Innovation in Practice: Lessons From the AI Frontier

Hi, everyone. Hello. Thank you so much for coming. For those I haven't met yet, I am Kristen Stoller, one of our editorial directors here at Fortune. This is going to be such an exciting session. This is on innovation in practice, lessons from the AI frontier, hoping to bring a lot of lessons, a lot of case studies. And we have a great group of speakers who I will introduce in just a moment. But before we get started, I would like to thank DoorDash for hosting today's session. We very much appreciate your partnership. And a quick note that this is really meant to be a group discussion. We want this to be very interactive. We want to hear your thoughts, your questions, your comments. So whenever you have one of those, please just raise your hand and someone will bring a microphone over to you. And another quick reminder that this session is on the record, which is very important to note. So without further ado, I would like to introduce our discussion leaders here today. So we have Liz Jarvis-Sheen, who is Chief Corporate Affairs Officer at DoorDash. We have Erica Burkhardt, the Chief Legal Officer and Corporate Secretary at Yum! Brands. And then we have Ami Badani, who's the Chief Marketing Officer at Arm. Thank you three so much for being here. Thank you for having us. Okay, so to start, I just want to ask the obvious question, because this is a session in AI. However, none of you run tech, engineering, AI, but all three of you touch AI strategy very deeply. So I guess just to start, what does AI look like from the marketing, the comms, the legal seat that the CTO of a company never actually gets to see? And Liz, let's start with you. Well, thank you, everybody, for being here. Yes, as the sponsor, we love you at DoorDash. Please let us know. This is our team over here. If you have customer complaints and feedback, let us know. EJS at DoorDash.com, send me an email, let us know. We always want to get better. And in some ways, that gets to the question about AI. So yes, I have the good fortune to lead communications, public policy, government affairs, those kinds of things, internal comms. And certainly in sort of the early era of AI, it was not like the obvious place to necessarily add a lot of value. But when we talk about it for our teams, it's not use AI for the sake of using AI. It is use AI to do the things that are sort of rote or laborious that allow you as a human to put your superpowers elsewhere. So, like, let AI do the first pass at a briefing document. Let AI do the first pass at a press release. But don't then ship it out into the world. Once that has happened and you've learned the tools and refined the tools, then have the human come in, get the right voice, make it specific to the thing that you're talking about, zhuzh it up a little bit. But for us, we think about it internally as sort of how does it help you be better, not how does it replace the work that you do. And then I would say externally we think about it as like, does it help our customers? And if it doesn't help our customers, If it doesn't make it easier for you all to get a DoorDash delivery, if it doesn't make it easier for a merchant to grow their business, if it doesn't make it easier for somebody who's dashing to earn the money they need, then we're using it wrong. I like that. Erica, for you on the legal seat, how are you using it? Well, first of all, I have to say happy Taco Tuesday on behalf of my Taco Bell team. It is also National Taco Month. We are very proud that our Taco Bell team liberated Taco Tuesday a couple of years ago. You're welcome. It is now safe for you to celebrate Taco Tuesday without being concerned about trademark infringement. But back to AI, listen, I think legal, you usually think about our interaction with AI from a governance perspective. Like, okay, so Erica's in charge with her team of figuring out the governance aspect, and that is a very important part of our job. But we at Yum view AI as something we should all be using, just like you're talking about, Liz. So from a legal perspective, I view it as a very important enabler for our teams to do their best work. Just like we want to make sure that all of the teams across Yum are using AI to reach our customers in the right way, I want to make sure that AI is enabling my team to get their best work done. And Ami, as a marketer, I'm curious how you think about AI and content creation and where you think a human is still necessary in that process. Yeah, so from a marketing perspective, you know, I think about it a lot more about positioning. So for folks that don't know about Arm, we're a B2B company and we're building the kind of the infrastructure that folks like Yum! Brands and DoorDash are building their software on top of. So we license technology to a bunch of chip companies and then recently also now have our own chips that either large language models or agents are running on. So we're sort of the infrastructure underneath. And so we think about it both from an external perspective around how do we accelerate progress for all of the AI productivity we want to see in society. And then also we use it pretty much everywhere. I would say kind of our biggest value add has really been around the unlock with our engineering teams. So even though I don't run engineering, I spend a lot of time with the engineering leaders figuring out how we can accelerate progress to get products out to market much faster. And then within marketing, we've sort of used it everywhere. And I think one of you guys mentioned, you know, doing the first draft, not doing the first draft. And so we use it a lot for, you know, how do we actually have it as sort of a guide post for us when we sort of use it to kind of gut check our thoughts. And then also whether it's, you know, content creation or campaigns or communications, pretty much every aspect of marketing I think is being changed because of marketing for the better and I think it really makes many functions more productive with the use of AI. Absolutely now I want to dig into those case studies those specifics because I think the most valuable thing we could do is talk about you know the AI initiatives that are really doing well and ones that may not be doing well and what you learn from that and what you take and how you innovate on that and Liz I'm going to put you on on the hot seat first because I don't know if many people know this story but I followed it very closely but earlier this year in January there was this viral AI generated Reddit post that accused an unnamed delivery app of exploiting uh both drivers and customers yes DoorDash was not named in this post at all but you responded anyway yes talk us through that Why did you think it was important for you to respond even when your company hadn't been officially called out and and what does that what did that teach you yeah um it was the day after New Year's. So really fun way to kick off. Was it also a Friday? It was a Friday. Oh, yeah. It was amazing. But yeah, it's so I think in particularly how it relates to AI. So this thing, it pops up on Reddit, which is where everything starts. And then eventually makes its way in particular over to X. And it was really when it made the leap from Reddit to X that just sort of like took off and as it was happening so number one we weren't as you say we were not named but I do think you know when when there's anything that's calling out an unnamed delivery platform like we are the market leader here in the U.S. I'm not speaking out of turn check any third-party data but it's like in some ways we are synonymous with it and so and then in particular when you sort of looked at what was in the post it had these claims of like oh you know because essentially was talking like I'm a you know I'm a security engineer that I'm leaving and like here's my DoorDash ask me anything but it sort of made these allegations of things like oh the company uses a desperation score to figure out how much to pay drivers and this sort of this kind of thing it wasn't about us but like everything that was alleged was so antithetical to our values and our mission and because we are so synonymous with delivery like we have to make clear this isn't us and where the AI component comes in is like I don't care if it was AI generated or not because the people who are reading it on Reddit and who are reading it on X and who are wondering is this DoorDash they don't care and they don't know at that point in that moment whether it's AI generated or not like the the way I sort of think about it is like if your house is on fire it doesn't matter if the fire started by an arsonist or an accident. You have to grab a hose and put the fire out. You can figure out who started it afterwards. But I do think that for communicators and lawyers and marketers and everybody, what has changed with AI is the pace at which content information and misinformation and disinformation can be created and then distributed. And I think that that requires a different approach and also a higher risk tolerance to be able to say, Like, I don't care if this is fake. It's important, you know, or generated by AI. It is important to know that it's not representative of our company and our values. I feel like from both a comms perspective and, I mean, I'm also curious for you, from a marketing perspective, I think that corporate communications now kind of has two audiences. It's both the human audience and then the machine audience. And do they need different messages? How do you go to both? And whoever wants to take that, yeah. Yeah, I mean, so I run communications for the company. And one of the things that we've noticed over time is, you know, you sort of look at your traditional forms of media, I think those have evolved a lot of people, more people use social to get out their message. And so we don't necessarily look at it as AI derived content, though, that there's a lot of that happening. we look at it as like what's our message who's our audience and then how can we sway different forms of media whether that's traditional media or social media uh to advocate for you know whether it's better education or being better informed about whatever it is that we're trying to uh sway the audience about so it's less about whether it's AI because i think the way you respond would probably be the same but to Liz's point i think the acceleration of allowing us to do it much more quickly. So, you know, for us, we sort of underpin pretty much every technology company that's building chips that are putting out announcements every day. I mean, Apple announced their home initiative today, which is actually built on ARM chips. That would have taken us to respond to Apple's, you know, announcement today. That probably would have taken us maybe a few days before. Today, it literally just went live about an hour ago, and we were able to respond immediately so I think it's just the pace at which you can respond and also get in front of the negative you know commentary out there much more quickly than you were before I think I was just going to say I I'm not in comms but I was saying earlier yeah no they're my close cousins so we spend a lot of time together we don't write for the bots but I'll tell you we're very mindful of what the bots will pick up I'd be curious to know how you think about that you I don't think You want to be writing for them, but you have to be thinking about what will AI pull out of this, and are we okay with that? Is there a different way to approach this if we aren't? Yeah, and I think that it's – there was a period of time that I'm sure you're familiar with. A few years ago, it was sort of like, oh, my God, traditional media, nope. Everybody's just like, all you need to do is focus on social. Social is where everybody's consuming information and news and blah, blah, blah, blah, blah. And I think in a lot of ways, the LLMs have actually been, have taken us back to the everything era, where, yes, you need to be on social. But those large language models are training off of your work. That is how they are getting educated. It's not, in addition to that is often what folks on social are bouncing off of and talking about and is generating content. But, like, the so-called earned media is more important today because of the LLMs than I think it was even three years ago. You're giving me a little bit of hope, so thank you for that one. Thankfully. And, Erica, I want to go to you next because Yum has been piloting some really cool things in the AI space, which I'd love to hear you talk about, too. But customers have put them to the test, a lot of these things. stress tested it in ways that I think no QA team could even have dreamed they would. And including some deliberate trolling. I recall seeing someone ordering 18,000 cups of water. Like, what do the pranksters teach you about all these pilots that maybe the test didn't? Listen, this isn't new. Customers always teach you about your business and about your product, right? So we just have to shift our focus to learn faster in this environment yeah we we have a lot of stuff going on especially at Taco Bell in the U.S. with our voice AI drive-thrus we're in about 900 by now and the customers are teaching us that it can really elevate the customer experience and we can bring some things to customers that we couldn't before through that but we we learn through trial and error right so there were some hiccups and you can go read about them just like the one that you mentioned but the important thing for us is to keep in mind that that's a learning process and that makes us better and one of the things that we did at YUM a couple of years ago is our then CEO had our senior management take a course with someone from the Harvard Business School and his intent was for all of us to be digital forward to make sure that we were all approaching our jobs with that digital mindset and it really helped galvanize us because when something like that happens then the organization isn't like oh I told you guys that we weren't ready for that that's scary customers there's gonna be backlash we all said okay right understood that happened now how do we pick up the pieces and do it better next time and so having that mindset really helps when you hit those blocks that you will when you're innovating as quickly as you have to in this space and have the whole organization behind you and on On the positive side, is there any innovations that you recently rolled out or you're testing that you're really excited about right now? Well, I'm going to go back to voice AI in the drive-thru because there's something really cool about this that I think helps when you're having these conversations, especially at home. I feel safe here. We're all talking about AI and technology in a positive light. And then I go back on my socials last night and I'm seeing all the data center noise and all the things that people worry about with AI. And there's that constant tension, right? One of the things that we've learned through the drive-thru experience is that AI is actually increasing retention among our team members in the restaurant because it's making their jobs easier. It's taking away the tension points of the ordering process, the loud headsets. They're able to focus on the job with the product and the hospitality with the customer. So your first instinct is, oh, that's taking the interaction away from the customer. It's actually creating space. our team members to have that moment with the customers and has been a really important learning for us. I'm going to turn to the audience after this, but I mean, I first want to go to you to hear about what innovations you're testing out, trying right now that's really exciting you. One of the things we did back in March is we actually repositioned the company and evolved our business model. I can't tell you, usually it would take us probably, I don't know, six months or so to to kind of think about, hey, what's going to be the reaction? How are our customers going to react? How do we stand up a campaign? I mean, you guys probably all know all the different facets that go into repositioning a company. We did it in 60 days, everything from testing out the message testing, segmenting, who are we going to go after, our pricing, because we had to reset our valuation to the street. So how do we model from a financial perspective that campaign sort of after we announced it how do we continue longevity and make sure it has a longer shelf life everything we had done with AI now that's not to say that we had you had no humans in the loop we actually had you know a huge team behind that but i think it streamlined again making sure that we could do this and stand up something so quickly uh and i think you know what we see is uh from kind of the promise of AI i was at Nvidia earlier when the ChatGPT moment happened and you know there was obviously a lot of promise and still continues to be with the large language models we're now moving to kind of a world of agents our software engineering team i would say 99% of them use pick your agent to help with uh with debugging of software code which has asked which has led us release products into the market much faster than we previously intended. What I'm waiting for is sort of that next iteration of agents to kind of connect all these tasks, because it's still very much like take a small task and, you know, repetitive small tasks and throw an agent at it. But what you don't have is like connecting all the different facets of, you know, your function or the company together in a way that, you know, really helps streamline a lot of the productivity. Like I started using Muse and, you know, I think it's great for a lot of different purposes, but it's still quite not there yet that, you know, where I'm not checking my email and have to have Muse, you know, respond on my behalf. Like I still have to go in, gut check it, make sure it's verifying all the right information. So I think by the end of the year, over the next year or two, we'll get there. I keep saying that there's probably in three years time, there's probably going to be more agents than there are humans, you know, checking the Internet. So that's how many agents I think we can imagine, whether it's software agents doing verification work for coding, whether it's personal agents being sort of our personal assistant. But I think that's sort of going to be the evolution over time. So we've seen a huge value add with kind of the repositioning that we did back in March to streamline a lot of kind of our day jobs and allow us to move much more quickly. But I think that promise of doing everything, and I don't think it's ever going to replace someone's job. I think it's just going to be additive and help with human productivity and allow us to do things that we probably weren't able to do before. And so that's what I think will happen over time. Questions, comments, thoughts, fears. I'm in the back. I'm curious how you think that changes the ultimate end state of the business model where you're right now, your customer bases are very, you know, human oriented. And when even if you just use the example of like personal assistance, that means that there's like an agent acting on your behalf like how much does that change your marketing strategy how much does that change the way you're building your systems and your interaction points for customers I think for us it's about customizing right creating that custom interaction for that person and exactly what that looks like in a year two or five years I can't tell you but it's still about that connection with the customer and it's leveraging these tools to make sure that we have that ability to engage with them in our own unique way so it's bringing our unique brands to that unique person and bringing it together in a way that we haven't been able to yeah and i think i mean i agree if any of us knew what this was going to look like in three to five years um we'd be on Wall Street or Polymarket right now um but i i'm very rich Exactly. I think for us, part of how we are thinking about it, and it is going to change and evolve with agent commerce, is it's not about telling a customer who's opened up the DoorDash app, like, oh, my God, we're going to use AI to help you. It should just make the experience more delightful and more value additive in the moment. We have DoorDash Ask. So now you can just hop in there and sort of have that conversation, and we are seeing that means that 50% of those customers who are using DoorDash Ask are ordering from new restaurants, for example. Or if you're a local business trying to get on the DoorDash platform, we have a ton of technology that is built into that process to help with the self-service, and so these restaurants and other local businesses are getting onto the platform 35%, 40% faster. but we're not doing it with like flashing this is AI and it is helping you, it simply ought to make the experience better and more delightful and customer centric as opposed to being about us and the technology that we're delivering like that's kind of not the point, that's the input but really the focus is on the customer, i think that you know i sort of look at what the world of marketing will look like three years from now and you know i don't know what it will look like but I'm anticipating there's going to be a lot of change. And what I know of the function today won't actually look like the same function three years from now. I actually think it will be broader from an expertise perspective. You know, we're orchestrating kind of our org chart that there will be agents, you know, across our entire org and the different functions within the company. I also think the customer experience will change. So today we all interact with apps. So take the DoorDash app, for example. we, you know, obviously order through the DoorDash app, I think that experience with a bunch of agents sort of acting on your behalf will change. I don't know if it's going to be an app or I don't know if it's going to be, you know, a common UI across a bunch of different apps underneath. I have no idea what that's going to look like, but it's definitely going to change. And I think we're all thinking about what that looks like as the world evolves towards, you know, that change. We have a microphone coming right over there. And if you want to say your name and where you're from. Hi, Annie Young-Scribner. I'm on the board of Yum. Good job, Erica. So question for you guys, the pace of change is happening faster and faster. Transformations for every organization needs to take place. What's one or two best practices that you've deployed in your organization to help with that? I can go first. One of the things we have, you know, which I probably didn't exists before was we have AI engineers in every function. So it's not centralized, you know, there's obviously product engineering, which is sort of centralized with those are the products for the company. But for marketing, for example, we have AI engineers within our function that are deeply rooted and ingrained with sort of the AI mindset. And, you know, these tools are changing all the time. So it's not like they have expertise or experience for many, many years, but they have the mindset. And then we partner with OpenAI. OpenAI and Claude both have fellows where they sort of can be deployed in your company. And then they have best practices that they sort of bring along with that. So for us, OpenAI is a close partner. And so we partner very closely with these OpenAI experts. And then we have these AI engineers that sort of train and upskill the entire organization around sort of that AI mindset. And so we do hackathons once a month where we sit down together and it's literally, you know, hands-on keyboard. Like, this is what I want to do. Help me build this workflow with kind of these OpenAI experts and the AI engineers on our team. That's helped a lot with, because a lot of it is what we see as mindset shift. And so that's helped a lot with, like, not being scared, kind of understanding how these tools can be helpful, not hurtful. And so that's something that we've done that I think we've embraced across the company to really help adopt AI. The mindset and practice is so important. So we're always going back to these like training moments. I mentioned the Harvard Business School example just three weeks ago. Our global legal, not legal, the global leadership team of the whole company was together. and we dedicated an entire morning and had trainers come in to work with that team in all different functions these are not technologists these are lawyers these are marketers these are folks in HR to work with us on agents and we spent that time and dedicated that really precious resource to making sure that our leaders were learning but also it's the mindset so they understood the importance of this and how each of us is responsible for incorporating this into our functions and into our daily work to really transform the organization and be as quick as we can to adopt the tools and to use them you know for the business yeah and i think similarly um you know so as a for example at DoorDash we have built it into performance evaluations AI fluency is a requirement, and everyone, you know, we just finished sort of the mid-year cycle, we'll have the annual cycle, everyone that is among the criteria that folks performance is based on, and everybody has a clear understanding, sort of like, what is entry level, but then what does excellence really look like? And it involves things like, how are you incorporating agents? What are the new technologies that you are bringing to the team that are helping you and your colleagues think about different ways to leverage this technology? You know, I've mentioned, like, yeah, sort of first pass at a press release or first pass at a briefing document or sort of some help with the blank page challenge. That's entry level. It's, like, it's what comes after that and how are you then really thinking about getting enormous leverage out of the tools. And then, yeah, I think almost every off-site we do has some kind of hackathon at this point. Most teams, my teams included, have, we've got, like, a huge Slack channel that's just about AI tips and tricks. Like, hey, has anybody used this product? What have you found to be valuable? Hey, guys, I actually ran this experiment. This is what worked. This is what didn't. And so that's folks across 30-odd countries around the world sharing all these kinds of things on a weekly basis, if not a daily basis, that are specific to us because the ways in which we're going to derive value from it across communications and policy and government affairs is probably going to be different than how the product team is. and so there's some like really being able to use the specific use cases and the things that each person on the team is trying to accomplish versus having you know technical folks would be like this these this is how it works for me it's like well my team doesn't code but we do a bunch of other stuff and how do i actually leverage whether it's Muse Claude um any of the tools to be better at what i'm doing we got a lot oh wow yeah hi i'm sandy carter i'm a ceo of a company called Equis i'm an agentic ai company um one of the things i'm really curious about we were actually in my town we were just doing an ai governance meeting um with community and a bunch of people tagged all the cars that were in the parking lot because they knew there was an ai meeting there so there's still a lot of backlash to ai and i'm really curious um erica yes i'm really curious because like when you're when you're now training workers in a restaurant who maybe really don't understand ai has the story changed on how you train them and then also for doordash do you share those tips with your drivers too like talk to me about out of the corporate office and down to like the real workers how do you guys tell that story and has that changed yeah so i think it's different you know when you're what we call above restaurant and the the training and the mindset that i've been talking about that's where people are accessing the tools and doing so with the intention of using ai for their particular work or project once it gets to the restaurant level we've embedded it into the technology right so the training isn't so much about the appropriate use of AI at that point the training is about how to use the tool and so we help them understand and are are clear about what the tools do and that AI is involved but it's talking to them about how this is going to help them engage with the consumer and do their jobs so it's a little bit of a different discussion depending on on where you are from a governance perspective like I said you know when I got home from the dinner last night I on social media and seeing all the kinds of things that you're talking about. And it was such a shift from where we've been, right? There is a divide that we all need to manage through. And part of my job in legal is to make sure that we have the appropriate guardrails so that when we are having these sessions and we're talking to our teams about AI, I can credibly say we've done our job. We've done our job to make sure that this is safe, that we're using information appropriately, that we're training people to do it the right way. And that is really important. it's the foundation of everything that we do and you can't lose that because you need to have that to have the credibility first talk to people in your organization and then when you're talking to people in the community yeah and i think similarly for us when we think about it with regard to say dashers it it is that it's embedded in the product like Erica's talking about so as a for example um we have a new product for for anybody who's like i want to be a dasher i've never done it before there is there's kind of this cold start challenge people it can feel intimidating of like oh okay you know how do I know to like deliver it to the right place or take the photo correctly or you know how many hours should I be working or where you know where should I go to be able to get the the best delivery options that I can choose to accept or reject and so we built a tool called Dash Buddy just for new dashers and there's a ton of AI technology that is embedded within it and you know because it's everything both in the dasher app you can text it There's a bunch of different ways that you can access it, and it can produce answers in real time and have a dialogue with that person who is interested in starting to dash in a way that scales beyond what we might be able to do with customer support or dasher support, and you know we're seeing great results. About 95% of questions get resolved in real time, and that then enables folks to get on the road and earn what they want, but again, it's it's embedded within the product versus sort of having a conversation that I think is a bit different about like, you know, hi, we're excited that you're interested in dashing. Here's an AI tool to help you do it. It's just like, it should just be easier and better. Hello, Pamela Passman from APCO Worldwide. We're hearing from our clients that boards, C-suites are asking them to be more predictive, to be more proactive. Whether Whether it's geopolitics, it's public policy, it's risk, it's marketing, consumer trends, are you using AI in your various roles to be able to be more predictive, to be able to hopefully look out and understand different trends and risks and opportunities two, three, four years out? You want to start with that one? Yeah. So I sit on the board of Corning, so I have sort of a board seat and then also obviously run a function within a company. And that's absolutely, I mean, something I do pretty much all the time, which is, and I think, you know, from a marketing perspective, I think the best marketers are ones that can anticipate. You know, I always say you market one year ahead. We don't know what the world's going to look like a year ahead. But you can predict, you know, based on what's going on in the world, how the world might be shifting, how the landscape is shifting, how your competitive landscape is shifting, et cetera. So we use predictive analytics both within the company and then at Corning, we use it a lot. It's a manufacturing company and we use it a lot to figure out, you know, how do you improve manufacturing because so much of the company is around manufacturing. And so how do you get products to market faster based on better manufacturing operations? And so when we look across the companies that I'm involved with, one of the things we think about is you can have every employee use AI, but what are the big levers that are going to give the company lift? And for a manufacturing company versus a software company, those look very different. And then once you figure out what are those two or three levers that are going to help improve company productivity or help unlock some innovation, then we spend a lot of time figuring out with predictive, thinking about from a predictive perspective, how we can help improve those different, whatever those are. And so we think about it all the time and we use different AI tools for predictive, whether it's macroeconomics, whether it's supply chain. A lot of those aspects are specifically for my job today. That's kind of how we use it. I want to be able to get to more, but I see one over here. Hi, Michelle Peng. I'm a senior reporter at Twitter. and i was wondering how y'all are thinking about developing entry-level workers especially as AI takes over those messy slow tasks that previously was where they got their reps in yeah so i don't have an answer yet this is something i think a lot about as the tortured young associate who learned by doing all the grunt work right like this is how you become a lawyer at least in my experience so it's something i've been thinking a lot about um i'm hopeful that AI is going to help us, right? And help us train people and help us solve that issue. I was telling these folks when we were getting ready for this panel that when I think of the narrative around lawyers and paralegals and that being one of the jobs that's going to go away, I think about the first partner that I worked for way back in the 90s. And he was teaching all of us about the billable hour, you know how to track your time what you you know how you should be spending your time and and he said something that stuck with me he said don't hesitate to write down your feet on the desk time uh and he he was saying that you know when you lean back put your feet on the desk that's when you do your best thinking and you should charge the client for that you should absolutely charge the client for that because when you're a young associate you think well i must be writing something or forming a corporation or something for it to be valued to the client and it was very important for me to have heard that from him to have the permission to charge my feet on the desk time and so I think about the AI and how that's changing the jobs for people in my profession it's creating more space for that feet on the desk time but you've identified hey but how do we get people skilled up to the place where they know how to do the basics if they're not physically doing it, I think we're just starting to scratch the surface. I think those of us who know how to do that work and know how to do those jobs have a responsibility to work with new entrants into the workforce to help them and to make sure that we're not skipping steps with them. But I would say that it's a challenge and something that we need to be solving for. And Liz, maybe in your answer too, because I want to add on to that question. I believe I was talking with Bex over there at Okta yesterday about AI burnout and fatigue, especially with young workers and for me I was going back to like the last Friday of every month I am so excited to do my expenses because I can turn off my feet on the desk time and and just zone out and listen to Broadway and and do it and I feel like AI is going to eliminate some of the that task time so like how do you balance both I would love AI there are some things too complex for you yeah yeah um so i think maybe taking this in a few parts firstly to directly answer your question i actually i will be honest i worry less about the the younger the more junior folks on my team right now because many of them are already excellent writers they may not have the reps that the more senior folks do but they you know part of what we work on is recognizing what great looks like what does great writing look like what does great content look like and helping them learn pattern recognition and those kinds of things and so i think for those folks particularly who are more sort of often they are digital natives they are more ai fluent i think they are actually set up to succeed the place where i get worried is actually the entry-level workers five years from now six years from now where so much of their time in education has actually been powered by AI and so they are not doing the writing in high school and college that teaches them what great writing is. You know, I'm not an academic and but have a lot of friends who are and so like return to blue books like all the kinds of things but the thing that looms out there that concerns me sort of more is like for those folks if they don't, if they are entirely reliant upon AI tools and aren't able to understand, as a junior associate, what makes a great litigator or some of those kinds of things. I don't have an answer to it, but that is probably the thing that I worry about. But I do think that fatigue is very real insofar as there's just ... I think the important thing for at least our teams is to differentiate between using AI for the sake of using it, which I think is part of where the fatigue comes in, versus actually finding ways to make you productive and more efficient and gain leverage. Like, there are days where I will be honest, I don't use, deliberately, like, use any of the tools because I'm in a bunch of meetings, and maybe I'm in meetings where, like, I have to turn Granola off and a variety of other things. And then there are days when I spend most of my time kind of going back and forth with Claude. And so I share that both here but also with the teams, is pick and choose the moments when you're using the tools. You're going to have agents eventually running in the background, but not every single task probably requires the use of a tool deliberately. And you're going to want to practice and get better at finding the ways that you can use it strategically versus just using it habitually for no real output. That makes sense. And unfortunately, we're almost at the end of our time, but I'm going to give you the last thing to kind of build on what Liz said. Is there anything you are now refusing to let AI do even though it could I think judgment you know I think using good judgment uh and curiosity those are both things that I don't think AI will ever really replace I'm sure you guys are seeing it so you know you sure you can prompt or you can ask an agent to do whatever task you want but ultimately it's judgment that's gonna you know make you go from good to great and I think that is something that AI can't replace the the conversation about you know younger workers we're actually or entry-level workers we're actually looking at different programs so like USC has an AI business program which they combine business and engineering and they have a very specific AI program and back to the mindset we're really hiring much more for mindset because i i often say that uh it's easier to hire for culture than it is for skills you can teach someone the skills on the job but it's a lot harder to teach someone the culture and so we're really looking at people who have judgment and curiosity because I think those two things are super important and that job will change but if you can apply judgment and curiosity in ways that and creativity is probably a third in ways that AI can't do I think it really sets someone apart in terms of what they can do and what jobs they can do in the future absolutely well thank you three so much Thank you all for joining us, and thank you, DoorDash, for hosting this session. Big round of applause. The main stage conversation is going to start shortly at noon in the ballroom, so please make your way there now. Thank you.