking robot operations, we can establish better governance, which is essential considering potential AI bias. These systems are only as good as their data samples, and imbalanced datasets or inherent human biases can manifest in unintended outcomes.
For instance, software used to understand human expression might struggle to “read” people of different ethnicities if trained with homogenous data. Bias can develop over time, such as conversational AI that might emulate negative language – a hard lesson for Microsoft and its Twitter bot, “Tay.”
To avoid unintended outcomes, it is up to humans to govern the machines and identify conscious and unconscious biases. Diversity in the teams working with AI can help. Bringing in team members with different skills and approaches leads to more balanced datasets. They can also train the machines to avoid negative ramifications.
Understand AI’s limitations
There is a misconception that AI is more mature than it is – particularly in business. Understanding AI’s limitations can help with digital workforce planning and knowing where to apply employee domain. For instance, experienced staff with knowledge of the multiple exchanges within a process can make sure that automation works from end to end, and orient it to a specific business goal.
The intersection of domain and digital is what will drive success. We need people who can translate their experiences to these new applications. There is an opportunity now to reskill employees on this intersection between domain and digital, breeding not only greater talent, but also better employee goodwill and engagement. At the same time, we should focus our recruiting programs to find “bilingual” people who have both domain and digital knowledge, and can bridge the connections.
The time to start thinking about managing a hybrid workforce is now, not tomorrow. We need to make sure we have the right processes and structures to manage the emerging digital workforce, with strong change management and governance protocols. We must also start planning – across education and talent acquisition – to reskill existing employees in materializing the benefits at the intersection of domain and digital.