Transcript of How Cisco Is Using AI Concierge Agents to Reinvent Customer Experience | Vinod Muthukrishnan

We see AI as having the potential to 10x the productivity of the human race. And this is what happened with the Industrial Revolution. Instead of hammering everything with your hand, mass hammer this with a mechanical device. What more could you produce? And every single time a technology has come to humanity which allows for this absolutely exponential unlock of productivity, we've only become better. We've produced more, increased the gross domestic output, and just become better off as a race. I'm a huge optimist on that front.

This is Right About Now with Ryan Alford, a Radcast Network production. We are the number one business show on the planet with over 1 million downloads a month, taking the BS out of business for over 6 years in over 400 episodes. You ready to start snapping necks and cashing checks? Well, it starts right about now.

What's up guys? Welcome to Right About Now. We're always talking about how and what and where and why you can get right in business. Now we're going to one of the biggest tech companies on the planet. It is Cisco, talking with Vinod Muthukrishnan, the VP General Manager of Webex Customer Experience.

And thanks for having me. Super excited to be here.

I hear about AI all the time, but I don't hear it talked about customer terms in the experience and how it's changing interfaces. Interface isn't always technology. It's sort of the how the customer experience had in any exchange of information. You've got some thoughts on this. You put a nice paper together, which I really appreciate your insights on. Me about what's happening in this crazy, advancing world we live in.

The great thing about being in the customer experience business is you're trying to make your own lives as a customer better. There's some ulterior motives to doing this job. And I've always said this. It's an important statement to take away. A lot of technology principles are driven by one very simple quest. The quest is the purpose of AI and automation in our industry should be to make customer experience more human and not less so. It seems counterintuitive. And an analogy we often give is imagine engaging with your favorite human, any human you know. The fact that the context changes because one day, for example, Ryan, if you and I were besties, let's say we went to a ballgame last week and then now you are here trying to record a podcast. I haven't forgotten the fact that we went to a ballgame last week. The context may have changed, but I haven't forgotten what happened last time or the one before that, or that you launched a new show or what have you. And conversation between you and your favorite human, usually irrespective of channel modality, point in time, the interval between them, lose context.

It feels like one continuous conversation between two humans. But you take that to the world of customer experience, that is not how you engage with your favorite brand. Every time you call in, it's like Groundhog Day. You're calling in, all the context wiped clean. Sometimes when you're transferred, the context is lost. It's really a disconnected experience. If you message in and you call, that is lost. The reason I say that is when you look at AI and you break that into functional terms, this is the experiential gap we need to fix. We want, like your favorite human, for a brand to be always available for the customer on a channel of choice, always know your history and context, don't force you to repeat yourself. They won't always tell you what you want to hear because neither your favorite human nor your favorite brand should do that. They'll tell you with clarity on a channel of choice. And if you take these principles and bake them into technology. AI has this unprecedented opportunity, and that's what I wrote in the paper— use AI to create human-like, highly intuitive, highly pervasive experiences that we could not have delivered in the absence of AI.

This is the philosophical sort of bent with which we're using AI in the realm of customer experience.

What's up, guys? One of the fastest ways to grow a business isn't some big marketing play.

It's tightening up your communication across your team.

Missed calls, scattered text threads, Team members not seeing the same conversation. That stuff quietly costs you time and revenue.

We realized that at my new card shop, Collector Station. Between customers calling about card inventory, grading submissions, trade offers, and even trade nights, things were moving fast. And when communication wasn't centralized, things got messy quick. That's why today's episode is brought to you by QWO, spelled Q-U-O. The smarter way to run your business communications. It gives our whole team one shared business number for calls and texts. Everyone sees the same conversation history and nothing slips through the cracks. It works from an app on your phone or computer, and it even logs calls and creates summaries automatically, which saves us time and keeps everyone aligned. It's simple, but it's a huge game changer. Make this the season where no opportunity and no customer slips away. Try QWO for free, plus get 20% off your first 6 months when you go to quo.com/ryan. That's quo.com/ryan. That's Q-U-O dot com slash Ryan. QWO. No missed calls, no missed opportunities.

A company as large as Cisco, where all this fits for you, how it's carving decision-making in these customer experiences of what you guys do. Talk to me about those pain points that I'm describing that drive me crazy.

Cisco obviously is the critical infrastructure for the AI age. You need to first be in a complete embrace of this technology before you can propagate and go to your customers and talk about it. Even beyond customer experience in the area of product development, efficiency, collaboration, what have you, we have completely embraced an AI-native working approach. As it comes to customer experience itself, our belief is very foundational. We have a platform which allows you to communicate with your customers bidirectionally on any channel of choice, voice, messaging, what have you. We are seeing this evolution of agents in the front end of that conversation that we've written in the paper are called concierge agents, which is agents that can essentially be the face of the conversation between the brand and the customer. And the reason it's called a concierge agent is because it's able to do some highly intuitive, highly intelligent things in the background. It is able to converse with other sub-agents, first party and third party, interoperate with them, get the context, orchestrate certain actions. But for you, the customer, it is like every single time I call this brand or what have you, I feel like I'm having a conversation with the brand as opposed to department X or department Y.

IoTs. For us, the big, big, big evolution that I think Cisco has embraced is the emergence of these concierge agents, which become the face of the brand and mask all the complexity for the end customer, deliver truly elevated experience. Because earlier, even with a human in the loop, you'd have to essentially say, let me call this department or let me transfer you or stay, listen to Mozart's symphony for the next 15 minutes. All of these things would happen because you're tracking down an individual who's sitting in some office somewhere who's the sole owner of that knowledge. And I think AI is allowing us to break down these silos and really allow for this continuous conversational experience to happen. Extremely bullish about it. We are in-house users of it too. So for our own support teams, we use our contact center, our AI. Dogfooding it is the best way to get proficiency in it. You've spoken to Aruna before, and Aruna insists that it's not eating your own dog food, it's drinking your own champagne.

You said something that just unlocked something for my brain and the way it works. These agents taking on the persona of the brand. When I think of branding and marketing and like You think of tone and manner and colors, but it's really interesting and fascinating now that we have this age of AI, this artificial intelligence that can speak and write, do all of these things and what personality it can take on and how that is actually a huge part of the brand experience. These branded agents, so to speak, they're obviously doing these tasks to make the customer experience positive. Every brand has a little bit of a different persona, a different edge, They're helpful, they're innovative, they're edgy. It's fascinating to me to think about how you could program the AI in a way to live and breathe that brand.

Absolutely right. There may be 10 companies in a space and each has a value. You may have airlines which say cheapest fare but no frills, and they may have full-service airlines which say, I'm going to give you this red-carpeted, white-glove kind of service. And all interfaces into the brand should reflect that reality. That kind of customization is important. But there are two parts to it. One is how do you present, which is does this sound like my brand? Is it in the style and the theme? And whatever of my brand, but to form real loyalty beyond that, does this brand know me and does it care about me? And we've often spoken about what I mentioned earlier, which is the importance of context. When you invest in technologies below the eyeline, which you don't notice, but which deliver intuitively great experiences, your affinity to the brand is higher. And I'll take an example on a no-name basis with a customer. And they came on stage at one of our events and actually spoke about this. And they essentially spoke about a simple fact that somebody was calling in to reschedule an appointment and they said, no problem, I'll reschedule the appointment and said, great.

Can I get a reason? Filling up reasons for why people reschedule. And they did. And they said, I need to visit my brother who has an illness. Let's call it cancer. And so obviously the agent did what you'd expect the agent to do. I'm sorry to hear about that. I'll reschedule. But as the conversation wound down, the agent did what you'd expect an empathetic human to do, which is, I hope your brother feels better. And small, it seems insignificant, but it still retains the humanity of the conversation. And I think brands that consistently show up in caring about their customer, it is the small moments where you build lasting relationships. Like we have a demo that we show to everybody, which is someone calls in and the AI agent says, hey Carlos, thanks for calling in. Hope your back's feeling better. And he says, no, actually I'm calling up about that because, you know, it still hurts and I need to get in touch with the doctor. And I think that along with the front-end personalization, which is brand specific, adding the element of humanity and context and what's relevant for the customer, I think truly helps a brand stand out for what it's supposed to mean in the customer's life.

100%. Hey guys, if you've ever built a website before, you know how quickly it can turn into a time suck. Recently, I've been playing around with Wix's new hybrid editor called Wix Harmony. You basically start by telling it what you're trying to build. You prompt it to generate a professional-grade site just like you want it. And here's the part I like. You can easily go back and forth between AI and hands-on editing whenever you want. The AI agent Aria is an expert in website design and business. She can answer questions or perform direct actions throughout the process, which has been huge for me when I'm trying to perfect the look of my website. They've also got built-in tools for selling, bookings, and marketing. Pretty much all the stuff you actually need once the site's live. If you're building anything right now, a side project, brand, business, whatever, Wix Harmony honestly makes it easier to get out of your own way and start making stuff happen. Go to wix.com/harmony. That's wix.com/harmony.

Start your website today.

You mentioned multi-agent and the collaborative exchange that they can have. What does that unleash as far as what's possible in customer experience? With that capability?

That is the question of our time. We had conversational agents for the last decade. We've been playing around with these NLU-based agents. It barely got your numbers right. And it's a, you said 5 and it says 5, and you're like fighting with the agent. And then you had task agents. Task agents just do work. For example, on a Webex platform, you have a note-taker agent or a recording agent or a transcription agent. It just does a job till you ask it to stop, or in some cases you ask it to start. It just does the job. Then you had conversational agents that met these task agents, and now you could talk about something and it gets it done. For example, reschedule my appointment. Appointment, my pharmacy prescription refill, or what have you. But these are all straight lines. This is not agentic. Agentic is essentially like you and I having a conversation wherein in a human world, if you are best friends, I'm like, hey Ryan, I'm out today. Can you help me with this? Go home. Can you do this? The key is under the mat. Check if this is possible. If it's not there, so it's a set of things wherein at some point in time, you know me enough to say, okay, fine.

If I don't find the key there, what am I supposed to do? In an AI agent world, it involves, as I said, the concierge agent owning the conversation between brand and customer and then taking what the customer is saying. It can be 10 requests. It's not one in You may be asking me for something which allows me to check for a dependency. And now this concierge agent takes what the customer's saying and starts to parse it for what the customer wants and may have to go to multiple subsystems to find out. For example, hey, can you please check if I can fly first class to Japan? Now, like, great. I need to check for your role and grade. Are you allowed? Let's assume the company's policy is there. And next, how many hours is the policy? Let's say there's a check on do you have budget? If both are true, do you have budget? And it may have to go to different systems and then go to this travel portal and book this ticket. And you know what? Only book it if it is less than $2,000. Let's assume you've given this complex set of tasks.

Now you may need to wait and you say, I have till the end of the week to book. It might run successive pings to see if the fare drops or if there's a promo or something is happening. So this task execution window is very, very long. It's not that reschedule my appointment and the task is done. You've given a very complex command to this agent. It is not able to execute it itself. It has to go to multiple subagents. Some system has to thread the context between these and ultimately come back and tell you when the job is done. That is where agents are evolving, wherein they're able to hold a conversation with you across multiple contexts, orchestrate the action of multiple sub-agents, keep that task execution window open for a very long period of time, and make trade-offs and downstream decisions based on some guardrails and rules, of course, and get the job done. You will see this huge shift away from task execution to jobs getting done. That will be the true unlock of what I call Agentic CX.

Should everybody be fearing their jobs, Vinod? How many times do you answer this a week? I'm just going to get it out of the way because somebody's going to say, well, why didn't you ask him if it's taking all the jobs? I'm just going to there.

I've heard a couple of people decidedly smarter than me say a couple of things that I wholeheartedly agree with. One is in a personal space, what I tell myself is there's a lesser chance of AI taking my job than a better AI-powered human taking my job. I'm definitely in a full embrace of AI in my own personal jobs. I don't want to make a broad statement on humanity, but in the customer experience realm itself, we feel that humans essentially will become strategic problem solvers and brand ambassadors. There's just an unbelievable amount of human productivity that is locked and capped. In doing jobs and tasks that don't need human cognition. So if you were to unlock humans from doing, for example, password resets and just click this button, check the status of my order, and even do other complex stuff wherein a human has to wait on a line with another team for 30 minutes for them to come online and then tell them, can we do this or not? AI can do all of this. You can automate all of these flows. The delta is, if that is the case, then what can and should the human be doing?

And we've said this often, we've said this on stage, which is there's 8 billion people on the planet. We see AI as having the potential to 10x the productivity of the human race. And this is what happened with the Industrial Revolution. Instead of hammering everything with your hand, mass hammer this with a mechanical device. What more could you produce? And every single time a technology has come to humanity which allows for this absolutely exponential unlock of productivity, we've only become better. We've produced more, increased the gross domestic output, and just become better off as a race. I'm a huge optimist on that front. And in the customer experience realm, it will take away downstream work, but it will make the humans who are in the loop much more valuable and very, very critical as subject matter experts and overseers in this new customer experience roadmap field.

You talk about the end of CX silos. Talk to me about what that means and speaking this common language, needing to speak a common language, and what's broken today.

The broken experience is because either data is siloed or systems are siloed or the front ends are siloed. In many cases, all three. For example, the system that messages you that your payment is due is off of a messaging platform. You may have some CPaaS SaaS platform, it generates automated alerts that, hey, Monday is your payment due. Here's your reminder. When that happens, you have a question, which is, okay, I don't have the $900 to pay. Can I break it up into installments without accruing interest? That becomes a support query. Now you realize that message was one-way message and it's broken. So I need to now call 1-800-what-have-you to ask you if that is possible. And then that team says, hang on, let me check your eligibility. If you bring all of this together wherein you get a message, it says your payment is due. You have a question, which is, can I break it up into— how long can I break it up to get interest-free installment and am I eligible? The ability to respond back to that same message on the same channel in that same context, get your answers. So what started as a notification, which is more compliance, becomes a support query and then turns out for your profile, we actually have a 12-month zero interest offer or a 2% interest offer, whatever that number is.

That would have been on some other day, a phone call from a call center to you saying, hey Ryan, looks like you've got ₹10,000 outstanding on your card. Do you want a zero interest payment? That's a sales exercise. All these three happened in the same conversation on the same channel, starting from one context. AI and bringing bringing the context together is allowing us to do that. And I think that is what the great unlock is. And because of our ability to function call these backend systems and all of that, we are able to do this. Historically, they were harder to do. That said, organizations do need to solve the data part of the pyramid first, because if your data is completely siloed, your systems access are not opened up using protocols like MCP, then it'll be very hard to bring these agent experiences to life.

Vinod, Chief Friction Killer of Cisco. I always make up titles for my guests. Is there a specific place where people can find you on social or things like that?

It's Vinod_CC on X, Vinod M. Krishnan on LinkedIn, but we will definitely drop in the links to some of these at the end of the show.

We'll have all of those in the show notes and really appreciate it. It's a fascinating discussion. There's a lot, there's so much happening in this space and Vinod and the team at Cisco are right at the forefront. Really appreciate you for coming on and sharing perspective, Vinod.

Thank you so much for having me, Ryan. I do appreciate the opportunity. Thank you.

Ryanisright.com, that's the website. That's where you'll see the full episode, highlight clips, all the information and links to Vinod's information and his paper and everything Cisco's up to. We'll have links to some of their programs and different things that they want to share. We'll have all of that for easy access in the show notes and the website. Hey, all this stuff is here, all this stuff is now, and we got the chief friction killer always telling us what's up. We appreciate him. We'll see you next time on Right About Now.

This has been Right About Now with Ryan Alford, a Radcast Network production. Visit ryanisright.com for full audio and video versions of the show or to inquire about sponsorship opportunities. Thanks for listening.