9 Ways AI and Intelligent Automation Affect the C-Suite
Members of the C-suite need to contemplate a broader spectrum of issues than technology alone in implementing AI and automation.
![](https://eu-images.contentstack.com/v3/assets/blt69509c9116440be8/blt80ddd73019814be7/64cb547db3ea4f0b1223530e/automation-intro-pixabay.jpg?width=700&auto=webp&quality=80&disable=upscale)
Today's organizations are upping their competitive game using various forms of machine intelligence. From chatbots and virtual assistants to robotics processing automation (RPA) and deep learning, businesses and entire industries are transforming the way they operate, although the results are mixed. Proceeding responsibly requires members of the C-suite to consider the potential opportunities, challenges and risks, including the impacts on people and processes.
For example, McKinsey Global Institute recently analyzed more than 2,000 work activities across 800 occupations. It concluded that 30% of what 60% of all occupations do could be automated today. Apparently, 5% of occupations could be automated entirely.
Meanwhile, more types of businesses are utilizing and experimenting with different forms of AI technology to increase sales, improve forecasting, discover the previously undiscoverable, and increase efficiencies at scale.
One of the biggest impacts AI and intelligent automation have on the C-suite is the expanded scope of responsibility: Executives are no longer responsible for only people; they're responsible for people and intelligent machines.
"That's a very different model than we've had historically," said Steven Mills, associate director of AI and Machine Learning at global management consulting firm Boston Consulting Group (BCG). "You have to understand the technology and have the ability to effectively manage it. In the end, [you] are responsible for it, including any issues or errors that happen, so you need to understand the limitations and how you have to manage the business with this new workforce. I don't think we've done a lot of work preparing for that, yet."
For one thing, AI and intelligent automation are often regarded as technology initiatives, when their transformative potential is must broader than that. "Most CIOs would say, 'I'm going to put my lead architect, technical implementation lead and a data analyst on a team," said Sanjay Srivastava, chief digital officer at global professional services firm Genpact. "I recently sat across the table from a CIO who said, 'I need a VP of HR on that team because that's one of the issues I have to solve for.' What a thoughtful CIO!"
The states of maturity vary significantly from organization to organization and industry to industry. For example, management and technology consulting firm Booz Allen Hamilton conducted a flash poll in February that showed half of federal employees participating in the poll had no awareness of automation. By comparison, four out of five employees in the financial sector reported using RPA in 2017, according to Brett Fraser, director of Automation at Booz Allen Hamilton.
Regardless of where a company is on its journey, following are a few ways AI and intelligent automation affect the C-suite.
The definition of "business as usual" continues to shift as human-machine partnerships evolve. RPA, virtual agents, chatbots, and intelligent interactive voice response systems (IVR) technologies are all being used to automate call centers, although the same technologies can also be applied to many other functions including HR, finance and procurement.
"Leaders are having to consider what to do with their labor sources," said Booz Allen Hamilton's Fraser. "We're finding [that those in the federal sector] are not focused proactively on what they can do better. They may have middle management thinking of ideas to improve workflow processing, or the call script, or the contact response methodology, but if they were to adopt more proactive means of augmenting their workforce with technology such as shadow agency using virtual agents powered by AI to always give the expert answer, they'd be able to spread their workforce, no matter the skill set, and provide the expertise that a citizen might like."
When one part of a value chain leverages machine intelligence, other parts of the value chain may be impacted. For example, truck drivers fear the advent of self-driving trucks. However, in the past year, trucking industry automation hasn't advanced much, according to Jake Tully, editor-in-chief of trucking career website TruckDrivingJobs.com due to the initial human involvement required to get it started, the state of safety legislation, hours-of-service issues and infrastructure improvements that need to be addressed.
"The obvious thing is the impact on the 3.5 million U.S. truckers. As you start thinking about the value pool, you see these forward and backward linkages," said BCG's Mills. "Trucking impacts 5.5 million additional jobs at truck stops, hotels and restaurants, so it has amplifying effect in the economy. But there are also backwards linkages such as companies producing these trucks and developing the technology. It behooves executives to ask what the future will look like in their industry."
Responsibility = Opportunities + Risks
AI and intelligent automation enable new, competitive opportunities; however, they also introduce new forms of risks. Executives need to consider both in considerable detail to minimize the risks of unintended consequences.
"There are some tough choices the C suite faces of how and when to introduce specific technology. It's important to understand what the goals are," said Josh Elliot, director of Machine Intelligence at Booz Allen Hamilton. "You need to look at your overarching strategy and understand what the risks and implications of that particular technology will have on your organization, your customers, shareholders or whatever is an important thing. We're seeing some organizations get to a [point where] they realize they don't have the governance in place."
Genpact's Srivastava advises his clients to think through the obvious "gotchas" and plan for them in advance.
"It's hard to apply change management when you've got a black box that doesn't explain how it made a decision," said Srivastava. "You need the right expertise that can contextualize the results, go through machine learning and distill the insight from the data."
Businesses have been advised that they need to become more Agile to stay relevant and competitive. Using RPA or RPA and more advanced AI techniques, companies can become even more agile than they are already and get even deeper insights into customer behaviors and preferences.
"If you can create an organization that's more agile and adaptable than your competitor, that's an advantage," said BCG's Mills. "If you implement a technology correctly, you'll suddenly have access to far more information and insights than you've ever had before and in theory that should help you make far better, more informed business decisions."
While it's important to consider what can be automated using existing technologies, executives should realize that the balance between human and machine labor will continue to shift as technologies and use cases become more sophisticated.
"As we increasingly integrate more advanced capabilities into RPA, that will only create even less of a static situation, so when you combine RPA and computer vision or RPA and [natural language processing], you're just upping the sophistication of what you're doing and that means you can't just do it once and forget it," said BCG's Mills.
One common obstacle to RPA is that the people tasked with implementing it don't have detailed knowledge of process they're supposed to automate. Someone in the organization probably knows what the process is, but, for a myriad of reasons, the change between the RPA planning and implementation stages.
"The two keys to automation success are planning and communication. This is what we're planning on doing, here's our current status and next steps," said Booz Allen Hamilton's Fraser. "You cannot view this statically, you've always got to be open to change."
Some organizations approach RPA from the standpoint of workforce reduction when workforce evolution may be a more appropriate strategy.
"On the commercial side, they say, 'With the advent of 50% automation, I should be able to reduce 50% of my staff, but who's going to build the automation? Who's going to quality assure the automation? Who's going to run the communication plan and the projects and programs that are spawned by the automation?' " said Booz Allen Hamilton's Fraser. "If you have a proactive approach you're going to want to find out, of those resources that are going to be impacted, where will we put them, who should be cross-train? Automation makes humans better at what they do."
Genpact's Srivastava said that AI discussions have increased 30% in the past year, but the interest doesn't necessarily mean companies know what to do with it yet.
"I think a lot of executives are confused about the topic right now. I tell people don't look at AI as a solution looking for a problem, start with a problem first and figure out what's behind it, test a couple of ideas, and then once you have an idea, they you outfit it with the right technology whether it's AI, IA, or something else," said Srivastava. "Don't start with, 'Where should I apply my AI?"
Industry disrupters tend to have a privileged dataset that they leverage for competitive advantage. BCG's Mills suggests other companies build their own privileged datasets using combination of their own data, open data sources, and partner data.
"By pooling your data, you create something unique that no one else has," said Mills. "If you're just aggregating data for data's sake, you can have a huge investment without any return or idea of how you're going to use it."
All companies have data. The question is whether they can reposition and govern their data assets for use by machine intelligence. Without labeled data, it's nearly impossible to train a model, said Booz Allen Hamilton's Elliot.
There's a lot of information and misinformation about AI and intelligent automation. To avoid unnecessary angst and confusion, members of the C-suite are encouraged to learn the basics.
"Read what's out there. You need to understand what it can and can't do, and the state of the art in your industry," said BCG's Mills. "Then, assess where you company is so you have a clear idea where you're starting from, what types of advantages you have and where the gaps are. That will inform your long-term strategy. Then, it's a matter of prioritizing specific use cases and that will be coming out of business units."
Booz Allen Hamilton and NVIDIA jointly offer an executive-level course, for example.
"It's important to understand what the terms mean and what the risks are," Booz Allen Hamilton's Elliot. "It's critical to be able to trace back and be accountable for the decisions that some of these machines and algorithms are going to be producing. [Executives don't have to understand] the nitty-gritty of the tuning of the parameters, but they should understand how an algorithm works, and what it takes to train a model and produce an output."
There's a lot of information and misinformation about AI and intelligent automation. To avoid unnecessary angst and confusion, members of the C-suite are encouraged to learn the basics.
"Read what's out there. You need to understand what it can and can't do, and the state of the art in your industry," said BCG's Mills. "Then, assess where you company is so you have a clear idea where you're starting from, what types of advantages you have and where the gaps are. That will inform your long-term strategy. Then, it's a matter of prioritizing specific use cases and that will be coming out of business units."
Booz Allen Hamilton and NVIDIA jointly offer an executive-level course, for example.
"It's important to understand what the terms mean and what the risks are," Booz Allen Hamilton's Elliot. "It's critical to be able to trace back and be accountable for the decisions that some of these machines and algorithms are going to be producing. [Executives don't have to understand] the nitty-gritty of the tuning of the parameters, but they should understand how an algorithm works, and what it takes to train a model and produce an output."
-
About the Author(s)
You May Also Like