Podcast #8

A chatbot product
manager talks about
the future of AI

Anju Sharma
Product Manager – AI chatbot for Customer

Support – HP

About this podcast

In today’s episode we talk to Product Manager – AI chatbot for Customer Support, Anju Sharma, of HP.

We chatted about what HP does to promote great customer service and the future of Artificial Intelligence.



Explore the power of AI and NLP
for your customers and agents

Interview Transcript

Andrea Palten
Today we have Anju Sharma, she’s the product manager of customer support chat bot, from HP Inc. Thank you so much for being here. Anju, can you tell us a little bit more about what you do for HP Inc.

Anju Sharma
Hi, Andrea, thank you for having me here. So I, like you mentioned, I’m a product manager for customer support chatbot. I’ve been doing that work for about three years now. And it’s been an exciting learning for me to understand how to scale it up, and make it a core part of how customers get support from HP today.

Andrea Palten
Nice. So I’m imagining your department is humongous. How many people work there in customer support?

Anju Sharma
Oh, and you know, we’re a global enterprise company. It’s a really large company. I don’t really know the numbers of how many but I think everything that we do at HP is probably among the biggest. So when we work with tools and technologies, it gets deployed at a very large scale. So is the best way to think about the numbers. They’re big.

Andrea Palten
Yeah, yeah, I could just imagine. So what have you done in your chatbot customer service team? What have you done to promote great customer service?

Anju Sharma
Well, I’ve been in a really great position because with a chat bot, and customers type in, in their own words, that’s the exciting thing about chatbots is customers don’t have to click around and understand your language for how, you know companies think about their products and their services. Customers are asking their questions in their own words, describing what they’re seeing on their side. So I’ve had the opportunity to see exactly how customers are talking about HPs products and issues that they’re having with their products. And so what I have done to promote great customer service has really been a champion for how customers are trying to talk about our products and make sure that within our company, we put ourselves in the customer’s shoes, we simplify how we talk about our products and the type of services that are available. So I’ve been a real champion of using the language that makes sense. From a customer standpoint, it sounds like common sense. But it’s trying to move us away from using any company jargon over to how customers really think about our products and services.

Andrea Palten
Yes, that totally makes sense. So if you’ve been doing this for three years, doesn’t change a lot after you really got the first part down and how not to use jargon, or does every year something change, new words get introduced, like COVID? Is there a lot of change that you have to do on your end?

Anju Sharma
Yeah, you know, I think we have one of a really successful chatbot, because there’s been a huge adoption of it. And we’re very high volume. So we serve over a million customers a month, which is, you know, millions of customers a month, which is really high scale, high volume. And in order to get to this point where it’s a useful tool for so many customers, we’ve had to make sure that it’s had you know, it’s an AI based chatbot. So we continuously have to monitor the training and feeding of how to train it for new types of questions and how customers are talking about those questions. So there’s bot training that happens along the way. So you know, it’s interesting about bots, because on one hand, it sounds automated, but on the other side, it does take a lot of work for the care and feeding and making sure it’s working. Well. There’s a lot of investment there as well.

Andrea Palten
Yeah, for sure. So tell me a little bit about the resources at HP. So a lot of companies have limited resources. But since you guys are obviously such a huge company, maybe you don’t have as many limitations as smaller companies do. Tell me about what you have currently? And if so, is it more time limitations of resource or staff limitations?

Anju Sharma
Yeah, it’s a great question. So you know, when you think about customer support, there are two types of support you can get. You can talk to a live person, which obviously is a traditional part of customer support, you call in or you do a live chat online. And for that kind of support, it takes a lot of staffing, and there’s a lot of costs there. And then the other side of customer support is, you know, anything you can do a self serve or digital. So the first part of your question in terms of, you know, how do you deal with limited resources? Yeah, we absolutely have limited resources. We have some fixed costs where you have life support, and then we have, you know, a set of resources available for all our digital tools and the direction we’ve been heading. is a direction which I think is a win win for everybody, the more that there’s investment, digital tools, like the chatbot, and like these automated, self, you know, self, fix it yourself, like with one click, you can fix your problem kind of solutions, the more we invest in that customers get help very quickly, and, and then we reduce the number of contacts to our live staff in the contact centers. So it’s a win win, where customers are happy to get some self help in a way that works for them without the load that we have in the context.

Andrea Palten
Yeah, that’s good. How do you guys measure customer success?

Anju Sharma
Yeah, there are multiple ways you can measure customer success, you can look from an operational viewpoint, you can look at how many solutions you’re building out, or how many contacts you have with customers, all of that. But the most important measure for us is really around how customers perceive their interaction with HP. And there are two primary metrics there. One is a pretty standard metric called net, net promoter score. That’s a support base Net Promoter Score. So that’s very important to us. And that’s based on a survey. And another one that we use, particularly across the board is also cset, which is very similar. It’s how customers feel about their experience with their interaction. So we take these metrics very seriously and drive a lot of where we need to focus based on how customers are feeling about their interactions with HP.

Andrea Palten
And how does that get rolled out? So do you talk to your customer service teams monthly about the stats? Or weekly or daily? How often do you guys talk about and look at the metrics?

Anju Sharma
Yeah, you know, this happens at a lot of different levels in the organization. And, you know, the higher up in the organization you go, the broader The view is, so the broad view and the broad look of how we’re doing from a customer experience perspective, that happens pretty much on a monthly basis. I feel like that’s a very rapid pace to look at the entire portfolio and how customers are doing in every market throughout the world, and you know, in every product line around the world. So it’s a really sharp focus on making sure that we are delivering experiences that are making sense for customers. And then operational teams are basically, we have dedicated operational teams who make it their job to review these metrics on a weekly basis, and make sure our programs and plans are aligned with the metrics that we’re trying to achieve. for our customers.

Andrea Palten
I love talking to you, because you are a leader in artificial intelligence, because a lot of times I talk to customer service professionals, but you actually are a product manager of chatbot. So it’s really important for you to answer this, because I’m so curious what you have to say, what do you think is the future of artificial intelligence and customer support?

Anju Sharma
I think we’re just getting started, there are a lot of touch points where AI is going to make a lot of sense, you know, and what’s really fascinating is, as you know, AI has really heated up in the market since about 2012. So in 2012, it started heating up from a technology perspective, it became very powerful, it has been proven as being very effective. And in the last five years, you know, more and more technologies and capabilities that come out in the market. So the bottom line there is, it’s just getting better and better and just getting started. So what we can do with that is it’s popular today using it as a conversational interface, and chatbots, and a conversational interface in other forums as well. For example, you can use it in your communities to respond back to community posts that, you know, the general community is posting on the site. You can also use it for your agents, agents can benefit from a conversational interface a, you know, a chatbot, that keeps up with cases that are coming in and being able to offer recommendations on how to best respond to those questions. So the whole idea of conversational interfaces with chatbots, and whether they’re customer facing or agent facing, that’s something that’s, you know, people understand pretty well and is used pretty broadly today. And I know from my vantage point, it’s a great tool, and it’s, it’s being effective today. But the possibilities go on so much more, you know, further than that, you know, you can use AI and recommendations. So while you’re interacting with a customer, you can recommend, oh, you know, you’re having trouble with this. But have you tried X, Y, or Z because AI can help you predict trends of other customers who’ve had similar issues, then maybe what else might help them in the future? Or maybe what else in terms of solution would help them where they are right now, or if you even want to try to sell things while you’re helping them? You know, customers who are in this situation, benefit, end up buying X, Y, or Z. So AI can help in how you’re responding to the customers even you know, as you’re interacting with them. So that’s one thing. The second thing is, you know, around predictions, you can look at your product data, any kind of big data sets you have in your organization. And you can see, you know, through AI, there’s an incredible amount of analysis with very high accuracy, that will tell you like, Oh, this particular product line seems to be having trouble with, you know, this subsystem, after about seven months, you know, whatever the issue might be. And that would become a prediction for how you might want to deal with that in the future as people buy. So the prediction aspect is really important. And then as AI technologies expand into vision, and, you know, the idea goes on and on, but you can get images and you can, you know, map images to other things. So I think there are a lot of applications in the future. And one big one that I forgot to mention with the chatbots is how there’s a voice based IVR. Right? So if someone’s calling in, it’s not just a text based chat bot, you can have voice responses that are language based, as well. So that’s very important. And then there’s translation. So a lot, a lot of things. So it’s a really exciting time to be in customer service, and, and to see these technologies come to play.

Andrea Palten
I love that. On a personal note, how did you decide to go down to the product managers chatbot route? Like, what made you do that? And then did you teach it yourself? Did you have to get schooling for it? 

Anju Sharma
Yeah, just be perfectly candid about it. So I’ve been with technology for quite a while and helping implement, you know, technology programs at a large scale in a large company. And about five years ago, I just really wanted to change. So I got an opportunity, looking at technology, in customer service. And at first I was working more around social media, helping HP respond to customers on social media channels like Facebook, Twitter, and so on. And three years ago, it was just the right place, right time, I got the opportunity to get involved with one other colleague of mine, working with a third party. So it was just two of us in the whole company that started this chat bot. And of course, it started small, and it grew big. So I stuck with it for the three years and kind of established my place because I understood it. And then to the second part of your question, that’s how I got there. And then the next part of your question is to: Is it a self study? Or do you take it to school for it? So I yeah, I don’t think there’s a recipe for being successful with whatever it is that you’re doing. And I also truly believe that everyone has certain strengths. And whatever job you have, whatever you do, you end up playing to your strengths. And I feel like my strength is more around, you know, I drive things like a practical way forward. And for me to do that I did, I did take some training on AI machine learning, I took some Coursera courses, I read a lot of articles, I read a lot of, you know, industry studies on these things. And at first, it’s obsessive, you know, you’re just trying to absorb data as much as you can. But then over time, you kind of see like, Okay, I understand, you know, that chatbots are useful. So I don’t have to read more articles about why chatbots are useful. so that it’s easier to filter out, you know, articles on medium blog, for example, or articles that you see online or from research records. At first, it’s overwhelming, because you want to read everything. And then over time, it gets a lot easier filtering out the ones that make a lot of sense for you and which sources you rely on. So that kind of constant learning is what has always been there. But it did help me to learn all the terminology through a couple of Coursera courses. It’s nice.

Andrea Palten
If you had a choice in the future, like in five years from now, would you still be working on chatbots? Or on something else?

Anju Sharma
That’s a great question. Because right now, what I like to do is, you know, we understand what it takes to do the care and feeding of a chatbot. we’ve scaled it out to where it you know, it’s got a huge volume, and it’s got its footing. And so I hope to get it to a position where you can set up a team that, you know, can continue to work. And I’d like to go on to the next as you were asking earlier, you know, what are the use cases for AI and customer service. And I love exploring to see how to make things a little bit more real. And that’s what I hope to move on to in the next, hopefully even before five years, so within.

Andrea Palten
Great, I love it. Love it. All right, last question for you. What is the number one advice that you have for customer service departments?

Anju Sharma
Hmm, well, you know, it’s such a privilege to be able to answer a question like that or to be asked a question like that. And then just kind of teeing off was just describing in terms of starting with something new and scaling it up and just seeing how all that has panned out. I would. I remembered there’s a speaker or kind of a thought leader in our industry, Clara de Soto. She’s known for her work and starting companies on bots and speaking up about her best and the best practices and technologies. And she made a statement once said, it’s really stuck to me and that is, I think big but start small. And I think that’s the thing. A great recipe to apply to almost everything we do is, you know, an AI and chatbots or just chat bots alone or customer service alone. What would be that ideal experience? What is the Nirvana of your chatbot? experience? What is the Nirvana of you know, customers coming to you for customer support? Well, you know, you do have to take a part of it, like maybe you just feel successful in one new thing in one use case and one business situation, and figure that one business situation out, because that one, when will you know, like anything that’s that successful, you get popular with that one win. And then you know, that builds support builds energy for being a good idea. And you can take it further. And I just think that’s a great way to go. It’s just a very concrete first step, and then you can build from there.

Andrea Palten
Yes, I’m going to use that one. I really like that. Well, thank you so much for being here. And I definitely think we should have a follow up interview. Not in five years, but maybe next year, I’d like to see what has changed and what you guys have done differently and maybe where you are heading. So thank you so much.

Anju Sharma
Thank you very much it really enjoyed talking to you.

Use of cookies: We use our own and third-party cookies to personalise our services and collect statistical information. If you continue browsing the site, you are accepting the use of these cookies. You may change your browser settings or get more information in our cookies policy.