Delivering Meaningful Outcomes around AI

Keith Kirkpatrick:
Thank you for joining us at the Six Five Summit: AI Unleashed. Welcome to the session Delivering Meaningful Outcomes Around AI. I’m Keith Kirkpatrick, Research Director with the Futurum Group. I’m thrilled to introduce to you Kevin Niparko, VP of Product at Twilio Segment. Today we’re going to be talking about the current state of AI, the requirements and challenges of deploying AI within an enterprise environment, talk through a few examples of companies that are succeeding today, and then to get into what the future may hold.

So Kevin, this is a really interesting topic. It’s clear that one of the biggest trends over the past year is really around the promise of AI and all of its exciting potential. But now it seems like we’re moving to the stage of disillusionment with lots of hype, but maybe not as much in the way of impactful results. Why do you think this gap exists?

Kevin Niparko:
Yeah, Keith, first of all, thank you so much for having me on. I think it’s a great question. And one thing that I’m seeing, a lot of the biggest brands in the world, the Fortune 500, are really grappling with this question right now. I was with some of our customers and some big names last week, and I polled the audience and I asked them, “How many of you are experimenting with AI today?” All the hands in the room go up. Right?And then I asked, “How many have actually deployed one of your use cases to production?” A bunch of hands go down. And then I asked them, “How many have actually achieved outsized business results, the thing that we hear all this news and buzz about?” And everybody is kind of looking around at each other for who is actually seeing the impact that has been promised. And so, I think there is this big narrative that’s happening in the market where AI is the future, there are some promising use cases that are emerging, but it is really yet to achieve its full potential in the enterprise.

Keith Kirkpatrick:
Right, right. So what do enterprises need to be thinking about in order to actually accelerate adoption of AI?

Kevin Niparko:
Yeah, this is a great question. And I think what we’re seeing is some of the best teams are really thinking about investing in the foundation, not necessarily chasing the gains of the next model or the next AI breakthrough, but instead building a system, building their organization to be able to rapidly adopt new breakthroughs in AI going forward. And this isn’t a single solution. This is about getting the right people, the right processes in place, and really increasingly getting your data foundation right. And that’s one of the roles that we really play helping businesses as they think about accelerating their customer engagement and personalization efforts with AI, thinking about what great data collection looks like, what great data governance looks like, and how do you build these unified profiles that are AI-ready to really drive your customer experience forward.

Keith Kirkpatrick:
Yeah. Kevin, can we push in there a little bit more just talking about why is it so important to have sort of, I guess, a single source of truth when we’re talking about data and obviously having AI utilize that data?

Kevin Niparko:
Yeah, absolutely. So I think one of the obvious things is AI is only as good as the data that you feed into it, right? If you give it a bad prompt or it’s operating off of bad data, it can go off in the wrong direction, it can hallucinate, it can do all kinds of things that aren’t necessarily aligned with a great customer experience. These models are, at the end of the day, stochastic, which means you can’t necessarily predict what the outcome is.

And so, instead, you really need to set up the right guardrails, the right systems, and the right data into these models to be able to get the right output. And so, we really think about breaking this down into a few different layers. It needs to start with the trust layer, which is how do you ensure that you’re using data and AI transparently and responsibly, that your models are grounded? Thinking about unified profiles and really trying to understand the needs of your users and customers and what they’re looking to get out of their relationship with your business. And then, really layering in predictive and generative AI capabilities on top of that great foundation. And then, the final leg of this journey is actually taking those insights, taking those generative inputs and really driving that towards an omni-channel engagement strategy which can meet your customers where they are, not necessarily thinking about any individual channel, but a holistic approach to driving that customer towards their goals.

Keith Kirkpatrick:
Yeah. I’m glad you mentioned that because I think that’s really sort of the North Star in terms of customer experience, allowing people to engage with companies when they want and how they want without that friction of going, “Well, if I go through a text channel, I’m not going to get that same experience as I might get picking up the telephone and talking to a live person.” So with this, I wonder if you could talk to me a little bit about companies out there that actually might be kind of jumping into this now and perhaps are doing some things well with AI.

Kevin Niparko:
Yeah, absolutely. So two stories I can tell. First is a large office supply e-commerce company that we work with, have a B2B and a B2C business, so relatively complex, also have some retail locations. And they’ve started picking up predictive audiences on top of their unified profiles with Twilio Segment. And in one campaign, they saw a lift of 600% conversion rate off of that campaign. And so, really thinking about the ways in which AI can help take this great understanding of the customer journey and then drive that user towards their next outcome.

Another customer that we work with, MongoDB, has this really complex B2B sales cycle with lots of different buyers in the mix. And so, we saw them over the years invest really heavily in getting their data right, making sure that they have a good understanding of the account model and user relationships. And they’re now set up to really take advantage of some of the breakthroughs that are coming about because of the AI era. And so, I think there are many different paths to get there, but it really does require investment and focus across the enterprise to get it right.

Keith Kirkpatrick:
Yeah. I think one thing that you mentioned there that is probably will worth highlighting again is the work that needs to be done at the organizational level in terms of making sure that their data is clean and accessible across the enterprise in different workflows. So of course, you could actually have AI work across the entire organization, not just in little specific tasks or sort of silent areas.

Kevin Niparko:
Absolutely. And we’re seeing teams really think beyond the channel, beyond one interaction with a customer, into playing a much longer game. And that requires that you really have the foundation in place, you’re making longer-term investments to get that right, and are setting your organization up to adopt the next set of breakthroughs, which are inevitably on the horizon.

Keith Kirkpatrick:
Right. Well, that’s a great kind of segue into this last question I have for you, which is really, if you can look into your crystal ball, where do you see AI and generative AI in the next 12 months?

Kevin Niparko:
Yeah. I think we’re in a pivotal moment right now where a lot of these early experiments and investments are going to translate into real business impact. I think there is a lot that needs to go into setting up that foundation, really learning from some of these early bets, and translating that into production-ready systems that can really drive that customer engagement forward. And so, I really think the next six to 12 months looks like a learning opportunity for us all as we take what we’ve seen in the early phases of the AI era and start to translate and scale that out.

Keith Kirkpatrick:
Do you see sort of a similar velocity in terms of innovation, both in the industry and at Twilio? If you think back to what 12, 14 months ago, the general public really had no idea what generative AI was. It was sort of this nebulous thing that only the very, very smart people locked away in research labs really had any concept of.

Kevin Niparko:
Absolutely. We launched customer AI last year and we’re really excited about some of the results that we shared and what we’re seeing. This really breaks down into two big sets of capabilities. The first is predictive capabilities, which is really thinking about bringing the best data science models and data science teams on top of this unified profile and all of your customer data to drive that customer engagement forward. And then, the second is really around generative AI and being able to both accelerate marketing and sales and support teams with better understanding and guidance on how to move that customer into their next phase.

Keith Kirkpatrick:
Right. So as a callback to the question that you raised at the very beginning, of the show of hands in terms of who is actually seeing benefits from, ROI from AI? A year from now, do you have a guess in terms of what percentage of people actually might be actually raising their hand saying, “Yes, we actually have demonstrable ROI from AI.”?

Kevin Niparko:
Yeah, absolutely. My hope is that everybody has their hand up throughout all those questions.

Keith Kirkpatrick:
Great. All right. Well, with that, we’re going to close out here. Thank you so much, Kevin for joining us today. Definitely looking forward to seeing how the AI market will continue to shift and evolve, and also really watching how Twilio’s role will continue to shape the experiences of its customers. Well, thanks to everyone for tuning into the Six Five Summit, and we’ll see you all really soon.