#121: "People, product, process, platform" - Cindy Turner, Chief Product Officer at Worldpay

24:43Okay. Yeah. So we've had a role within crypto from the beginning. and a lot of the big crypto clients are very large customers of ours. So exchanges are customers of yours, like the Coinbase of this world. Okay, I didn't know that. But then also Stablecoin, which is kind of a subset of the crypto industry, to me is really increasingly there's a lot of call to action for it to play a role in the B2B transfers on the back end of payments. So two primary use cases that I think there's a real benefit to using Stablecoins to solve. One is for all of these marketplaces where we move the money to our clients, which might be a payfac.

25:23If you think about a business that has a lot of small businesses that sit behind them, we are their processor. We get paid by Visa MasterCard. We make those funds available to our client, who is the payfac. And that payfac is then trying to make the funds available to their customers as soon as possible. Small businesses are trying to have money same day. They want to close their terminal at the end of the day and see the money show up right away afterwards. There's use cases in marketplaces. If you talk to an Uber driver about when they get paid, there's a lot of Uber drivers who are taking instantaneous payouts.

25:55They're paying 1 % of their earnings to push that money into a debit card so they can go fill with gasoline, right? So the acceleration of funds through the value chain in these platform businesses and marketplace businesses, stablecoins is a really good fit for. So if we can do instantaneous transfers and we can make those funds available while we're still holding them in a U.S. denominated account, it's a huge advantage to some of these platform use cases. The second one, of course, is if you can avoid FX. So again, to use some of these examples of platforms, there's some platforms, particularly the creator platforms, where the majority of their providers and creators are sitting in Southeast Asia or relatively small countries and currencies.

26:43And those providers, if they're getting paid in their local currency, they have to take a three-ish percent, like 300 basis points hit in FX in order to get paid. Yet the cost to them, if you think about that provider's cost, their cost is actually frequently buying ad spend within like a meta platform. So their cost is denominated in US dollars and their revenue is coming from US dollars and the US consumer. And the same example could be said in the UK or Europe. But and every time they have to make a round trip of those dollars into their home currency, they're losing three to 4 % of the top line.

27:19So if we can hold all of those funds in a stablecoin on their behalf, they can spend in dollar and they can receive in dollar, but they don't need to have a US bank account to do that. Those are the places we've exposed us so far from a payments perspective. Technically, it's in our funding. We can fund in stablecoin and then we can pay out. So we can expose payouts in stablecoin. So do you just use stablecoins like on the Ethereum network or something like that then? Or have you done anything on the Bitcoin network? work? No, we're just using stable coins in those two cases. Yeah, just stable coins.

27:51Yes. At least you have some partners behind the scenes. We don't operate on wallets, for instance. So we have some partners on the wallet side. But yes, we can either fund in stable coin or we can facilitate the payout to stable coin from our clients who are moving money that way. Yeah, absolutely. This episode was brought to you by Be Digital. Be Digital support leadership teams to optimize cost and get more out of technology investments. BDigital and the team have unrivaled expertise with technology license management and data remediation and are therefore perfectly positioned to help prepare organizations for AI technology capability.

28:32And on the last point, BDigital have just developed a cutting edge AI readiness assessment, which provides tech leaders with the platform they need to make well-informed decisions about AI adoption strategy in 2024 and beyond. Go to bdigitaluk.com to find out more and get in touch.

28:55I wanted to ask you about agentic commerce, okay? Something you've spoken about in the past. For those unfamiliar, what does that term mean and how might AI agents transform how consumers and businesses transact in the future? Yeah. First of all, thanks for starting at that point is what does agentic mean? And it's a very highest level. It's the idea that an agent is purchasing on behalf of a consumer with their instructions. I joked, I did a LinkedIn post a couple months ago. I have a wonderful nanny for particularly when I travel. And she's been my agent for a really long time. And we do it outside of network rules.

29:31She has my own chase card in my name in her wallet. And there was a day she was like, she texted me and she was like, hey, like a bird poops on the car. so I'm going to take it to get a car wash and then I'm going to rent to the grocery store because you need like milk and juice. And I was like, I've been living in an agentic world for a long time and it's wonderful. Welcome to it, everybody else. And she's an autonomous agent for me, right? And that she's making those purchasing decisions. As soon as we do this at scale with an AI-based agent, we need to - I love that analogy, by the way. You've done this before, haven't you?

30:03But we need to protect our merchants and their exposure to potential liability for erroneous purchases, right? So whether it's true fraudsters or just accidental purchases, the merchants, as they're thinking about how they can expose their inventory into an agentic world, an agentic shopping world, how can we do this in a way that remains safe for them? I'll also, I'll just give a kind of another top line. So we went to our client event. We have this big rethinking event in Europe every year. We had 250 of our clients, largely in the enterprise space, a trillion dollars of payment volume represented in the room.

30:38And we did a survey of how disruptive do you think agentic commerce is going to be? So as discovery from a consumer moves into an LLM and the consumer asks chat GPT for X item or asks an agent that's using an LLM behind the scene to do their closet replenishment or book their trips, how disruptive do you think this is going to be to your business? Asking our clients. They were like 85%, nine or 10. I think this is going to fundamentally disrupt my industry, the consumer experience. the value chain, like how my company makes money. Very worried about it. Second question that we said of 250 people, how many of you have tried an agentic transaction so far?

31:18Four people out of 250 had actually initiated one. Perplexity has had them for three or four months. OpenAI has had them for two or three weeks. So this is still very, very new. And you've got these payment geeks in a room not having even tried it themselves, right? So we're very early in this. And then the crazy question, the third question was, how many of your companies have a strategic initiative at a C-level to develop your approach to agentics, so to the consumer moving discovery into an LLM? And only 30 % actually were. And I'd say that that's resonating as I'm talking to a lot of merchants is there's a fear, like this is going to fundamentally change everything about my business.

31:56I'm very worried about it. I haven't tried it yet myself. And even though I'm worried about it, we're not really working on it because we've got a lot lot of other stuff going on. So that's the moment that we're in. We're very early. But all of these parties are trying to kind of land grab of what does, what can Agentic mean? What role can I play in it? For the people who are listeners who are spending much time in LinkedIn, you know, following Agentic, there's a new protocol announcement like every single week. So we had Google announced a protocol, Visa has a protocol, Massacre has a protocol, OpenAI did a protocol.

32:28So we have all these different kind of protocol announcements, which is, they're trying to say this is how I I believe the industry should work. But really, this is how my LLM is going to work or this is how my card can play in it. Our view from a well-paid perspective is, again, we want to normalize this for the merchant. So we want to be able to have one set of tools, one set of APIs that is wrapped around our regular payment processing that lets the merchant, what does the merchant need to be able to do differently? First of all, they need to expose their goods into an MCP. That's a little bit outside of what we do typically, but we can expose those MCPs within our environment.

33:02Second, we need to give a signal to the merchant of, is this a trusted agent? So is this an agent that we believe has good intention and not just a fraudster? Because the fraudsters are all agents too. So the fraudsters are coming in, they're going through guest checkout, they're trying to take over accounts. That is also an agent. And so we have to be able to discern between a good agent and a good actor and a bad actor. Secondly, even within good actors, fraudsters come in and use those too. If I have a good actor and this is an open AI agent, how do I know that when a card was stored in that agent and given the ability to spend, the right set of fraud authentication was done at that moment?

33:45Did I bind a token to that agent with a established way of assessing whether or not there was an authentic consumer that was tied to that card? The storage of that credential, again, network tokenization back in my history there, but network tokenization is a critical component of that. And then the most interesting is going to be once an agent checks out, how can the merchant know that the agent is buying the thing that the customer wanted them to buy? So that's a new API, right? The merchant today, if they get shopped as a guest checkout, the assumption is that the thing being purchased was in line with what the person wanted to purchase.

34:21But if an agent went and did that purchase and there's a mismatch, the merchant's liable today. So how do we help protect the merchant from knowing and being able to validate what the consumer asks them to buy? And that's probably the area that needs to mature the most over the next couple of months. Tokenization is going to get so important when we have autonomous agents representing our interests left, right and center. Do you know what I mean? Because we're not going to know which ones are real, which ones are authentic or not. Otherwise, are we? So, yeah, I hadn't really thought about that.

34:51How important that is. Yeah. And in the early days, what we have is we have agents shopping on a website still. So an agent is mimicking a human in a guest checkout environment. I actually think eventually it will mature to an API to API. So an agent to agent based transaction, right? where the merchant exposes an agent that validates digitally the agent's instruction from the consumer and that becomes the transaction. You don't really need a website if there's no human on the website, right? But if you think about if you're a merchant, you want to establish brand, you want to establish customer loyalty, you want to be able to upsell, you want to be able to hold that consumer information so you can come back to them later.

35:34You need to be able to catch that consumer in the LLM, sell to them, But think about how to influence and drive consumer loyalty to you as a merchant without having this traditional front door of an internet that they came to shop in. So, you know, it's wild west still. But it's a really, really interesting space. It's coming thick and fast though, isn't it? Yes, very much so. We can't be too far away from the majority of people having their own AI avatar agent who does shopping for them and stuff like that. That is crazy. Have you got any other tools that you use regularly on the Agenda AI side of things, Cindy?

36:14Yeah, we're working with OpenAI and we're already pushing into our DevCenter their APIs. So one of the things I really like about OpenAI and their approach is it's multi-acquire from day one. So the assumption is that the merchants will integrate into their MCP, expose the inventory, allow for those purchases to happen in the ChatGPT environment, but the transaction will flow through their regular processor. So they're really just exposing that endpoint and saying, here's where to send that transaction to. And so then the processor can do everything else, BAU to the merchant. Because if you think about these huge merchants, they have a treasury operation that needs to be able to do reconciliation, to close the books on the second of every month, to be able to make sure they get paid.

36:57They don't want to view this as like a net new front door. They need it to flow through the backend that they have set up at scale. So yeah, so it's a new discovery process. And that discovery process has a friction-free consumer authentication happening in it. But that needs to flow through a set of rails that is consistent with the rest of their processing relationships. In terms of governance and the ethical element of it, okay, in terms of AI deployment in a company like WorldPay, what sort of principles guide your approach to using AI responsibly within the WorldPay's product set? Yeah, so there's probably different areas of both the risk of exposure and where we embed the AI.

37:41So we have both for internal tools, right? So we have our own instance of an LLM that we can use internally. The critical aspect of that is we don't contribute information back into the source LLMs. It's actually multiple LLM day one. So you can choose to use the open AIs. You can choose to use any one of four different LLMs behind it. But we don't contribute any information back into their source code. And so we're able to leverage the best of it without exposing any of our data from a risk perspective. We have that. We also have Microsoft Copilot. A lot of workflows are really able to benefit from a productivity standpoint with just embedded meeting note-taking, to-do lists, a lot of code checking in workflows for an engineer.

38:26We also use AI in our products, embedded within the product. that's largely in the big data aspects of our products. We have AI in our fraud tools. We have AI in our retry. So if a transaction fails and you want to go retry it, we have AI that's anticipating where to send that transaction next. And then we also are using it in some of our routing optimization, being able to anticipate where the lowest costs are going to be embedded within our products. And then when you think about the consumer using the AI, That's the last kind of element that I think is the least mature still. But if the consumer is using AI, then you're in an agent-based transaction.

39:05I don't know that the processor needs an AI tool. They really just need an API that can accept that transaction. But it's certainly something that we're tracking really closely and working with all the same parties. Yeah, of course. So let's look beyond payments now, Cindy. Yeah. I'm talking business, society. What are you most excited about in terms of the proliferation of AI technology in any manifestation, agentic, generative AI, just in terms of innovation generally around AI, true AI? What excites you the most? I mean, I just love the amount of change that's happening in the industry right now.

39:41And disruption allows for opportunity. So I was talking to an analyst yesterday, and he was saying that he thought Visa was going to be the winner of agentic. And I was like, I don't know that I believe that. I actually think if someone, and in particular, if one of the LLMs embeds a pay-by-bank payment method in the U.S. where pay-by-bank has struggled to scale, and we have a really friction-free pay-by-bank embedded within that LLM, all of a sudden you have the opportunity for that to take off and become the way to pay in that type of an environment. I don't know where it's going to go. I'm not sure what's going to be disruptive.

40:17But anytime you have these major title changes of change, I think that we're just opening ourselves up to a new way to scale really quickly. And whether that's a way to pay, a consumer experience, platforms that don't exist today that become the consumer preferred way and trusted way to shop for XYZ as a category. I think there's going to be name brands in our lexicon four years from now that no one's heard of yet because of their ability to scale so quickly. Yeah, no, absolutely. I've got to ask you the inverse slant on that question, Cindy. What are you most... Concern's the wrong word. What should we be fearful about?

40:57What can go wrong? I'm worried about some of our merchants, right? So I sit in the merchant seat a lot. I haven't actually really ever worked for one. They're very nervous, right? If consumer loyalty is to the brand of the good getting purchased and discovery moves to take any category, right? So women's clothes, pantry items, like any one of these categories. If what the consumer cares about is making sure that the physical good that they buy is from a brand that they trust, And their discovery moves from walking into Nordstrom and then shopping their curated selection to having an agent that curates a selection on their behalf.

41:40There is a real concern that the retailer and the merchant, the person who's doing the selling, their value add of inventory curation becomes less important. And, you know, I want to represent them and help them really push themselves and push their inventory in the right way into those LLMs in that moment of change. That's probably the thing that I think we're in the most risk is that you're going to have new scale platforms start to pick up those same goods and the revenue share of who's actually doing the selling is going to redistribute. I've got to ask you about this one, Cindy. jobs displacement anyone who's used ai can see straight away that this is going to take people's jobs it's absolutely inevitable as with most technology innovation going back to the industrial revolution generally there's a hit on the amount of jobs created by any industry and then it generally recovers and then you see a net gain in jobs because the technology has just elevated the industry generally.

42:41Is this another iteration of that, in your opinion? Or do you think this is different? I think it's certainly another iteration of that. It'll take different jobs and it'll shapeshift in a very different way. And look, these are people too, right? So, you know, highly conscious of that. So roles in an organization like call center and QA teams and technology, like that feels, you know, that feels scary and that does feel very risky. Where am I an optimist? I think that the thinking is like the AI only knows what it knows. So if you either create knowledge or your role in the ecosystem is to add intelligence to decision sets, then AI can make you more productive, but it can't replace like the core function that you're doing.

43:30I think I'm most bullish on those types of jobs. A lot of it is leadership roles, product roles, GM type roles, trying to set the thought leadership, the communications and PR jobs. Yeah, the roles with more of a human element, I think you're explaining that, aren't you? Yeah, they don't need to write anymore. But like, gosh, if you can actually differentiate yourself in terms of being able to influence the talk track, like that is so critical. Yeah. Did you envisage any new career paths will emerge? Well, I'm sure some career paths will emerge, but are there any new career paths that you think will be particularly prominent over the next couple of years?

44:08AI ethics, for instance. Have you got any? Yeah, you're outside of my wheelhouse. I don't spend any time on that. I mean, I was actually interested. I read an article that was like, there's actually an entire kind of sub-industry right now of these people who maybe used to work in a call center. They're now going into these kind of data farms and validating pictures. So they just spend all day. Is this a picture of a truck? Yes, no. Here's a map. Is this like a stop sign? Yes, no. So the human validation of the best guess by the AI tool to train the model. So this model training, I don't think that's a long-term industry and it sounds like a terrible job.

44:49But like who even realized, like I didn't even realize that was a job right now, right? But it is and it's happening at scale. So, yeah, I mean, I think the people who are trying to, you know, very customer centric in what they're trying to solve for, like, I want to stall for a friction free consumer experience of trip planning, of closet, you know, management of any one of these kind of sub industries, you know, the pantry management. You know, I think that there are going to be scale brands in those spaces that, you know, I'm really excited to see what they can do. Yeah, no, absolutely. So just casting your mind back to the, I know you got a long way to go, Cindy, don't get me wrong.

45:35But I just wanted to ask if you were to have a chat with your 21 year old self, okay, someone who's starting their journey, knowing what you know now, what advice would you give to that person? Interesting. Yeah. You know, when I started my first job out of college, it was as a strategy consultant in Boston. And I knew I loved the consumer. Like I've always loved the consumer, but I thought that meant working for a consumer product company. I think the thing that I wish I had recognized sooner is that you can be consumer obsessed and work in the technology that enables shopping and enables the commerce.

46:11And that's both a more profitable sector and the CPGs themselves, but I think also much more dynamic and faster moving space. I didn't really know what tech was. Like fintech wasn't really even a thing. Product wasn't really even a job function. I stumbled my way into those, but I wish that I had appreciated that you could spend all day driving consumer technology and consumer experiences, but do it for a technology company that was driving that instead of for a CPG itself. Yeah, fantastic. No, I love that. That's brilliant. So just very quickly on product. So in terms of running a productive product team, what are the core fundamentals?

46:54What advice would you give to, say, an entrepreneur about spinning up and running a really effective product team? What are the core pillars that you have to implement in order to achieve that? Yeah, so first of all, you need a really strong product ops team. That is a muscle that is incredibly valuable. You can't run a product team at scale without a good product ops. It's not necessarily a skill set that I think is the strongest skill of someone who views themselves as a great product leader. The great product leaders are innovators and they're dynamic and they're really good at thought leadership and they're really good jumping into podcasts with people and talking about the future.

47:27They're not necessarily really good at rag status and tracking and PDLC consistency. And you can't run a good product organization without doing that really well too. So product ops is critical. We've run this transformation since I joined 15 months ago, two months ago in WorldPay, we did people, product, process, and then the platform. And so we have made massive change against each one of those four pillars. So from a people perspective, it was like to your first question, how do you lead putting the right people in charge of a domain and then giving them autonomy to go make the decisions but support where they needed it?

48:02So getting the right leaders in place. Process is rag statuses and tracking and really good road mapping tools and communication to all of your stakeholders about what's getting done, what's stuck, what's coming when. So you have to be in a position to communicate that. It's so many different levels in an organization, externally, internally, within the teams, within the engineering orgs. You have to be able to share that in dashboards and reporting tools. And there's some wonderful software that does this really well now. We implemented Aha! and it's transformational. No one sits around in PowerPoint making a roadmap document anymore.

48:36So we just look at a hub, we view it there. Is that A-H-A? A-H-A exclamation point, actually. Okay. We'll have to put a link to that in the show notes. But yeah, but thanks for sharing that. That's really useful. Yeah. Having a really good road mapping tool. And I've had prior organizations use Monday. So there's other ones out there too. But yeah, you need a good, consistent way of being able to track what you're working on. A good product organization is transparent at all times about how they made the decisions they made, how the work they're doing is going, if it's going to hit the milestones that were previously communicated, and then where we plan to go next.

49:09So there's an infinite ideas in the roadmap of what we could do. How are we deciding what we're going to go do next? And so being able to communicate that to at the right altitude to your stakeholder with the right timely information and transparency is just so critical. So the process of product and then the product itself, picking a couple big bets. How do we stay on top of Agentic? How do we decide what to build? How do we think about global expansion and modularity of our tools. That's the stuff that, you know, product people just love doing. And that's kind of within ourselves that we do it.

49:40But influencing that within the team, having an opinion on a couple of things, making a couple of those bets. And then platform, look, WorldPay is the series of a bunch of prior acquisitions and our platforms are, it wasn't just built, but having a view of like how we pull that together, how we deliver a best-in-class developer experience and best-in-class backend for the merchants and shield them from any of the complexities in the scale of who we are as a processor. Fantastic. That is a brilliant answer, by the way. So let's finish on this, Cindy. We always ask our guests to recommend a book or it can be, if you're not much of a book reader, it can be a piece of content, just something you were inspired by recently.

50:19Okay, I'm going to definitely go podcast. So the Acquired podcast, I am absolutely thoroughly obsessed with it. Acquired, okay. Acquired is the name of the podcast. And then they have a second one too, I think ACQ, ACQ2, something like that. But it is absolutely phenomenal. The acquired one, they do these four to five hour deep dives on a company. And like how the company runs beginning to end. So they did one on Meta, they've done a couple on Google, they've done one on Microsoft. Oh, wow. And it's what problem are they trying to solve? How have they scaled? How have they evolved? They just did a series on Alphabet.

50:55And then the last one of the series, I mean, we're talking 12 plus hours of content here. on Google and like what's made Google, you know, how Google's first growth, like the best product that was ever invented was search, right? Yeah. And then how they scaled beyond that, even like some of the Waymo, you know, things that they've done. But then this last chapter, which was a four hour chapter, is on, you know, the role Google should play, can play in AI. So they were the talent incubation of all of the current LLMs and then the role that they can play going forward. It is the single best education that I've ever had.

51:30And so I deeply, deeply recommend the Acquired podcast. That is some recommendations, Cindy. We are absolutely going to have to check that out. We'll have to put that in the show notes. But look, thank you so much for coming on. I mean, I've had such a blast. I mean, yeah, you clearly live and breathe payments, which is so inspiring. You're clearly passionate about what you do. So thank you so much for sharing your story with us, Cindy. I really enjoyed talking to you. Thank you so much, Gareth. I really enjoyed this. Thank you.

52:04This was one real avalanche of insights, a proper masterclass in how payments have evolved and where product strategy fits into that bigger picture. There were so many moments worth highlighting here, but it's really hard to look past Cindy's take on what really makes a great product function tick. That was just gold dust. She talked passionately about the power of product operations, which I thought was really interesting, and then broke down creating a good product function into four Ps, which I thought was genius. Product, staying close to innovation and evolving customer needs, people, making sure the right talent is in the right roles, easier said than done, platform, building a strong technical base and a great developer experience and finally process having the right tools and roadmaps in place to bring it all together you get those four working in harmony and you've got a blueprint for a truly world-class product organization i thought that was just fabulous and great insight so a huge thanks to cindy ford for joining us and thank you so much for listening i hope you found it as insightful as i did until next time thank you so much for tuning in to the tech leaders podcast Don't forget to subscribe and I shall see you on the next one.

53:44organizations for AI technology capability. And on the last point, BDigital have developed a cutting edge AI readiness assessment, which provides tech leaders with a platform they need to make well-informed decisions about AI adoption strategy in 2024 and beyond. Go to BDigital UK to find out more and get in touch.