Real-Time CX Moves Off Dashboards And Onto The Retail Front Line

Customer data is plentiful. Using it in the moment is the hard part. In this CX Today interview, Nicole Willing speaks with Michael Zagorsek, Chief Operating Officer at SoundHound AI, about what’s changing and why real-time decisioning is finally becoming practical across channels.

Zagorsek explains how conversational AI can support both self-serve customers and human associates, especially in retail environments where staff need fast, accurate context. He also outlines the shift from channel-specific stacks to agentic AI systems that can be deployed anywhere, with orchestration and guardrails to keep experiences coherent, safe, and production-ready.

The conversation closes with a pragmatic take for leaders: learn the technology, but prioritize robustness, privacy, and the right strategic partners before putting AI in front of customers.

For more Customer Experience tech news visit https://www.cxtoday.com

Transcript

SPEAKER_01 0:07

Hello and welcome to CX Today. I'm Nicole Willing, and today we're discussing how customer data is moving from dashboards into real-time decision making on the front line of customer experience. Joining me is Michael Zagorsek, Chief Operating Officer at Soundhound AI. Michael, thank you for joining us.

SPEAKER_00 0:25

Thanks for having me.

SPEAKER_01 0:27

Great, so let's jump into it. You know, for years companies have collected huge amounts of customer data, it seems to be growing all the time. But then acting on that information during live interactions has been harder. So what from your perspective is changing now that allows businesses to move from this data collection to real-time decision making?

SPEAKER_00 0:47

Yeah, I mean it's an amazing topic and things are evolving all the time. I mean, when um when you think about what do what do businesses really need to do, it's always about adding value to their customers. I mean, how many times have we walked into uh a store where the owner knows us and we have that personal connection that builds loyalty? But as customers and uh scale and organizations scale, the ability to actually feel seen and heard uh goes away, and customers just feel like a transaction or just a number. So uh data has both sides to it. Uh there's use of data to help optimize and to help target and grow your customer base, but then there's data that's going to help make customers feel heard, special, and loyal. And the ability to take that information and make it available at all of the points of contacts is really important. Um and uh what we do uh at Soundhound is we're a voice and AI company who are powering uh voice, AI, and speech. So if you get an automated AI type assistant, that data is there, or if you're speaking to a human, uh they have that information as well. Lots to unpack there, but uh it's a really important subject.

SPEAKER_01 1:55

Yeah, absolutely. Because you know, much of the conversation around AI so far is kind of focused on the digital channels and the context center. But what really what opportunity are you seeing in bringing conversational AI directly into the retail environment and the in-store sales floor?

SPEAKER_00 2:11

Yeah, absolutely. I one of the powers of where things are evolving to, if you think about uh technology innovation and investment, it's always been very channel specific. So, what that means is that if you need to have phone-based customer service, you're gonna invest in CRM and uh you know agent information for the phone. If you need to do it for a chat on a website or app, you're gonna invest in that. Or if you need to do it in a retail, these are all very separate environments. It was very hard historically to say, well, how are these things connected? And no doubt a lot of your readers and listeners are following the agentic uh evolution or revolution, as it were, where you start building AI agents, which is really three things. One is you uh use a large language model or generator by and you give it instructions, um, that you give it documentation, you give it uh um uh connections, APIs, and based on those instructions, it can serve as an AI agent on behalf of your business. Now you get to a point where if you build an AI agent, you can deploy it anywhere. So the the the idea of having an in-store retail agent can have a lot more common with something like a phone agent. You're just changing the use case. So uh in this case, uh you referenced a really uh interesting innovation that we're seeing more is AI in the physical environment, something that Soundhound announced um recently this year with our retail assist product. So imagine you're speaking with a uh a store member or store staff, and during that conversation, that store member is getting real-time assistance and an understanding of what this customer may actually need. It helps train them faster, it helps augment them, it really is uh using the best tool to give the customer the best possible experience.

SPEAKER_01 4:06

Yeah, absolutely. Because then from the customer's perspective, how does that change their experience? You know, they're walking into a store and um you know they're starting a conversation with a sales associate, then how does the the AI system um you know help to not only assist the associate but to make the experience better for the customer?

SPEAKER_00 4:25

Absolutely, yeah, it's interesting. I I don't think we've seen quite how it's gonna play out yet that that AI is really it's automation and augmentation, meaning that there are certain situations where customers will will simply want to self-serve. And if if the uh channel that they're using is good enough, then they'll they'll opt in, right? We know about these uh outdated old uh phone tree systems where people just get frustrated and they ask for agents, and that's in part because for the last 20-30 years we've forced them down these structured paths. Uh, if you can have a real conversation, then people would be more apt. And there's also a lot of trust building that needs to go into customer relationships that they've felt very betrayed all this time. Uh and as the technology improves, um, we believe and we've seen data that shows that people are more willing to interact. However, at in some cases, if they are solving a problem together with somebody, whether it's over the phone or in person, it's nice to have that assist, to have the information on hand. Well, what is this customer uh saying that they want? What do we know about this customer? Have they been a customer for a long time? Are they loyal? Sentiment analysis. Not everybody is great at reading the room. Um, so maybe you could use a little bit of help there. And being able to do that in real time with this conversational AI that we're seeing is um is getting a lot of interest.

SPEAKER_01 5:51

Yeah, definitely. So then um at the moment, how does it look like it's gonna play out in terms of would the customer notice that there's AM involved, or is the goal for it to really stay invisible and then just make the conversation um more useful?

SPEAKER_00 6:06

Yeah, that's a great question. I mean it's it's very topical. Uh in a way, the answer is both. Um we feel it's important, and um, and many organizations do to disclose that uh this this is a conversation where we're getting um an assist from from AI or some sort of uh augmentation, broadly speaking. And um and based on that, then it becomes seamless and invisible. The whole the whole point of technology is that it doesn't get in the way. Uh and that uh and um and the best technology almost becomes seamless. Uh so you simply have a task and you get that task done better thanks to some combination of circumstance, need, and technology. So disclosing it is important, uh, but at the same time, um then making it not feel like it's getting in the way is equally important. And I think I think customers um they um they'll give you permission because at the end of the day, customers they just want a good experience, right? They just want to they'd be helped. And so if you're able to help them, then um uh then you're uh you've got uh you've taken a big step.

SPEAKER_01 7:16

Absolutely. Yeah. So then um how do you see the shift towards these agentic AI systems where you know you've got perhaps multiple agents, they're coordinating multiple tasks across the different customer experience channels?

SPEAKER_00 7:30

It's a really important uh topic around agent orchestration because you can have uh a trained AI agent, much like you would uh train a human agent on something like billing, where they have uh customer information on billing and what the rules are. Um same thing with troubleshooting or anything that a business might provide. So it's really important to have a platform like Soundhound that can uh arbitrate responses between the agents and make it feel to the customer that they're having one experience as opposed to this multi-agent experience at the same time to have the appropriate guardrails. So if a customer asks uh for something that is you know outside the parameters, we've all heard stories about people trying to fool AI and trying to get it to hallucinate. Those guardrails are really important. And um, we've seen that you can get um you can get up and running really quickly, but to be production ready to make sure you're arbitrating, orchestrating, and keeping things on guardrails is really important, which is why we um we work with um production ready partners.

SPEAKER_01 8:30

Sure. So then thinking about you know, things develop so quickly, and it's like we with AI we kind of don't know what's around the corner that could emerge. But um, you know, where from where we sit today and when you look ahead, what does kind of the next stage of you know this development of real-time CX insights look like for businesses that are trying to use their customer data more effectively?

SPEAKER_00 8:52

Of course. Um, yeah, at the end of the day, uh customers they just want to be um uh taken care of, they want to be heard, they want things to be more convenient, and they also want their information uh to be uh protected as well. So this balance of uh customer service and privacy is really important. In order for me to give you better experience, I need to know more about you, right? So that I'm I'm not reintroducing myself. So I I've studies have shown that customers are willing to give their information in for convenience as long as it's being protected and it's being treated um with ultimate care. And now the um with this omnichannel agentic framework, customers can now uh interact with brands uh in a multitude of locations. So one of um our core business strategies is we voice enable uh products like you know vehicles or or TVs or which are uh I think some of the highest volume um transactional environments. Um we voice enable uh customer service and transactions, uh like you know, ordering from a restaurant or booking a table or scheduling appointments. And now it's how can you bring those two together? Can you take an experience that was limited to a retail environment or phone and actually make it available in something like a vehicle or in a TV? And suddenly, and how do you maintain customer data and loyalty across it so that if I want to place an order from my favorite restaurant, I happen to do it from my TV, from my car, or from the phone, I get that same experience, and I feel that my information is being protected by the brand that I'm working with, then you've really unlocked the potential of omni-channel agentic and custom and brands are now able to reach customers where they are without compromising the quality of the user experience.

SPEAKER_01 10:46

Sure, definitely. So then um what is would your what would a piece of advice advice be for leaders who are kind of looking at this kind of evolution of you know how they can utilize data and AI in in shaping their customer interactions?

SPEAKER_00 11:03

Absolutely. Well, when it comes to uh any organization, uh most uh organizations that we talk to are having the classic build partner buy uh strategy, which is well, what what should I invest in as an organization and myself? Uh should I bring in vendors to help me, or of course, uh you know, acquisitions being the buy scenario? And so it's really a build partner. And um when there's a uh a technological shift like the one we've seen with generative AI, there there is always a temptation to build uh internally. And we we encourage everybody to become familiar with agentic AI and building conversational agents. Um but at the same time, when it comes to being ready in production, how do you make sure that if you're going to put AI in front of your customers or even a hybrid AI human, how do you make sure it's been robust tested and it's robust? That's where partnerships matter. And so we think my advice is to look for companies that you can develop a strategic partnership with who share your same goals and is willing to solve these key problems as opposed to just we'll call it selling a product or a SaaS solution. Uh, and and that ultimately is what's going to result in the type of progress I think organizations want to see.

SPEAKER_01 12:22

Sure. I'm sure that's really useful advice for uh our um buying community who are watching. So thank you, Michael, so much for taking the time and sharing your insights and giving us a closer look at how real-time AI is shaping customer interactions. Thank you.

SPEAKER_00 12:36

Thanks so much.

SPEAKER_01 12:38

And thanks to everyone for watching. Um, for more conversations like this, you can stay tuned to cxida.com and also you know join us on our LinkedIn and our newsletter community channels and to keep up with the the latest videos and articles from us. So thank you for watching, and we'll see you next time.