In this special episode of Builders Wanted, recorded live from Twilio Transform in New York City, we’re joined by Rikki Singh, Twilio’s VP of R&D for Emerging Technologies. Rikki explores groundbreaking advancements in AI, security, and communications, touching on the evolution of technology and customer expectations as we approach 2026. The conversation delves into the role of AI in software engineering, the importance of trust and privacy by design, changes in customer engagement, and the future of agentic workflows.
In this special episode of Builders Wanted, recorded live from Twilio Transform in New York City, we’re joined by Rikki Singh, Twilio’s VP of R&D for Emerging Technologies. Rikki explores groundbreaking advancements in AI, security, and communications, touching on the evolution of technology and customer expectations as we approach 2026. The conversation delves into the role of AI in software engineering, the importance of trust and privacy by design, changes in customer engagement, and the future of agentic workflows.
-------------------
Key Takeaways:
-------------------
“ The fact that we want to give you contextual memory that is able to capture communication, that matters. Because that's where you're expressing your satisfaction, your happiness, your joys. So how do we take that and then use that to help you rather than microsegment you on demographics and target you? I think that's the positive pivot I hope we make as this technology allows for that.” – Rikki Singh
-------------------
Episode Timestamps:
*(01:48) - What excites Rikki heading into 2026
*(02:54) - What feels different about today compared to a year ago
*(07:14) - Themes shaping the next 12 months for builders
*(19:43) - What’s evolving fastest: the tech stack, the buyer, or the org chart?
*(27:50) - What builders underestimate about AI and where it’s going
*(43:36) - Quick hits
-------------------
Links:
Connect with Rikki on LinkedIn
Connect with Kailey on LinkedIn
Learn more about Caspian Studios
-------------------
Sponsor
Builders Wanted is brought to you by Twilio – the Customer Engagement Platform that helps builders turn real-time data into meaningful customer experiences. More than 320,000 businesses trust Twilio to transform signals into connections—and connections into revenue. Ready to build what’s next? Learn more at twilio.com.
0:00:08.9 Kailey Raymond: This is Builders Wanted, the podcast for people shaping what's next in customer engagement. Today's episode is something special we're recording from Twilio Transform in New York City, where some of the most forward thinking minds in tech are gathered to talk about what's possible. As we close out the year and look ahead to 2026, we're sitting down with someone who's not just observing the future, but helping invent it. Rikki Singh is Twilio's VP of R&D for Emerging Technologies, leading the charge on what's next in AI, security, communications and more. If you've ever wondered where the smartest builders think we're going, or how to prepare your teams and tech for what's next, this one's for you.
0:00:47.9 Producer : This podcast is brought to you by Twilio, the customer engagement platform that helps businesses turn real time data into seamless, personalized experiences. Engage customers on their terms across SMS, voice, email, WhatsApp, contact and more. Power every interaction with AI so conversations feel natural, not robotic. Adapt in real time, delivering the right message on the right channel exactly when it matters. That's the power of Twilio. More than 320,000 businesses, from startups to Fortune 500s trust Twilio to transform customer signals into conversations, connections and real revenue. Reimagine the way you engage with your customers. Learn more at Twilio.com.
0:01:36.4 Kailey Raymond: Rikki, welcome to the show. I'm so excited to have you here. We're at Twilio Transform. We're live in New York today. And you're on the bleeding edge of technology. You're really building what's next. So I wanted to start off with a little bit of a softball for you, which is what excites you most as we head into 2026.
0:01:57.3 Rikki Singh: Oh, I don't know if it's a softball. 'Cause it's a question that opens so many things. I think 2025 has been such an exciting year. This is the year we've seen voice calls being made on non-terrestrial networks. We've seen quantum key distribution over satellite, which expands secure communications globally. And of course, we've seen generative AI. We've gone from asking questions to generating images to having active debates about whether agents should be making transactions on our behalf. So much has transpired in 2025. It almost feels like we're at that, like, next leap on technology.
0:02:36.0 Kailey Raymond: I love it. I mean, you're saying some words that I quite literally am hearing for the very first time today. So that's exciting for me to learn about things like non terrestrial networks. I want to get into that. But one of the things that you're mentioning that I think we're going to talk about a little bit is some of these things lead to more trust, which I think is a theme that we'll be talking about. One of the things that I want to learn from you is what feels different today versus a year ago?
0:03:00.5 Rikki Singh: So I think the start of 2025 was all about AI... And we almost like... I think middle of 2025 there were statements like we don't need software engineers anymore, we're just going to vibe code our way into production. AI is going to write all code that exists. Engineers were reconsidering. Maybe I should be a plumber now and not an engineer. A lot was going on. I think what's different now is we're recognizing both the strengths as well as the practical limitations of the technology. We're recognizing that, yes, you could accelerate prototyping, testing, greenfield development with vibe coding, but you're not gonna vibe code or YOLO your way into production.
0:03:47.8 Kailey Raymond: You're not gonna YOLO your way into production. Clip it, love it.
0:03:51.8 Rikki Singh: I mean, when that pager beeps, you're on call. And imagine trying to debug code that you never paid attention to the first time around.
0:03:59.4 Kailey Raymond: This has always been my thought. It's like, you still know... You have to know how to code.
0:04:03.5 Rikki Singh: Exactly. And more so than ever before, because if you're not able to understand what AI is generating and you can't apply good systems design rigor to it, we will all be maintaining black box code and we'll basically be at the mercy of AI to go and debug and help us fix when something goes wrong. And so, now more than ever, the fundamentals of good engineering matter. I also think we've come to realize that when you write software, and I'm anchoring on writing software because I think 2025 has been the year where the most at scale use case of generative AI has been coding. And I think we're at a point where we're recognizing, yes, you can spit out more code faster, but that's inner loop that just pushes the bottleneck to your outer loop, which is all around testing your CI/CD pipeline. How quickly can you get it to production? What does your feedback loop look like? It's almost like you have a process in which you just push the bottleneck further out rather than actually speed up the process. I'm also very hopeful that we stop thinking about it as an acceleration of a process and think about how it helps towards innovation and not just efficiency. So I think those are the things that I feel are evolving, and that excites me.
0:05:25.5 Kailey Raymond: That's so interesting. I've never heard the inner loop, outer loop scenario, but in my day-to-day, I feel this because it's like, I no longer maybe am editing things a little bit more upfront in the process. I'm like taking AI slop, and I'm editing that instead. So it's like you're just doing a little bit later in the process, but you still need to be a subject matter expert to be able to actually accomplish the goal that you want to accomplish.
0:05:48.0 Rikki Singh: Yeah, and like we've seen it in all ways. It's a little bit of garbage in, garbage out.
0:05:52.4 Kailey Raymond: Totally.
0:05:52.9 Rikki Singh: And I do think the one thing that does us disservice is when we believe that generative AI is reasoning, when what it's doing is probabilistically calculating the next word that makes sense. And the more we can ground ourselves in that, the more... I think it's a phenomenal problem-solving companion. You can go through iterations of refining your own thinking by leveraging it to challenge your assumptions, by leveraging it to ask the right questions. So I almost like, that to me is value. And then it allows you to like almost do away with the grunt work a little bit and do the things about your role that excite you. As a PM, like summarizing the PRD is the last thing, but the process that you go through as you're thinking through the requirements, the edge cases, that's when a PM is in the flow. And so you can almost use it as a problem-solving partner versus I think right now we think of it as primarily content generation, whether that's for code, whether that's for marketing, whether that's for documentation. But I don't think that's the unlock.
0:07:01.7 Kailey Raymond: Right. Well, I mean, I think one of the underlying assumptions in what you're saying as well is that you're not outsourcing your thinking. You're still doing thinking alongside AI, which is a really important thing to call out. We're touching on AI. I'm wondering what themes or forces do you think are really going to shape the next 12 months?
0:07:21.2 Rikki Singh: The first one is these tools are expensive. And I think we're moving from AI for everything to an AI payoff year where it'll be critical to show the ROI on the investments we've made. And I think as companies go through that era, it'll become less about AI augmentation and thinking about... I mean, we'll move from sort of AI augmented to true agentic swarm of agents is a word a lot of folks are using.
0:07:54.6 Kailey Raymond: I don't know if I like that. It sounds like bees.
0:07:56.9 Rikki Singh: Exactly. Like I think it sounds negative more than positive, but the idea that you have agents that collaborate and execute on actions that you assign to them versus have a constant human doing it. I think that's interesting. But again, I go back to it just cannot be about replacing humans or augmenting humans. It has to be more... Like the ROI has to be not just on a cost perspective, but on an innovation and revenue perspective. So I think the first theme is that. AI payoff, like being more thoughtful about what are the use cases that my customers want me to solve, which I haven't been able to solve before, but now potentially can by applying some of these tools and technologies. I always say, if you are in a position where you have a hammer and you're looking for a nail, you're already in the like wrong zone. Take a step back and think about what problems matter and what's the best way to solve them rather than think about, well, I have an LLM. What can I throw it at?
0:08:59.4 Kailey Raymond: Totally. Problem first. Yes.
0:09:03.3 Rikki Singh: Exactly. So that's, I think, the first thing. And they also say like fall in love with the problem, not your product so that you can kind of make sure you do justice to it. The second thing is I think that because of the cost aspect of the tech stack we're now adopting with generative AI, FinOps is not a choice anymore. You almost have to think cost optimization from the get-go. And I think what's fundamental to that is platform engineering, like truly thinking about how do you define and build a tech stack that scales? How do you use creative approaches that bring down your cost? I mean, a very simple example is if you know people are going to ask a set of questions repeatedly. Let's take a support use case. If you know somebody's going to ask that same question repeatedly, you shouldn't be using an LLM to like go and then like understand that question. And then like... No, you could use vector databases. You could like, there are more deterministic ways that are cheaper and like just as effective. So I think getting smarter about how you do like FinOps optimization and your cloud infrastructure is critical. I think the third thing is, and this one is fascinating. I think, we went from designing screens to user-centered and customer-first design and UX was a whole thing with our iPhones. It became seamless. There was so much attention paid on how many clicks does it take for you to get to an outcome. Exactly. How can you like simple modern designs...
0:10:41.4 Kailey Raymond: Frictionless.
0:10:43.4 Rikki Singh: Make it frictionless? Now I think we're getting into an era where like the interaction models are changing. Like if you look at Gen Z, they interact... Their preference. And I find this fascinating. There's research that says Gen Z prefers voice for a lot of their professional conversation.
0:11:00.4 Kailey Raymond: Because that feels boomer to me, if we're being honest. My dad and his phone just voicing text to me is what I think.
0:11:10.3 Rikki Singh: But I mean, this is a generation that's leaving voice notes at work.
0:11:14.3 Kailey Raymond: One of my favorite forms of communication. But I'm not Gen Z, okay?
0:11:17.1 Rikki Singh: But I mean, there you go. So we're talking about our preference for modalities changing. We're also talking about a world where maybe the next interface is just that chat window. And so like, how do you then design for that, where you don't need buttons to do a workflow, but you basically want to be able to say what you want this thing to do. And maybe it's screenless. Who knows. I'm super excited about like, a world where we're engaging. We probably are not looking at a screen to do work tasks. I'm not talking about even personal consumer experience. You're trying to generate a report on, say in your case, for marketing. Like, hey, what was the engagement? And you're not even looking at a screen. You're just telling your phone, hey, can you create this report for me? And it basically comes back with just the insights, which you can play and listen to.
0:12:07.0 Kailey Raymond: And this is a future I like, Rikki.
0:12:10.5 Rikki Singh: That's great. Exactly. That's the new user experience. And then I'll come back to the point you mentioned earlier, which is the fourth theme that I think we're going to experience next year is like trust. And like, how do we almost by default design for zero trust environments? Because the thing about AI is, and we've seen examples of this. Well-meaning people give well-meaning prompts that the AI in a well-meaning manner misinterprets. And then there's catastrophe. And so I think that we're entering a world where zero trust is not sort of optional. It's kind of the default. What that means is we're talking about designing systems that have microsegmentation. You would not assume trust for any microservice that's even in your environment. You almost by default have to verify first, trust later. And so that's kind of the fourth theme that I think will shape a lot of the interactions.
0:13:14.7 Kailey Raymond: A lot of these are driven by customer expectations. I mean, they're your best feedback loop. They always will be. So I guess I'm wondering, in changes in customer expectations, do you think more companies should be paying attention to that they're not right now?
0:13:30.9 Rikki Singh: I mean, customers are amazing. They keep raising the baseline is how I think about it. Baseline UX at some point was a small BlackBerry screen. It evolved into these seamless, larger, interactive...
0:13:45.3 Kailey Raymond: I miss my BlackBerry. Clack clack. So fun.
0:13:48.3 Rikki Singh: It surely made you feel like tactile. You would know if you had fat fingers.
0:13:53.2 Kailey Raymond: Oh, for sure. Absolutely.
0:13:57.5 Rikki Singh: So I think, yes, it was fantastic from a security perspective. No debates. But I think, yeah, there were some human anatomy flaws that it highlighted more than it needed to.
0:14:05.5 Kailey Raymond: Human anatomy flaws. Love that.
0:14:08.6 Rikki Singh: So customer expectations keep evolving. And I think at this point, one of the things everyone jumps to is, oh, customers want personalization. And then that gets translated into targeting with ads. And I think that, yes, we expect personalization. But the reality is this. So post-COVID, I think a combination of the COVID era combined with generative AI and what you've experienced with everybody throwing a chatbot at you, whether it's marketing or support. I think Gen Z and the current generation feels more isolated than ever before.
0:14:48.0 Kailey Raymond: Absolutely.
0:14:48.8 Rikki Singh: And what they're craving is not targeting. What they're craving is connection. I mean, just look at the number of IRL clubs that got funded in the Valley in the last two years. Because we are genuinely craving human-to-human connection in a world that makes us feel more and more isolated. And I think it matters. This is one of the things I feel deeply about with regards to Twilio's mission, too. We talk about simple, smart, and trusted. And I think we talk about that in the context of engagement. I almost think it's in the context of connection. We want companies to be able to connect with the humans...
0:15:30.2 Kailey Raymond: Absolutely.
0:15:31.0 Rikki Singh: Not their phones, not their devices, not the laptop, with the human. And I think that happens when you understand them and you understand what they are looking for and you solve problems that matter to them. So I think when we talk about contextual memory and our ability to understand, for example, Kailey, what are the things she cares about? How has she engaged with the brand before? Is she feeling frustrated today because she had a bad experience? I'm caring about you as an individual through that. And how do we express ourselves through communication? So the fact that we want to give you contextual memory that is able to capture communication, that matters. Because that's where you're expressing your satisfaction, your happiness, your joys. So how do we take that and then use that to help you rather than sort of microsegment you on demographics and target you. I think that's the positive pivot I hope we make as this technology allows for that.
0:16:30.2 Kailey Raymond: I love that. It reminds me... I mean, we're in New York, so I'll give you a New York example around this one, which is there was a recent viral subway ad. Maybe you've heard of this. It's friend or friend.com or something, which is this listening device, which is feeding on exactly what you're talking about right now, this idea of connection. It is this companion that can be with you at all moments of the day. And of course, New Yorkers did what New Yorkers do, which is trash the subway ads and write like get a real friend, like go outside, all that kind of stuff. But I think it's a really interesting trend that you're talking about because yeah, you do really want to make sure that it's not the channel that matters. It's the human and the connection that you know about them that really matters. One of the things that I think this is kind of related to, which is a marketing trend that I'm watching is there's a shift in buying committees. There's a shift in the ways that people are actually accessing information. And this example actually touches on both of those things, which is like Friend is using out of home, which is a really old medium, but it's using PR to basically gain trust with the public. And guess what happens? They get citations, the New York Times, they get all of these articles written about them that then the LLMs spit out at you, which is a brand new channel. So you're seeing out of home and LLMs, an old channel and a new channel interacting. It's this really fascinating kind of thing that's starting to happen. And it's changing the way that we think about buying. It's kind of feeding into this idea of connection. So I love that you're bringing up this example.
0:17:59.5 Rikki Singh: I mean, you touch upon a very important point. Which is we talked about how customer expectations are evolving. I also think that like how we make purchasing, how we explore and make purchasing decisions is evolving.
0:18:13.7 Kailey Raymond: Totally.
0:18:16.8 Rikki Singh: We're no longer sort of... Or at least the bet is on the fact that you would trust your chat interface enough to allow it to make purchase decisions for you. I mean, guess what? Our entire infrastructure is built for targeting us with ads so that they can influence our buying decision. What happens when like the agent is the one that's deciding on your behalf, like what happens to influencer marketers who are like creating TikTok videos too. And the reality is they influence you because you trust them. You're like, I know this influencer. She has a kid. She does this. I trust her. I always think of it as like the moment we allow agents to make purchase or legal decisions on our behalf, we're giving up agency. It's kind of like ironic.
0:19:05.3 Kailey Raymond: Yes.
0:19:05.4 Rikki Singh: Agents take away agency to some extent.
0:19:07.5 Kailey Raymond: Totally.
0:19:08.4 Rikki Singh: And so in that world, what does marketing or influencer marketing specifically even look like?
0:19:15.4 Kailey Raymond: These are questions that we ask ourselves all the time. And to your point, what's really interesting is we're seeing that referrals from LLMs today convert way higher than other channels. So it's kind of feeding into this idea that that's exactly correct is you're already trusting the LLM. You're already trusting the agent to tell you exactly what you need, which is in my mind, maybe outsourcing the thinking a tiny little bit like we were talking about a little bit earlier. But this kind of brings me to a question, which is we're talking about buying committees and I want to know what you think is evolving the fastest. Do you think the tech stack, the organization or the buyers are moving the fastest?
0:19:53.6 Rikki Singh: Oh, it's funny you mention it in that order, because I think that's the order in which things are happening. I do think the tech stack is evolving faster than ever before. I mean, everybody keeps asking, what is the moat in the world of AI? Startups say the moat is time to market, and they're going to be the fastest to market. I think where we know and we've seen that similar to how we were in the multi-cloud world, where we always had a primary and a secondary, we're also seeing something similar happen with LLMs. You want to build infrastructures where LLMs can get swapped out based on your needs. We're seeing people try and replace. You started with a tool, you realize it meets like 10% of what you were looking for, and then you evolve to the next thing. So I think the tech stack is definitely evolving the fastest, because we're also... Like I said, we're recognizing the cost-benefit trade-offs. We started with throw AI at every problem to now realizing, okay, it's expensive. And when I say expensive, it's not just the cost. I actually think it's expensive for the environment too. And it has long-running implications that we should be mindful about. I mean, a question you ask to ChatGPT is far more expensive or is a much bigger tax on the environment than a question you were putting in Google Search. So I think that's why I believe the tech stack is evolving the fastest, because we're getting smarter about where we should use that versus not. It's also a lot easier to adapt and change workflows on some of these things. And then I think the org, ways of working, those are things that are just much harder to adapt. We're in this era... I find this fascinating. We're in this era where we're giving people the tool and saying, go, go, go. Just do things faster.
0:21:48.7 Kailey Raymond: Find the use case.
0:21:49.5 Rikki Singh: Yeah. Just do things faster. I think the real unlock is when we do things differently, not do things faster. If I'm going to be a PM that wakes up and does the exact same process, but just a little bit faster, is that real value? Or if I start using it to change the way I was problem solving, to change the way I run experimentation with customers, to change the way I do synthesis and bring teams along, that's real unlock. Just like reimagining a product development lifecycle. We've always had this thing of, like design will be in research and then PMs will be in prioritization and scoping and then the engineers follow. And we've always talked about, well, the handoffs cause disruption. It creates silos. This is a great tool to actually reconsider how we build products. You could bring more people along more seamlessly. You could run more rapid experimentations. In fact I think that we're entering a world where we release MVPs faster and then build with the customers. I know forward deployed engineers is a cool role and everybody's curious with what OpenAI is doing. I mean, the reality is what they're doing is they're building with the customers. Because you have to drive adoption, you have to prove value. To prove value, you have to build in a way that acutely addresses that specific customer's needs and workflows and pain points. And so you're almost building the most table stakes thing, deploying with the customers, and then building with them so that you can create most value, maximize the value for them.
0:23:28.0 Kailey Raymond: Yeah. In some cases, I think that that's dangerous territory, like the new Sora release. It's like, okay, now they have to go back and they have to talk to all of these actors and all of these agencies to understand if they can use the likeness of those folks. And it's like, you probably knew that was a problem that was going to happen, but you're building out loud. Love that.
0:23:46.7 Rikki Singh: I think, so you're hitting on a very important point, which is what are the things you do by design versus what are the things you're okay experimenting with? I don't think copyright, trust, privacy, compliance are afterthoughts. They have to be by design. And then there are aspects like tweaking a workflow, tweaking a user experience, like changing maybe even nuances of how you build constraint prompts or how you're doing RAG as you deploy with customers. That's completely fine. So I almost think of, like we say this a lot actually. It's trust and privacy by design. It's compliance by design. So I think all of those aspects have to shift left. And then you have to be very intentional about the things you're experimenting with the customer.
0:24:34.5 Kailey Raymond: Absolutely.
0:24:36.8 Producer : Great customer experiences aren't magic. They're built. And Twilio is the platform that helps you build them. Every customer action, browsing your site, opening an email, reaching out to support, triggers instant AI-powered personalized engagement across SMS, voice, email, chat, and more. No delays, no guesswork. Just the right message at the right moment. From automated messaging to seamless authentication to hyper-personalized customer journeys, Twilio's customer engagement platform powers millions of interactions daily, helping businesses drive loyalty, optimize marketing spend, and create experiences that people remember. Twilio is the ultimate toolbox for customer engagement. Ready to build experiences that matter? Visit Twilio.com.
0:25:29.1 Kailey Raymond: I want to get your thoughts. And we're talking a lot about AI. And you sit in R&D. And so I'm sure you have pretty strong opinions about this, which is, what's overhyped? And what do you think isn't maybe getting enough airtime?
0:25:41.2 Rikki Singh: I think that the whole race towards AGI is overhyped, not because I underestimate what it could do for us as humanity, but I think because of the practical limitations that exist from a technology perspective. I do think we either would need innovation from a power source perspective or quantum technology to enable the AGI vision we're all chasing. And so to me, I think this like, oh, in the next two years you will have AGI, I think that is overhyped.
0:26:19.8 Kailey Raymond: So your p-Doom is a little bit further out than other folks. Got it, got it.
0:26:23.8 Rikki Singh: Yeah. And I think on the flip side, I do think on underhyped, we're not talking enough about the innovation we need from a power perspective, both to combat climate change, but then also to fuel the technologies we are excited to adopt.
0:26:40.1 Kailey Raymond: Totally.
0:26:40.6 Rikki Singh: So I think, I mean, we're trying... There are startups like, I think it's Inertia, which is looking at how to scale nuclear fusion based power. And I think that's interesting because it's clean energy source, it's not as doomsday as fission, and it potentially can help us combat a lot of... I mean, I know we're in an era where we're debating going back to natural gas versus not and things like that. But the reality is the scale at which we need data centers and hence power is exponentially rising. And if we're going to build that many more data centers, I do think we have to be thoughtful about innovating on power sources. So to me, that's an area that I feel is underhyped relative to the hype on AI.
0:27:27.6 Kailey Raymond: I love that. I think that there's been recent studies around this, which is showing that in general, all of the expenditures around AI are already... They've increased the GDP by like over 1%. So that is a pretty massive shift in where a lot of the dollars are going. So we're already seeing that in it. Power stuff is a huge, huge part of this. So I'm glad that you're bringing that up. What are we underestimating about where it's all going?
0:27:55.8 Rikki Singh: This statement holds true in every decade. I think change management is what we always underestimate.
0:28:02.5 Kailey Raymond: You are so right.
0:28:04.2 Rikki Singh: Maybe it's because I started my career with Azure when like cloud wasn't a popular thing to be on. And I had to do a lot of sort of convincing on moving from on-premise to cloud. But even though we say change is the norm, we all enjoy comfort. And it's not because we're not aware. I think it's because of the way our brain is wired. We like comfort because it brings predictability. And in a predictable environment, it's much easier to make decisions. Think about ERPs. Think about CRMs. What have they done? They've converted well-defined business processes into software. They've almost like... The IP of these companies is depth in understanding the business processes that run any company. Whether that's how you do accounting, whether that's how you do supply chain, whether that's how you have your HR workflows, compensation, everything. And that predictability. It's stayed there, for decades. For decades, this is how we build product. This is how we procure software or everything else. This is how we like compensate employees. This is how we do financial announcements. It's all set. And now we're talking about adopting technology that will only unlock value if you rethink all of that.
0:29:30.2 Rikki Singh: So I take it to the other extreme. I know Satya Nadella said SaaS is dead, and people were reacting to that statement. But I think what he was trying to get at is SaaS as we've known, is going to change. We're not going to build software that converts well-known business processes or workflows into user experience. We're getting to a point where personalized workflows that adapt will be a reality soon. So for example, you and I could both be doing brand marketing, but could have two very different workflows that work for us. And we'll have a tech stack that allows for both of it. And it learns from it. And over time, it suggests what could be an ideal one for the size of our company, the scale of our company. I mean, I love to think of it this way. When you're a startup or an SMB, you may start your ERP, CRM journey on monday.com or like one of these places. And then like every big company you see is like on Salesforce, SAP. And that's kind of accepted. Like, yeah, we grew, so we changed our tech stack. That makes total sense.
0:30:44.3 Rikki Singh: We needed something that was more complicated and much better at adapting to our complicated setup. That's kind of how we justify it. We're actually entering a world where like you could have adaptable workflows that grow with you. But that also means you're changing your business processes. Which is very unsettling. Like, go tell a CFO that their financial workflows will like adapt to their finance or accounting team. And so I think that we're sort of underestimating the amount of change management and behavioral change and change in the way we hire folks, we upskill them, we support them on their job. All of that needs to evolve to truly unlock the kind of future we're all betting on through the investments that you talked about earlier.
0:31:32.8 Kailey Raymond: You're 100% right. I mean, it's the same kind of scenario we were talking about with tech stack or buyer organization. Organizations are moving much slower than anybody else. Because they are maintaining whatever processes have existed and trying to jam new technologies or AI into them instead of rethinking the entire process itself. I want to get to a real world example here. So I'm wondering what do you think AI maturity looks like for customer engagement teams?
0:32:02.0 Rikki Singh: So I think I mentioned this before. But I think it is going from... It's on a few dimensions. So, and I'll use our simple, trusted, smart only to structure it. So I think on the simplicity side, it is moving from assuming default being screens to other modalities. I think the simplicity is not just in a simple UI anymore. It's in a simple user experience, whether that is texting back that you want to pay a bill or leaving a voice note if you want to change stuff, all of those experiences being enabled. I think that's sort of one shift that I'm seeing. The second one, I think on the smart front is the point I was making earlier, like moving from targeting to real connection. If you don't know the humans you engage with, I think that's a miss. We now have technology that allows you to get to know every human in a very contextualized manner and we should be harnessing that. And then I think on the trusted front, as I was alluding to earlier, it starts with... I think low maturity is kind of what we're doing right now. Which is we've had frameworks for authentication authorization and the way we set up our backs, the way we do consumer identity and like kind of using that, but enabling agent flows through it to I think higher maturity on the trust front where you assume an agent could be acting on behalf of humans and then how do you enable that infrastructure? So some of the principles we talked about from a zero trust perspective verify first, trust later, those kinds of aspects I think are on the higher maturity side of the scale.
0:33:40.0 Kailey Raymond: That's really helpful. It's interesting that you say we're in a low kind of maturity curve right now. I was listening to Hard Fork, it's a podcast, a couple of weeks ago. They had a journalist on that just did this experiment. And this experiment was like, can I go two days without AI? And to be fair, his definition around AI was like pretty vast. He included machine learning within that. And so what he started to do was he literally collected rainwater because apparently in the state of New York, and I'm sure elsewhere, they use machine learning to basically understand where they should be distributing water. So this is my thought around low, high maturity pervasiveness of AI. I'm wondering your perspective on this is, is 2026 the year where AI becomes ambient, pervasive?
0:34:30.4 Rikki Singh: Yeah, I think there is a consumer lens and a business lens to it. Is it going to be around? Yes, but also I think of it as like it has been around. I mean, we've had AI/ML models for the last decade. I think what generative AI did is consumerize the technology to a great extent, then allowed for more, I would say, organic conversations. AI/ML models have been there in support for over a decade. You've been talking to chatbots before. I think the difference is the adaptability that has come into the flow. The non-deterministic nature also allows it to be more adaptable, which I think has been the change. So would it be everywhere? Yes, just like it has always been. I actually think the difference is going to be, we get excited about technology, we start using it, and then slowly like public policy, laws, everything else catches up. I think 2026 is when things will catch up. What we want to do from a data governance perspective, how do we want to protect copyrights. We're having that active discussion right now. How do we prevent AI slop from sort of taking over human creativity to some extent. Because if everything you've ever created that's original, whether that's art, whether that's writing, whatever it may be, content, if that can be ingested and then adapted by AI, how do we build a society that still incentivizes creative pursuit? And so I think those are the questions we will have to answer in 2026, given that the scale of adopting some of these things has grown exponentially this year.
0:36:21.9 Kailey Raymond: Just small little questions that you're asking, just the tiny ones, Rikki, and we'll solve them, no worries. I like this thought. I mean, it's already exists. It's already here. It already is pervasive, but it'll change a little bit as we head into '26. And some of the things that you're talking about that we really need to start thinking about have to do with this concept that we really started with, which is trust. So as we're heading into kind of the new year, I'm wondering how you think about trust within this framework? What does that look like?
0:36:54.4 Rikki Singh: Yeah, I think, I mean, it starts with the fundamentals. Which is first, just everything runs on data. So step one is actually just making sure that you have the right permissions to use that data. You are thinking about the right sort of compensation for the data you use for the folks that are providing it to you. You're thinking about the right segmentation, data governance, all of that. So I almost think of it as like, it starts there. There are some fundamentals we've always followed from a data protection and governance perspective. And I think we just need to get better at it. Then it gets into sort of as you use expressions of this data to generate things. I mean, even small things. And Perplexity and ChatGPT have started doing this, which is like they're watermarking things that are AI generated because it's not just... I think trust is not just about me trusting that you're using my data well, but it's also then me knowing when I'm engaging with AI versus human or me knowing when it's an interpretation by AI versus facts.
0:38:05.6 Rikki Singh: I mean, journalism is a great example. We trust the news we get because we trust journalists to go do the work. And now if I'm looking at news that has been spit out by AI, step one is just like be transparent, be transparent about the source of the information. And then as a consumer of that information, I will have to make intelligent choices about what I trust versus not. So to me, like I think trust is both in how you come up with the information, but then also how you communicate and then how you take feedback. The other aspect we haven't talked about a whole lot, but I think will matter next year is if you have agents acting on your behalf, I almost think of it as if you have agents for payments, if you have agents for legal actions, if you have agents for shopping, they're all different sort of use cases or expressions. It's almost like me having assistance for different tasks I want to do. How do you build the right infrastructure to make sure you recognize what I'm delegating versus not? How do you inform me when you notice... I mean, we're so used to getting that notification from a bank. Hey, did you ask... Did you make this transaction whenever there's anonymous behavior? Think about a world when you're making the transactions and your bots are making transactions. How do we manage identity, consumer identity?
0:39:31.2 Kailey Raymond: I was going to ask you this. This is the thing that's such a step change, which is agentic browsers, they already exist. Perplexity came out with one. And so it is changing what we need to think about as it relates to your point, identity and authentication. So can you just walk us down your thoughts around that?
0:39:51.2 Rikki Singh: We're at a time where we've had authentication, authorization constructs that have worked for human. For example, two-factor authentication and now moving to passkeys. We understand human behavior and we introduce verification at certain points to balance user experience and friction. So it's like always a trade-off between seamless user experience and verification and trust flows. For every single task you ask me to do or you were trying to do in an app, if it constantly asked you to I don't know, face ID or put in your password, you would be frustrated.
0:40:30.5 Kailey Raymond: Yes.
0:40:31.1 Rikki Singh: But with agents, that's not the case. For them, it's not really a task, like a tax on their flow. So I almost think we have to, in that UX versus friction or balance, we can add more on the side of friction to start with, where you limit the blast radius as much as you can. You start with microsegmentation of permissions such that an agent, at least to start with, can do one specific task for a human and that delegation and authority will expand over time as you build more trust. But I almost think we have to err on the side of caution.
0:41:07.8 Rikki Singh: Because we also have a very diverse set of customers. And their own maturity, like our consumer maturity is so broad. You have Gen Alpha, which is like just growing up by default with AI and Gen Z, which is relatively comfortable with it. But then you also have boomers who will be at the other side of it. Imagine sometimes they may not even realize they're in an agentic browser and kicking a transaction off. And so how do you bring everyone along versus like... And then I think the last piece on trust, I will say is, if we do go to a world where my awareness and purchase decisions are informed by these search agents, I do think we need markers that those agents are acting on our behalf and not on the highest bidder. Think about Google search. Search engine optimization is an entire business. The web is free because ads are monetized. And so how do we think about that balance as we shift to an agentic world where yes, maybe I ask ChatGPT to make a purchase on my behalf, but I want to have the trust that it's acting on my behalf and not on the next vendor that wants to sell me something.
0:42:29.6 Kailey Raymond: How do you not hear you? And my thought immediately goes to, okay, great. If you're building trust and friction and all of this by design, which absolutely, this is a new world. How do you not let that stifle innovation?
0:42:46.1 Rikki Singh: I think it goes back to there are things that are okay to move fast and break and there are things that the fact... If you break trust, I actually think that stifles innovation more than the fact that you went a little bit slow to put the right guardrails. Because earning back trust... I mean, think about our own personal relationships. How much time does it take to repair a crack in a relationship?
0:43:11.1 Kailey Raymond: There's a personal story here we'll get to later.
0:43:14.7 Rikki Singh: But I mean, versus like just not letting those cracks appear.
0:43:18.1 Kailey Raymond: Totally.
0:43:18.6 Rikki Singh: And especially with adoption of new technology, I think if we can almost trust by design, I always fall back to it because I think the cost to innovation and speed when you get it wrong is much higher than just investing in it upfront.
0:43:34.7 Kailey Raymond: Beautiful. I have a couple of fun, quick hit questions to round us out, okay? This has been great so far. I've learned a ton. If you had to make a bold prediction about next year or just in general, what would it be?
0:43:50.4 Rikki Singh: I think there are a lot of predictions out there about agents, AI. I was initially thinking about quantum as you were asking that question. But I think the most interesting one that I don't have a defined answer to is, what is the monetization model of the new web? We've been in this world where we've monetized... Internet is free, but we've monetized ads. And then from a content perspective, we're so used to paywalls. You want to read to the Wall Street Journal? You subscribe to it. There are paywalls for all the content you consume. And the paywalls have been subscription primarily. I almost think in 2026, we may see the collapse of the paywall economy. And the reason I say that is... The reason we have paywalls was because of the concept we were talking about earlier. Which is friction versus convenience. And for humans, if I asked you to transact every time you wanted to access an article, you would be unwilling to. So I went to subscription where you pay once and you forget about me for the rest of the year. For agents, that's not the case. So you could actually charge like cents per access and not have to get people to commit for a year. So I think the evolution of that sort of paywall economy to whatever this transaction-based content access could be would be interesting.
0:45:18.8 Kailey Raymond: What's something that you're personally excited to build with or prototype with heading into next year?
0:45:23.5 Rikki Singh: I think two aspects. One, I always think of Twilio as abstracting complexity. And I think for the next year, I'm thinking about how do we abstract complexity not just on the AI stack, but on the communication stack, where we have a combination of terrestrial and non-terrestrial networks. So what does it look like with the combination of LEO, GEO satellites coming together and like us being able to enable our customers reach their consumers across those channels? That's one. The other one is genuinely quantum. I think like we're thinking about the implications of quantum cryptography and communication for our business. And I mean, it's one domain where I think the physical and the digital, like it's fascinating. I think of quantum as like, we went from analog to digital and we're going back to analog. And so I find this vision where we're able to almost in the digital world bring the analog is like fascinating to me. So those are two areas.
0:46:28.4 Kailey Raymond: I love that. I mean, you've hit me with some analog thoughts today, no screens, you said, and I said, subscribe. So I'm here with you. My last question for you today, any advice for the builders among us?
0:46:41.9 Rikki Singh: I think everybody will tell you like, go learn about AI, become AI literate, be AI by default. I would say go back to fundamentals. They will never fail you. I absolutely agree. Be a learn it all, not a know it all. That's my default mindset for any era, any decade, any moment. But I actually think this is a moment where going back to fundamentals around obsessing about the customer problem, not the cool tech. Being thoughtful about how you solve those problems, falling in love with those problems, identifying the soft signals. Because the world is changing, not just for us, but our customers. And if we can get ahead on thinking about the problems that they face, not just today, but will face tomorrow, we can help them. And so I think if there are builders out there, yes, experiment with interesting technology, but please think about the problems that are pertinent to solve for our customers and not just throw cool tech at them. Because there's nothing that would make the world better rather than like us figuring out the most important problems to solve.
0:47:55.7 Kailey Raymond: I love that. Back to the basics makes a ton of sense. Rikki, you gave us a ton to think about for next year. So I really appreciate you being here.
0:48:03.3 Rikki Singh: Thank you.