Episode Summary: In this episode of the AI in Industry podcast, we interview Rajat Mishra, VP of Customer Experience at Cisco, about the best practices for adopting AI in the enterprise and how business leaders should think about the man-machine balance at their companies.
Mishra talks with us about how the executive team should be able to imagine the future of specific work roles that might integrate AI technology or envision how those roles will shift in the short-term. In other words, how will AI affect workflows?
Subscribe to our AI in Industry Podcast with your favorite podcast service:
Guest: Rajat Mishra, VP of Customer Experience — Cisco
Expertise: Analytics, strategy, customer experience, business transformation, product management, machine learning
Brief Recognition: As part of the Customer Experience (CX) executive leadership team, Mishra leads some of Cisco’s initiatives related to CX product portfolio, product management for CX Collaboration and CX Internet of Things, CX Partner Model and Programs, global strategy, acquisitions and incubation. Prior to Cisco, he served as General Manager of Mu Sigma and Associate Partner at McKinsey. Earlier in his career, he worked at Microsoft as an engineering manager and had a brief stint at Google. He earned his Bachelor’s degree in Computer Science and Engineering from the Indian Institute of Technology, Delhi, and his MBA from the University of Pennsylvania – The Wharton School.
Interview Highlights
(03:20) There are a lot of considerations that make AI slower to integrate into the enterprise. What are some of these reasons for business leaders to think about before starting on the journey?
Rajat Mishra: The tech is far ahead of the usage in the enterprise. And as I think of the main reasons why the adoption has been slower than the potential, one fundamental reason I see enterprise is struggling with is clarity on where they want to stand in the human condition. Debates around would machines dominate the future of problem-solving in companies or would it be people? Which is the strongest asset and what is the balance? Answering that question is fundamental to figuring out what your AI strategy would be.
Another thing that I see is which initiatives to focus on. Most companies don’t start green field, most start brown field. There are many initiatives in play around AI. With all the hype, sometimes it feels like AI is the hammer and people are looking for nails to hit the hammer with. Which initiatives to pick is a hard one.
There is also a question around how to get the right AI experts. This is a hard area to hire and retain. How do you get the right experts into play so you can develop and scale the business? Those are some top challenges I see in enterprises and customer desks.
(05:00) We’ve done some interviews around choosing the low hanging fruit AI applications. Talent and vendor selection are some of the considerations. What are the next steps to make sure we can see a return on investment (ROI)?
RM: That’s speaking from personal experience. Many of the questions around talent and vendor selection mask the fundamental issue of where the balance of the strategy between people versus machine is. We ran this question also and surfaced the assumptions as we were building our own AI products and services. What helped us was the Iron Man movie as the metaphor. Tony Stark is a brilliant engineer and develops an exoskeleton. Tony Stark himself is great, but the exoskeleton is wicked cool. When you think of Iron Man, it is Tony Stark and the exoskeleton coming together. That creates an amazing superhero.
The question we discuss at an executive meeting before we could plan which vendor, which talent strategy is, “where do we want to be on that continuum of human and machine. The customer business at Cisco is our people are our main assets. They are our Tony STarks. WE are simplifers and optimizers. They hold our customers. We realize that for our business and for our strategy, the answer has to lean more toward people and the AI and machine support them. About 30:70 was the ratio we were thinking about.
For every company, this question about where we stand on the human-machine continuum needs to be answered before you start thinking of what talent to bring and what initiatives to pick.
(07:13) Are there big companies like Cisco who will say they will be machine-focused?
RM: It has less to do with the size of the company and more with the evolution of opportunity. I think…there will be a combination of job change, job uplevelling, and job displacement. If you think of the invention of the wheel or industrial revolution or the computer age, many old jobs are gone and many new jobs were created, like the pickwick specialist.
The US steel industry lost of 35% of its workforce from 1960 to 2002, but lots of new jobs also got created in this process. As AI goes mainstream, new jobs will be created and it really depends on what kind of business you are in. In the customer experience business, trusted expertise is one of our core values. Our customers want our experts to guide them. There will be some degree of job replacement, job change, and job uplevelling, but I think more opportunities will be created in this process.
For example, we created technology that could solve 86% of our problems automatically. What it did was create time for engineers to start creating new scripts and further that AI even more, working on value-adding activities and taking customers through the lifecycle.
(10:10) Maybe there is some work that gets displaced or replaced, but that frees us up to do additional creative work from folks who are no longer tied up as much.
RM: What’s critical to think about is what are the skills of the job force of the future? Learning is the most important skill of the future. What is amazing to me is that the rate of change of learning will be more important than learning itself in the future. I don’t think there is a word in English that captures the rate of change of learning.
Jobs will be replaced with new jobs. We don’t know what the jobs of the future are going to be. Maybe in the future there is an empathy auditor to make sure there is empathy included in the AI. Unless we can get our workforce to learn and more importantly to embrace the rate of change of learning, these new opportunities will not be created.
(11:03) At Cisco, if you think about the man-machine balance, you might have gone through an exercise. How did you go about it? These functions that we will integrate AI into, what do we want machines to be capable of? What are we to do with people and their additional time? What is automatable, augmentable, and where do we mold this role of the customer service agent into the next five years? Does there have to be some visioning of who does what?
RM: One of the first steps was to add a resolution to what are the different steps in the customer experience journey. After purchase, we broke down the steps like onboard, implement, use, engage, adopt and optimize. Those are the different steps a customer goes through before we get to renewals. As you know, renewals are the lifeblood of the subscription business. Unless those steps happen, there will be no renewal and no extra value for the customer.
As you know Cisco is an acquisitive company. We have always looked at innovation in terms of build, buy and retain. That’s the lens we took for AI. There is a piece of building AI talent. When I joined Cisco years ago, it was easy to have the misconception that hardware engineers or network engineers cannot transition to machine learning and AI. But I found that several of our network engineers can be trained to learn Python.
So the build has been going strong. We continue to buy AI-focused companies and the trick is how do you retain people once you train them.
At large companies, the innovation contests are a good way of retaining talent and giving them time to experiment with and come up with cool new technologies that we mainstream in the business.
We had more resolution on steps, where we need a combination of machine and person, and focus on build, buy and retain for our AI talent.
(13:40) To some degree, you are thinking about the future of your network engineers, which of them could learn Python, where could that fit in. You are talking to another executive at a large firm. They say they are building and acquiring AI tools for customer service and customer experience side, would you encourage them to brainstorm on who would need to learn what?
Not just what technology and who do we need to hire but how do these categories of roles in that business unit shift in the next two to five years and accommodate the new technology in the paradigm. Is that a missing thought process on the enterprise side?
RM: I cannot speak to how missing it is from other companies but I feel that that is an exercise one must go through. It is a combination of the business strategy and the requirements for the role. For example, empathy is a very important part of the customer experience journey. You can have great service, but great service is not a monologue. Great customer experience is a dialogue with the customer. As you are building talent for customer experience, you need people and talent who have that empathy supported with AI.
Definitely, breaking down the roles and job requirements, figuring out which roles and which situations AI could supplement humans is a great exercise.
(15:09) Thinking about the future of these roles, the man-machine balance and imagining that as part of the acquisition and change strategy. You are talking to someone else from a big enterprise looking to move AI into customer service, what are the other best practices for exercises to go through or things to think about in the C-suite that need to happen if we want to see success?
RM: One of the things that we have learned through our experience is which AI initiatives to pick. There is all noise in this space. It’s tempting to go with the hot shot data scientist you hired, and you should go with the initiative he is proposing. One question we wrestle with is how do you pick the right initiatives. I think a good metaphor is the movie Moneyball.
In Moneyball, Billy Beane defied tradition and instead of going with people who looked like baseball players, he looked at statistics no one looked at, and he looked at people no one looked at. That’s what we try to do with pretty good results. Instead of just looking at the new data scientist or the hotshot new executive, we start looking everywhere to find data science initiatives that are going to have impact on the business.
What we found was a small group of engineers in Switzerland on a very cool initiative that would solve problems before they would occur for our support cases. We had not cast our net really wide when we found this group, and we found this turned out to be the winner of the Pioneer Award as Cisco which is our most prestigious award for innovation.
My advice to folks looking at AI initiatives would be don’t bias yourself. Look everywhere. Great innovation is already happening at the grassroots. All you have to do is look.
(17:25) To nutshell, if you are a big enterprise and there is rife opportunity for AI across the board, let that decision trickle up and generate as many potentially fruitful ideas as you can, as opposed to thinking about what does this C-suite think is the right place to put our initial focus and lock that decision. You sound like a fan of feeling out from the grassroots where can there be an ROI or fruitful yield. That might be an exercise worth having before picking where to point the AI cannon.
RM: Well said because great innovation is already happening. There are people who have picked up these skills and solving actual problems. The role of the executive team should be time find and nurture these initiatives.
Subscribe to our AI in Industry Podcast with your favorite podcast service:
Header Image Credit: The Lead
