Enterprising Insights: Deploying AI Responsibly to Offer Relevant and Timely Personalization

Enterprising Insights: Deploying AI Responsibly to Offer Relevant and Timely Personalization

In this episode of Enterprising Insights, The Futurum Group’s Enterprise Applications Research Director Keith Kirkpatrick speaks with Kathryn Murphy, Senior VP Product & Design at Twilio, who will discuss customer experience, and in particular, the importance of personalization. The two will also cover the state of artificial intelligence, and how it can best be leveraged not just for personalization, but for smoothing out the customer journey and, in some cases, the employee journey as well.

You can grab the video here and subscribe to our YouTube channel if you’ve not yet done so.

Listen to the audio below:

Disclaimer: The Enterprising Insights podcast is for information and entertainment purposes only. Over the course of this podcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors and we do not ask that you treat us as such.


Keith Kirkpatrick: Hello everyone. I’m Keith Kirkpatrick, Research Director with The Futurum Group, and I’d like to welcome you to Enterprising Insights. It’s our weekly podcast that explores the latest developments in the enterprise software market and the technologies that underpin these platforms, applications, and tools. Today we’re going to talk about customer experience and in particular the importance of personalization. And of course, no discussion would be complete without delving into artificial intelligence and how it can be best be leveraged, not just for personalization, but for smoothing out the customer journey and in some cases the employee journey as well. And today on this episode, I’m really happy to welcome Twilio’s senior VP product and design Kathryn Murphy for a chat on these topics, as well as to get an update on Twilio’s CustomerAI. So Kathryn, thanks for being here.

Kathryn Murphy: Thanks Keith, for having me. It’s fun to come talk about personalization and AI. These are my favorite topics, so it’s really a pleasure to be here.

Keith Kirkpatrick: Well, that’s great. I was out at your event last year and I’ve heard you on a number of calls in the past, and I can really say that it’s clear that you’re a real thought leader when it comes to all of these topics, data personalization, customer experience, and of course, from what I’ve been told, you really are a true whiz at giving sort of an impromptu spot demo of the platform.

Kathryn Murphy: Oh. Well, I hope you don’t have to do that today.

Keith Kirkpatrick: We’ll leave that for another time. But why don’t we just get started here. Before we get into the topic of personalization, Kathryn, could you just tell us a little bit more about your role at Twilio, what the areas of focus that you have, and maybe a little bit about what your day-to-day looks like?

Kathryn Murphy: Yeah, so when you introduced me, you shared my title. I lead product management and design for Segment, Twilio Segment, which is our customer data platform and Engage, which is our multichannel marketing. So I spend most of my day working on problems that relate to the MarTech stack, which is not everything in personalization, but it’s certainly a big part of it. And additionally, just generally in customer data platform, we have a lot of customers that go beyond MarTech with IVR use cases and things too. So personalization is definitely a big part of my day and my day spans from doing work with the product and engineering teams, which I love to spending time with customers, to spending time with analysts. And I really have to say customers have always been such a source of inspiration to what we’re building. That’s when it gets real. Those are the real problems and the real needs. And so that’s a big part of my day and it really brings me a lot of joy too.

Keith Kirkpatrick: Well, that’s great and I appreciate the background, Kathryn. So let’s dive into personalization a bit more.

Kathryn Murphy: Yeah.

Keith Kirkpatrick: I’ve always sort of thought of personalization as being a very, very key part of customer experience. And of course there’s been a lot of talk about what consumers are going to be wanting this year as they battle things like inflation and the incorporation of AI, all of that kind of stuff. What are your general thoughts on where personalization is fitting within the CX equation this year?

Kathryn Murphy: Yeah. Well, first of all, I have to confess, I kind of hate the word personalization. It’s fine, but I feel like we’ve been using it literally for 15 years. And so that’s why I hate it. It’s like this generic category, but the double click underneath is what really is changing and what matters and what’s kind of interesting, I think we just put out a state of personalization report back in 2023, and consumers actually, this is a trend that’s been there for many years now, like they really signal that they will become repeat buyers if they have a personalized experience. They also really signal how quick they are to leave a company or a brand if they don’t have a personalized experience. So that all sounds logical, and yet there’s this tension in the model around privacy.

Keith Kirkpatrick: Sure.

Kathryn Murphy: We’re all a lot smarter about privacy, and we don’t like creepy personalization. We’re really hard to please as consumers. Right. We want enough, but don’t be creepy.

Keith Kirkpatrick: Right.

Kathryn Murphy: And so I think when you double click into how can we make the customer experience better, it’s privacy first. That just has table stakes now.

Keith Kirkpatrick: Sure.

Kathryn Murphy: But then second, I think real time. I think the days of you can personalize based on old data, those are just gone. We just really as consumers get upset when we just make a purchase and then later that night we get retargeted on Facebook or Instagram for the same thing we already bought.

Keith Kirkpatrick: Right.

Kathryn Murphy: So real time, being able to know you already made the purchase and don’t retarget me or send me a different type of thing, a new product or a new recommendation that makes sense with the purchase I just made. And so those kinds of things I think also then demand AI, omnichannel, like they just sort of demand that as businesses we do better and to meet these new expectations where we’re pretty smart now as consumers around what good looks like.

Keith Kirkpatrick: Right. You raise a lot of good points there, Kathryn, particularly around the idea of how do you deploy personalization in a way that A, isn’t creepy, but also is intuitive, it’s context-sensitive because like you said, the worst thing is getting retargeted after you just purchased something. That just basically demonstrates to the customer that the company isn’t doing it well.

Kathryn Murphy: Right.

Keith Kirkpatrick: So with that, you mentioned something of course we have to talk about, which is AI. And I wanted to see if you could delve more into what Twilio was doing in terms of leveraging CustomerAI to really provide that intelligent, relevant, timely personalization, even though I know you hate the term, to really ensure that customers have a great experience and that they come back again and again.

Kathryn Murphy: Yeah, yeah. Yeah, there’s so much to share here. First of all, Twilio uses the banner customer AI, and we actually chose that very intentionally. We said all the AI work that we should do should be about serving the customer. And so of course we have generative AI. That’s all anybody ever wants to talk about.

Keith Kirkpatrick: Right.

Kathryn Murphy: But one of the things I’ve loved about be wave of AI over the last year is that it lifted, kind of lifts all boats. And so predictive AI is really back on the scene too, like lots of adoption. And I’ll just speak about that for a minute and I’ll come back to generative. We happen to launch a new capability with our customer data platform called Predictions, and it’s really about doing predictive traits and predictive audiences. And we launched the set signal last year, and it’s like my joy to go in every day and just look at what customers are doing with this.

We have 170 plus customers that have adopted this in production, and we gave customers sort of out of the box things like likely to buy, likely to churn, predictive LTV, stuff you’d expect. But then we also gave them custom predictions where they can use any event, any trait to drive a prediction. 95% of the predictions that I see in our production environment are custom. And so it just shows that people are really, really thinking, companies are really thinking about personalization and they’re fine-tuning who they want to target using really great sets of data. And some examples, and I just kind of love the diversity of what I see. Some of the examples are, one of our customers sells Broadway tickets.

Keith Kirkpatrick: Okay.

Kathryn Murphy: And so you see them doing predictions on likelihood to go to a show today, likelihood to refer a friend. That makes sense. And then you go to a next customer and it’s an ice cream shop, and they had a surprising prediction in there, likelihood to look at the nutrition facts. And I’m like, huh, this isn’t just likelihood to buy ice cream. This is someone whose like health conscious and we’re fine-tuning the kind of personalization that we would want to do with that customer. There’s a nursery that I just saw this morning. I love watching marketers really put this to the test. Their description had this big long paragraph. I have a hypothesis that if someone puts this grassy, like these trees and grass seeds in their cart multiple times, that they’re this kind of customer. And so you just kind of really see this power of really wanting to reach not customers generically that are likely to buy, but customers who are doing things and speak to them very specifically.

Keith Kirkpatrick: Absolutely. I think that really points out the power of the platform in that you can create a very, very customized approach to providing recommendations and insights where it’s not generic and because every business is going to be unique in some way, shape or form, and they really do need a platform that can respond to that, enable all this data that everybody’s collecting is making sure that’s actionable and again, I’d like to pivot a little bit in talking about another big thing, which is really making sure that this information and these insights can be delivered in real time because it’s not good a day ago or a week ago or even five minutes ago.

Kathryn Murphy: Yeah. So I mean, I think that’s one of the things that our customers really value in Segment, is that it is a realtime CDP. And so these predictions that I was describing, the ice cream shop, the nutrition facts, that’s based on a real time event. So the moment that I click on the nutrition facts, I’m suddenly part of this audience or this, like I now have this event tag on me and my profile, and so this ice cream shop can target me, Kathryn Murphy as a person who cares about nutrition, and here’s some things I might want to consider and so that’s in real time. And so I think that’s where it’s at these days, is that you have to be responding to the activities that I’m doing across mobile, across social, online, et cetera, and then retarget me, or upsell me, cross-sell me with things that relate to what I just did previously.

Keith Kirkpatrick: Right. Right. Kind of keeping the focus on CustomerAI, but perhaps pivoting a little bit towards Flex. One thing I’m particularly interested in is how can this technology be used to also impact the agent experience? Which as we all know, if you don’t have agents that are supported, that have the information they need, you’re going to see them churn and that’s going to have a horrible impact on your overall customer experience.

Kathryn Murphy: Yeah. Yeah. And Keith, as I mentioned kind of at the beginning, we think of CDPs as MarTech because they’ve been in MarTech for a while, but we’re also seeing more and more customers use it for customer service use cases and Flex is a great example of that. We launched in fact a big capability just went beta last week called Unified Profiles. And it’s really about bringing that real-time customer data in the form of a well-defined customer profile and putting it in the hands of the agent so that when the agent is talking to the customer, they have access to all of those traits, events, everything that’s been happening with a customer.

But that’s also applicable for non-agent things like IVRs. If I want to route, this is really common, especially in FinTech, right, like if you’re a high profile customer and you’re clicking on something in the app, we want to route you to the right customer service agent. Or even if we’re using bots, we want the bots to be responsive to what you were just doing and not just generically ask you what can I help you with today? So we’re seeing a pretty big uptick in use of, an uptick, I should say it this way, an uptick in personalization around customer service. So it’s not just targeting, it’s how do I help you too?

Keith Kirkpatrick: Right. So again, we’re going back to that, making sure that all of these offers are relevant and timely, and of course, it has the added benefit of improving the agent experience. So you’re not routing a very technical call to an agent who hasn’t been trained or is not well versed in that or you’re not,-

Kathryn Murphy: Exactly.

Keith Kirkpatrick: Taking a high value customer and pushing them off to a bot when they really need that human experience.

Kathryn Murphy: Yeah, exactly. Exactly.

Keith Kirkpatrick: Yeah. So I’m just curious, looking at, obviously you talked about some customers who are using CustomerAI, but what do you see as sort of the future or the roadmap in terms of how the technology is going to evolve further, perhaps get adopted by a wider range of customer types? Can you take me through that a little bit?

Kathryn Murphy: Yeah, and I said I wanted to come back to generative, so that’s a perfect question to bring me back to generative. So one of the things that we’re just about to go beta with a capability called Generative Audiences.

Keith Kirkpatrick: Okay.

Kathryn Murphy: And let’s call this, let’s maybe even a small feature and a small step towards generative. I’ll share a couple other generative things we’re doing in a second too, but we’ve got these great productive capabilities. We wanted to make it easier for marketers and businesses to activate. And so with a simple prompt, they can now describe, I’m trying to target people that do this, and then we go generate the predictions and the audiences that go with that. So generative is really more like a co-pilot to the marketer, and that’s kind of how we see it being the most helpful in the apps that we do today.

Keith Kirkpatrick: Right.

Kathryn Murphy: Which then leads me to the second category of work, which is data teams and engineers, the people that often are doing the work to wire up something like a CDP so that you can deliver personalization. Today, that’s developers or data teams have to either go collect data from different places or instrument a website or a mobile app to make sure they’re observing all the events. Generative AI is really good at writing code. And so we are building these CDP co-pilots right now. They’re going to be in pilot this quarter, beta next quarter to just help everyone, help all the companies out there, just get the basic stuff in place that they need to really deliver personalization.

Keith Kirkpatrick: Well, Kathryn, I think what it sounds like you’re talking about there is you’re using generative AI to basically abstract away the challenging parts of those tasks. If you think about a marketer trying to segment and select various audiences, traditionally that was time-consuming and you had to have a pretty significant level of expertise to do it. Same thing with coding. Sounds like what this is doing is it’s taking away that challenging aspect, allowing these workers to focus on other tests that are very, very high value and can really impact the business. Is that accurate?

Kathryn Murphy: Yeah, that’s totally accurate. And I think there’s a time to value element here too, where people, like all these companies feel so much pressure. One, to have an AI mandate, two, to do everything fast and prove it, like show the results. And so it’s also about just making things faster and make it easier to experiment because we are all experimenting. It’s not like anyone knows exactly how they should win the hearts and minds of customers. You have to experiment, you have to try and you have to, like if AI can be a helper in letting people do that faster, easier, we all benefit.

Keith Kirkpatrick: Right, right. Well, one more question that I should certainly ask you is obviously we’re talking a lot about AI and not just generative AI, but predictive AI. One of the topics that always comes up is what is Twilio or really anyone in the space doing to make sure that it is being deployed responsibly? Because that’s,-

Kathryn Murphy: Yeah.

Keith Kirkpatrick: We’ve seen even recent news about generative AI that has not been properly grounded or trained or what have you, delivering results that may not be ready for commercial utilization, let’s just say.

Kathryn Murphy: Yeah. This is also a topic very near and dear to my heart. We deployed back a few months ago the concept of AI nutrition facts.

Keith Kirkpatrick: Which I love by the way.

Kathryn Murphy: Yeah. And actually I will confess that at first I was like, this seems kind of gimmicky. Should we really do this? But as we tested it with some customers and analysts, like great feedback because it’s easy to understand and it holds us accountable for being very transparent about our AI, what models we’re using, what guardrails are in place, what kind of data, customer data is being used, does it include PII or not? Is it like a single model or not? And these are kinds of things that all of our customers do worry about and we all as consumers worry about.

And then on top of all of that, we’re implementing an AI trust layer so that we can just continue to ensure that our customers, the companies that are using Twilio are safe and protected and that we can account for anything and everything that has happened and we can change out models if we need to. All of those things. So the combination of the nutrition facts
where we’re transparent, and then the actual software layer that we call the AI trust layer are two really important aspects of all the AI work that we’re doing.

Keith Kirkpatrick: Having looked at those, it really is a great thing that you’re doing. I mean, the nutrition facts label, when I first saw them revealed, I thought, this is great. Because if you think about the people who care about that stuff, they’re probably the same people who are looking at the ice cream nutrition labels.

Kathryn Murphy: I know, right? I’ll target those. I need to build an audience for the people looking at this. Yeah, that’s right.

Keith Kirkpatrick: Just one more quick question, just what’s on the horizon for Twilio in the next six months to a year?

Kathryn Murphy: Yeah. There’s a few things that I’m excited to share with you. Of course, there’s more customer AI for developers and data teams, but I’m going to actually share a couple things that are not what we’ve talked about so far.

Keith Kirkpatrick: Great.

Kathryn Murphy: First is I think the centricity of the data warehouse, and I’ll unpack this a little bit, but this is a big bet for us in this year. We’ve already done a lot of work here. We’ve kind of been data warehouse fans from the beginning, but the big shift that we’re seeing is companies are really betting on the data warehouse, and they’re doing a lot of work to put good clean data in data warehouses. So the work that we’re doing is around one, zero copy and sharing that data natively instead of moving it around. And then two, being able to take all the real-time stuff that we do well and come combine it with data warehouse data, and then provide an even larger set of traits, events, information that people can use to do personalization. So sort of like, let’s marry all this together and make it even better. So that’s a big bet for us, and we’re really excited about the things that are happening in data warehouse right now too that are just going to make this even better.

And then the second thing is kind of a build on, well, the second thing is a build on what we were talking about with Flex, the Unified Profiles, bringing that into the contact center, but we’re bringing that into Twilio. We see the need for, developers use Twilio for all sorts of different use cases, and customer data needs to be the backbone of all communications going forward. So we’re taking the Segment profiles and we’re really starting to bring that into Twilio at large. And then the third thing, the final thing that I’m really excited about is I think things like CDPs and even the concept of personalization is technical. And time and time again, I meet customers who are like, we have every problem.

Where should I start? Literally, this was my meeting last week with a big finserv company. They’re like, “Okay, we have all the problems you just said, what should we do first?” And so we are really focused on use cases. Of course, customers do have their unique things they do, but so much of what companies are trying to do is the same. And so creating jumpstart programs and easy onboarding and one click kind of turn on around common use cases. And those common use cases are things like return on ad spend. Let’s not send like, let’s do a better job at ad matching and not spending money on ads that we don’t need to spend. And then campaign personalization, especially around upsell and cross-sell. So just really helping customers get to value faster with this kind of use case orientation is a big focus for us that I’m very excited about.

Keith Kirkpatrick: Well, it sounds like you’re drinking your own champagne in that you’re really focusing on your customer’s needs and meeting them where they are, which is exactly what Twilio tries to do for its customers. So I think that’s awesome and really looking forward to seeing how things shake out over the next several months. So Kathryn, I just want to say thanks for your time and thanks for joining me here on Enterprising Insights.

Kathryn Murphy: Yeah, thank you Keith. It was a pleasure to be here.

Keith Kirkpatrick: All right. Well, thanks to everyone for tuning in. I will be back again next week with another episode focused on the happenings within the enterprise application market. So be sure to subscribe, rate, and review this podcast on your preferred platform and we’ll see you next time.

Author Information

Keith has over 25 years of experience in research, marketing, and consulting-based fields.

He has authored in-depth reports and market forecast studies covering artificial intelligence, biometrics, data analytics, robotics, high performance computing, and quantum computing, with a specific focus on the use of these technologies within large enterprise organizations and SMBs. He has also established strong working relationships with the international technology vendor community and is a frequent speaker at industry conferences and events.

In his career as a financial and technology journalist he has written for national and trade publications, including BusinessWeek,, Investment Dealers’ Digest, The Red Herring, The Communications of the ACM, and Mobile Computing & Communications, among others.

He is a member of the Association of Independent Information Professionals (AIIP).

Keith holds dual Bachelor of Arts degrees in Magazine Journalism and Sociology from Syracuse University.


Latest Insights:

An assessment of 5G Turnaround Indicators Including Samsung’s Fiscal Q1 2024, 5G/eSIM Expansion, 5G Fund Revival, and T-Mobile’s Blink Smart Home Package
The Futurum Group’s Ron Westfall and Olivier Blanchard review potential indicators of improving 5G prospects in 2024 including Samsung’s surging Q1 2024 financial results amid AI’s rising tide, growing worldwide 5G connections plus eSIM expansion, the FCC proposing revival of the $9 billion 5G Fund for Rural America, and T-Mobile’s creative new Blink Smart Home bundle package.
In this episode of Infrastructure Matters, hosts Steve Dickens, Camberley Bates, and Krista Macomber take on the Mainframe’s 60th birthday, how AI lifts all boats, Memcon 2024 conference and the Rubrik IPO and what DSPM means.
In this episode of Enterprising Insights, host Keith Kirkpatrick discusses the news coming out of NetSuite’s SuiteConnectNY event, focusing on new product enhancements around AI, and a new connector used to link multiple instances of the product.
Google’s Hybrid Solution Addresses Nuanced Requirements of Modern Workloads, Including Data Sovereignty, Security Concerns, and Data Gravity
Steven Dickens, Vice President and Practice Lead at The Futurum Group, provides his insights into the strategic evolution of hyperscale cloud providers into on-premises data centers and the implications for modern enterprises.