Integrating AI at Scale: Stress-testing your organization - Hadley Ferguson, Uber

Speaker A: There's a bit of just walking it and kind of going along together and sharing the failures and sharing the learnings and then as a leader, building the trust with other teams around you so that there's support externally, especially for teams that are like service team maybe who are doing something for another group. Like I want them to feel comfortable that I've got their back, that all other teams are like expecting this and they're not going to be blindsided. Try to make things the absolute best. Like our vision for our team is creating world class knowledge for human machines.

Speaker B: Welcome to the AI Edge podcast for Enterprise Marketers. A show dedicated to sharing insights, strategies and experiences from a group of experts who have successfully implemented AI solutions in a large enterprise B2B software company. Uh, specifically in the context of global marketing and how that effort can connect to sales, IT product and the rest of the business. I'm editing Porter De Leon and I'm joined by my fellow host, the amazing, the wonderful Jessica Ria. Jessica, what have you been doing? Where in the world is Jessica Ria? That's the game I always like to play every episode.

Speaker C: I am at home and I've not been on a plane in almost a month. Can you believe it?

Speaker B: One month plane free.

Speaker C: I'm actually really enjoying it.

Speaker B: That is amazing.

Speaker C: Yeah, the kids sports are all overlapping right now, like three to four sports plus end of season all star stuff. So it's been crazy. And I'm loving my over scheduled kids life.

Speaker B: Yes. Just immersing.

Speaker C: So I'm like, I was the creator of it and I enjoy it so it's good to be home.

Speaker B: That's fabulous. And so there was one of those events, I think your laptop took a water bath. What happened there?

Speaker C: Yeah, well don't put your laptop in the same bag as your water bottle. Especially if you put the water bottle back in your bag like when the lid's not on all the way. So just little life lessons there from this podcast.

Speaker B: I know I did that once but it was just my kid's jacket. But then she was cold so that didn't work out well. So we do have a special guest. I'm super excited. Hadley Ferguson. And Hadley Ferguson. Before I let her speak, I'm going to do a whole little intro. Hadley Ferguson leads global content and Uber's community operations, overseeing a team of over 200 professionals who develop and maintain the critical knowledge resources powering Uber's worldwide support system. She has like people all over the world so her slack is like blowing up. With over 12 years at Uber. Hadley has significantly shaped the company's support experience. Managing content for support agents, self service help articles for customers and earners, and content optimized for AI driven support. She has been instrumental in supporting the implementation of generative AI tools to deliver personalized, contextualized support experiences that meet users precisely where they are. Hadley, now that I built you up a whole bunch, welcome to the AI Edge podcast.

Speaker A: Welcome. Um, thanks. Do we think AI wrote the intro for me? What do we think?

Speaker C: I mean, I think that's what this is all about, right?

Speaker A: Right.

Speaker C: I couldn't tell. It wasn't overly AI there.

Speaker B: No, I think I grabbed it. I grabbed it, I think from a source somewhere. I think they grabbed it online.

Speaker A: Did you?

Speaker B: Yeah, I think this is written somewhere.

Speaker A: I'm like, did it come? I don't know. Anyway, it might be. Who knows where it came from?

Speaker C: But did you like it? I mean, who cares if AI wrote it?

Speaker A: I love it, I love it. It makes me sound so important and so accomplished.

Speaker B: You are important.

Speaker C: I kept thinking about how easy it is to get a refund with Uber if something happened, and it was the most streamlined experience I've ever had. So thank you.

Speaker B: Wow, that's fabulous.

Speaker A: Amazing, amazing. We work a lot on those automations and right now those are handwritten, not AI automated, but there's a lot of machine learning behind them. But yeah, our content side is still done the old UX way and we are actually figuring out ways to make it more seamless and make it more user friendly.

Speaker C: Wow, Kudos.

Speaker B: Ah, that's fabulous. Well, your NPS score on this podcast is perfect right now, Hadley.

Speaker A: So amazing. I'm going to take that back and see if it can go into our CSAT surveys. No one star csat.

Speaker B: So listeners of the show know that each, uh, show we like to go over the following. You go for topics and objectives, strategy and tactics, teams and tools, business impact. At the end, we do a lightning round. So kicking it off for the objective today, it is really about leveraging AI is one of the things that people talk about a ton. Like we want to leverage AI, we need an AI strategy. Or we need to build AI into our strategy, or we need to be leveraging AI in our workflows then. So rubber meets the road. And both of you, Jessica and Hadley have a, um, lot of experience around this. So to you, Hadley, what does it really mean when your organization is leveraging AI?

Speaker A: Yeah, I was actually thinking about that. I was with a different leadership team this week. I was moonlighting on a US and Canada leadership team to understand what they're up to with some partners more in the business side of Uber, we were still within support and so it was really interesting. I thought about this question and I was thinking a ton of what it means for my team and like, why would we really do it? Is it to be faster? Is it to remove a task we have to do and automate it? Is it to make something better and ultimately make a user experience like Jessica's refund experience so pristine because the way that the copy is written, it's world class? Or is it a mix of all the things and is it also to make people excited? And so I was thinking about it because I was listening to how these teams are kind of wrestling with it as well. And you know, their use case is different, be program managers, um, in different roles. But it got me thinking for my side, what is it really? And I think there's kind of two things. There's one, we have this part of our strategy where we say we really want to develop like world class talent and we want people to want to join our team to become experts in the field. Like, come to content team with an Uber and leave and you'll be a master or something that's going to get people there. And then of course, we do want to make our customer support experiences so much better. Experience is just such a big piece of what we do. And I think because we grew so quickly, it was hard to master that and to even understand, to really kind of fine tune it and to understand what it should be. So when I think about it, there's really two pieces. Like, I really want our teams to, um, take AI as their copilot, to like, overuse that term. But truly, especially as a content team, there's so many things you can do to make your experience with someone you're partnering with, like with the business or in the work you do easier. It can be quicker. You can get to like that first round quicker. You can then get to something more important after that and kind of move on. So I really want it to be that they can grow skills in something that's new and transformational, because this is going to be a transformational time in everyone's career, no matter what you end up doing with it. But I want to instill that in and like, how can you use this to your best advantage and come and really work on those skills? And then the second piece is how can we use it to really push the envelope of making things that much better? We do huge scale in what we have to do. And so I think when we look at it from that perspective, it will evolve. But I really feel like your point on the comma Gemini, finding a comma mistake. We want to find every single mistake. Why should anything ever go out the door again that doesn't have the tone we want it to have, has a term that's not really fitting in where it is. And I think that there's so many tools out there to kind of assist with that. I think we really definitely value the human in the loop. But I think there's a huge part of leveraging it.

Speaker B: Yeah, I think there's so much in what you're talking about there. One of the big things though is I think building that into the way that you're operating I think is really interesting because like you said, you want people to come in and they want them to have sort of this world class, upskilling, level up experience so that they then go out into the world and do even greater things because they're masters of all this stuff. But I think there's a lot of stuff that needs to change within organizations for that to actually I've got experience with that where habits have been one of the biggest barriers in changing workflows, in implementing tools, in truly leveraging AI, like making it a part of how you work. And I wonder if you could talk about that because I uh, actually was just noodling on that on the last couple days were just habits. And I just see people who are deeply technical or love change or innovative but have not changed the way that they work and think and build and connect just because they've built habits over their careers. I don't know if you could speak to that. And some of the other things too that you felt needed to change within your organization to really leverage AI.

Speaker A: Yeah, this has been heavily on our mind too because I think we're, we see people use it a lot for parts of their job, but it might not be what we intended it for and not in a bad way. More just you might see someone using it more to write emails versus you know, you're a content creator and you're not using maybe m the tool that's in your toolkit to write the content. Super interesting. Like what is it? Is it you use it Once you lost trust quickly, you were like, huh, I can do that better. Is it uh, concern, like different process this is going to be. How's my job going to be? So I think we've got a really big mix of Those things and the pieces that I've really thought about, like what needs changes need to take place and make that happen. We are in an operations zone. So I do find that people, myself included, you kind of get really stuck to process. Like it's such a comfort zone. And you're like, well, that's what I do. I create a structure in a chaotic world and that's my livelihood. And so I think stepping outside of that to use a new tool can feel really scary. So some things that I've been thinking about is trying to push myself as well. What are the skills you need to do as? Is it more like critical thinking? Slash a little bit of the growth mindset kind of skillset, mindset type of thing. And then also that push and that, uh, support to feel like you can take a risk even if it isn't a risk. Well, on a broad scale, maybe it's not a macro risk, but maybe in someone who's in a very different position than me, their risk looks so different than what I might say. You know, mine might be like, let's do something crazy and go test a whole new thing. And I can do that. Theirs might just be to use AI for a request that someone has shared with them about creating an automation question, like, what if it doesn't turn out well? What if they get mad that I use it? What if they don't like it? And so I'm trying to put myself more in the shoes of the teams that are using it and think through, like, how can I help? How can we kind of break down those barriers? I think for leaders and this really feels like it trends like way outside of just what I do. Because I think we've overall as a company too made a big commitment of like, we are going to be a tech driven, tech transformation driven team, like AI driven. Like we need to be using it for everything. So when you ask yourself, how can I do something two times faster or whatever, 10 times faster, how can I do it tomorrow versus next month? How would you do that? What would you use? And so I think for my team, I'm trying to show as well how we're doing that and think about sometimes it's like the critical thinking skill. Like, okay, if I do need to do it quicker, what would I have to change to do that? But I think there's a lot of things and I'm trying to push myself to kind of model the behavior versus relying on a change management plan or like a really beautiful, like, we'll be here and then we'll be here and then we'll be over here, uh, then the later and it's this nice people evolution. You've got skills, you've got process change, you have marketing. I think that's helpful too. But there's a bit of just walking it and kind of going along together and sharing the failures and sharing the learnings and then as a leader building the trust with other teams around you so that there's support externally, especially for teams that are like service team maybe who are doing something for another group. Like I want them to feel comfortable that I've got their back, that all other teams are like expecting this and they're not going to be blindsided.

Speaker C: It's such an important aspect of all of this too. And you mentioned change management and I agree there's no like set time frame, timeline. There's a roadmap for getting things done, for piloting and scaling. But change management and change leadership is ongoing and it encompasses all of these things. Communications and all of that drives culture, which is a big part of it and this sustained effort from all layers of the organization. But I think you hit the nail on the head with the leading by example piece of it all, is that I feel like I say this a lot and we're, I see this a lot, but I still feel like it's not really taking hold yet in the majority of organizations that leaders need to lead by example and they're a big part of it. And while this can be a bottoms up, a middle out and a top down movement, it has to be all of those things. It can't just be leadership saying that you have to do AI. And I still heard uh, that this week from a brand that we're getting pressured to do AI. But you can tell the leaders don't really do it themselves. So they don't know how to talk about it. And so I think leading by example is just such an easy yet such an important way that leaders need to show up for their teams and really drive that culture. And just as you said that it's safe that it can have an impact in understanding what type of impact that can even have on your team.

Speaker B: Yeah, there's two things. One is that the leading by example, because I personally was experiencing this and I think we've talked about this before, Jessica, we're like, okay, CEO says everyone has to use cursor now. I don't care like what you're using it for, but you got to use it, you got to generate code, you got to build Stuff actually he's built like an 8 bit video game which was super fun, where you're like this little person you jump through. It was super fun. 15 minutes. I have a video game now.

Speaker C: Congratulations.

Speaker A: That uh, is crazy. I gotta tell my husband.

Speaker C: I don't have a video game.

Speaker A: My husband's gonna freak out. That's where I'm going after this.

Speaker B: Yeah. Gonna cursor create a video game. But the thing is, well, there has to be something behind it. So there was sort uh, of this gap on sort of this air pocket which kind of stalled because. Where's the follow up? Okay, great, everyone has cursor licenses. What do we do? But then it built up. Then it was like, then it was bottoms up. Like you're talking about Jessica, because it hasn't come from everywhere. Then everyone's like, okay, we're building stuff, we're building demos. All the developer relationship people started building demos and doing demos and showing everybody. And then other people just kind of came and said, hey, this is what I did with it. And there was sharing. And then that created the momentum. So there has to be sort of, hey, we've got top down and we got bottoms up. I like middle out. I'm still trying to wrap my head around what middle out is. Maybe you can kind of go over that Jessica as well. But then the other piece is the sustainable part and I think you mentioned that. What's that sustaining piece of it? And that's the tough part is like, hey, I did this one thing. Like someone said, I wrote this script or uh, use AI to write a script. And it did this one little job for me. Great, that's a one off and that's great. You're using AI. So I think there's a big difference between using AI and leveraging AI. So those two are very distinct. And there's a lot of using, I feel like in people who are embracing it, figure it out. I feel like I even fall in this category. I use it, but I don't feel like it's super, super integrated to the processes like you're talking about. Where are those processes you said Hadley, like that we hold on to and we love. I don't see it yet. Constantly built in to the processes of we do it this way. Then this AI thing does this thing. And then there's automation in here. I see tools like, hey, it gets loaded into this database or it gets loaded into this thing. But there's not that, hey, AI is a part of how we do this because like you, I Think. Both mentioned there's people who are afraid that someone might get mad at them or they feel like they're not good enough. That's why they're using AI. Or some people even maybe you got, you guys experienced this will say, this is my superpower now. This is. Now I've got this. Why would I share it with other people? Because now I can do this faster and do this really better and like, and I'm not even going to tell anyone I'm doing AI that. I'm just going to let everyone think I'm like super amazing or something like that. But yeah, there's challenges.

Speaker C: Well, and Hadley, you mentioned this idea of the co pilot, the thought partner. Right. You might say, and I love the upskill motivational message you have for your team and it almost feels very employer brand for you in terms of getting people on the content team to upskill them for their career. I mean, I think that's an amazing employer brand and strategy, uh, for your team. But I think it's easy in the co pilot thought leadership sense to get lost in that ad hoc in the really experimental. I still feel like it's experimental even though you're using it every day. It's like, oh, let me try this, let me try that. That cursor experiment for you had nothing to do with your job but was learning. But like, how are you getting your team to anchor back into the main jobs to be done? Because I think that's where you find those repeata parts of the workflow.

Speaker B: Yeah. And I think that we're starting to lay into the second part of the show, which is strategies and tactics and asking the question, what is the strategic intent of what you're doing? What is your strategy? And the part of it is people just haven't really laid out their strategy, really laid out their process. Well, so then how can you apply AI, uh, to it if you actually don't have something to apply AI to?

Speaker A: Yeah, that's actually a funny thing. I was thinking on one piece you said too. And I wonder for some of these big companies that are the biggest ones pushing it, Every tech company, every brand, probably every company, but many are trying to be. Yeah. Use it. They can accelerate. We can do so many things.

Speaker B: Well, Hadley, I work at Salesforce. We're not trying to do that at all. Just so you know. That's not, it's not part of what we do.

Speaker A: No, nothing. You have no strategy. I haven't gotten sold anything from you recently. I wonder, you think about it and like the reason people ask M me all the time, why are you still at Uber after 12 years? Like, I get it all the time. I do a lot of new hire things and I'll talk to people. It's like one of the biggest questions. Like, 12 years, wow. And, and I always say it's the people and it's the really smart people. Because I feel like every day I'm learning from my own team, I'm learning something from someone else. I'm challenged. I'm like, oh my God, you're so smart. How do I hang out with you and learn what you do? So it's like business school on steroids. Every day of people's lives at these tech companies. And so I wonder on, um, that one. Closing on last thought. It's a tough one too. When you're really smart. It's tough to feel like I think about this all the time. When I write an entire brief speaking notes for a presentation using AI, I feel a little bit like, am I losing intelligence? Am I not going to be able to do it again if all these tools just went away one day and a new thing happens? Do I have that skillset? Am I losing, uh, knowledge? I was a runner for a long time in my life and I set repetitive motion. It's like you gotta grind to get better. And so I do wonder if that's a thing too for some of these big companies and like ourselves, like a lot of people, your knowledge is your brand. And so how do you bring AI into that in the future, but outside of the strategic intent, of course, of what I'm doing?

Speaker B: No, but I think you're pulling on the right thread, Hadley, because the value of an individual, like, what is their power? What is their value? What are they bringing to the job? Not necessarily. Like, you're a really good researcher. You're really good at going on the Internet and looking stuff up and, um, synthesizing information. Maybe you're really good at asking the right questions. You're really good at finding the strategic intent, like the who's it for and what's it for and how do you know it's working part of it, and you know what good is and you have an instinct for what is going to help tell a story or help create a market or help improve a workflow. But that's a weird sort of nebulous thing because people are like, well, how many years experience you have in this system or that system? What have you done here? And what you really want is proof of impact of what did you do? How do you do it? And then what was the result? We're not really good at measuring that. When we're looking at people, we just look at sort of these proxies and I think people need to let go of like, hey, I don't need to be really good typist or I don't need to be really good speller because that job's done for me. Am I losing that skill? Does that matter that I lose that skill? Is it okay if I can let that go and just go fast? Because this machine is helping me do it. And the same thing is with using something like Google Deep research where normally check a whole bunch of different sources and I would cross reference and I would do all this stuff and is it okay to just let that do that part? And then I look at it and said that's bad, that's bad. Ooh, that's good. Let's dig deeper into that and then go that direction. But that's the skill. Where do you see that being a part of the dialogue in the community, within an organization of letting people understand that, hey, look, you need to move towards this as being the value create for the company and not worry about, hey, I'm really good at doing the managing of the chaos or I'm really good at spreadsheets.

Speaker A: Yeah, I think this one is a really interesting one. I don't know what you all feel like, how you use it, but for us, one piece, that's another one I'm trying to kind of instill and like talk about a lot with our team and even my direct group, it feels like especially for content, so maybe it's unique to us. For content. A lot of the skills you need to best leverage AI for what you want it to do actually are those core foundational skill sets of content that are the really big ones. Like, you still need to be a good content designer, you still need to understand really good UX flow of like or conversation flow. It's almost like maybe you're going back to the basics in some instances. But I think that that's the one piece I've been really trying to push on. Because if we think about the strategic intent of what we're doing, again, we need what we do to always continue to get better and it needs to evolve and then the better needs to get better because the better is every year it's going to uplevel. And I think that whether that's scale of what you do for us, like products, whether that's just like the world is changing and people have a different expectation because of many different factors. So that has been one really big piece. And so our strategic intent is to try to get better, like try to make things the absolute best. Like our vision for our team is creating world class knowledge for human and machines. That's it. Like we need to create something. And so the strategic intent of our team and how I think where it's like divided space a little bit where not divided, we're there but I think to just this point on like the tops and the middles and the bottoms, it's almost like I think we have a bit of a anchored North Star strategy for our organization that without Genai was kind of going towards a space that just enabled it anyways. The create ones publish everywhere model like super simplified ecosystem. We've got to get better. We need to leverage tech to accelerate like how we manage content. We need good metadata, good taxonomy. I'd already put that together which was awesome. And then you know, we had to tweak it as AI kind of ramped up. But the other piece is that we also have to say, well how can AI get us there faster? So when our strategic intent is to have that world class knowledge for human and machines, part of that is that we're now then like brain behind the machines that are being built. So we need to transform everything from its current state to its future state. And that's where we've, I think I would give us like a. We're nailing it as we're using it to do things 500 times faster than we could have done before. Like we're finishing something up this month that should have probably taken, would have taken two years and actually someone else in another role over the week. When I was in New York they were doing the same transformation with people and I asked why, I was like we're finishing ours in one month and we started in May and we're finishing at the end of May. What would have again taken us probably like two years but we're doing it in one month. Yeah, there's a long tail.

Speaker B: Wow.

Speaker A: Right? I know it's super cool. But it's using AI so we have this one part of a team and I think our strategy is being super accelerated by using it to transform, to distill information, to check. We're using it for assessments of things like risk assessments, different pieces. Like really trying to ask ourselves a question. How can you do it 500 times faster? Not two times faster, 500 times faster. And then the other Side is when we have the work that we do. And so we're still like doing our job, our normal job of we still need to create things for Jessica to do self help in the Uber app. And we still have to do that. And that's really, really important. And those are the biggest things, you know. So you're kind of balancing and why I say so strategic intent is again that North Star of like best in class. We're kind of now figuring out how you weave that into everyone's job. And so that's why I say we're kind of taking this like, what are the skills we need? It's kind of those basic ones, but we still want people to be really solid in those skills. Like, okay, maybe you are just like more so writing content beforehand. Like just really drafting now you need to be more of that designer. Your copilot is going to draft, you're going to do it. And even your intake from asking questions or getting things for a product requirement stock, that's all going to be automated because it can grab those things quickly for you and synthesize things better. And so it kind of always goes back to the improvement of the end experience and the end result. And so I think on the side of integrating it fully into everyone's workflow, I think you kind of said it, Jessica. It's like a bottoms up, middles and tops there where again trying to like lead by example, show these cool examples of how we're or show these instances of how it's being used, get people using it and then also kind of just like bring in the mindset of this will help us all move towards our goals faster and at the same time, I guess investing in some new skill sets. Like we're definitely trying to say if you're really interested in things like prompt coordination, we need to start doing that. Like we have to build and maintain our own models to build our content. So that's going to be a thing we need to do and we need to be able to review other models doing things for us. You need to understand guardrails. So I think there's that other piece of. There is a bit of literacy in some areas or skills in some areas that are. We're building in pockets and trying to kind of create it as like a have people filter through these roles to get the skills and then kind of go back into their spaces and then some stay if they're really into it. But that was a big discussion too recently of how can we make sure we're creating Kind of a internship program within our own team in a way to get that.

Speaker B: That's fabulous. I think this is great because it's a perfect transition to the teams and tools section of the show. I think having the skills part of it. And like you said, this internship like structure, rotation.

Speaker A: Yeah, yeah, exactly, yeah, rotation.

Speaker B: Dig into that a little bit. What does that look like?

Speaker A: Yeah. So again, on the side of like transforming what we have into this AI friendly version of itself, the uh, machine friendly, we kind of been trying to figure out our team structure and the ways we work to be able to bring people out of their day job if we need to and kind of put them into these strike teams and call it kind of like, yeah, little rotational program of groups of people who are either working on the transformation itself, working on continuous improvement cycles for that specific content once it's out and being used. Because once you have it in a new use case, you have that flow of information very quickly. And we want those spaces to be ones where you're unrestricted to timelines. Like you're going as quick as you can, you're able to really fail fast and learn. You're leveraging tools to do things and you're kind of like in the midst of all the change. So a bit of like maybe build and run. But I think of it as the build side of the team because it's kind of new versus the run side is like your regular things you're doing. So yeah, we just opened sort of like short term assignments and we kind of bring these groups over to pod, like these strike teams. We are hoping to just make it more. I think we can do more. Which goes back to also using AI to kind of unlock broader efficiency that we can still do our day job. People can do their day jobs but we can actually give them the space and time to go take this rotational versus it. Being like a manager might be like, ah, I can't. Like we have a huge Eats thing launching and announcement. We have to switch our Eats model in Brazil because we ramped down our eats, bought a different company. That's going to be so much work for our team. So much work.

Speaker C: Mhm.

Speaker A: And so if you think about it, that team person might say, I can't, I can't let anyone from m my team go right now. It's too much. We have so much going on. It's high stakes. We want to be able to make it where they can do that type of thing quicker. And that manager could be like, yeah, cool, I have a few People who really can transition over and work on these other areas for two months, three months, and I'll be fine without them.

Speaker C: It's a double edged sword for so many leaders because we have to acknowledge it is work above and beyond their day jobs. But at the same time it is work that is going to optimize and make their day jobs more efficient. So it's like if you can figure out this process now, it's going to make your work that much more efficient which then frees up time for you to do those other things and those rotationals. It reminded me a lot of our pilot lead process that we had at VMware and that I talk about a lot now. Right. Where you have to identify these pilot leads within these functional areas or projects. People who are willing to. In a lot of cases it is above and beyond. But I love this sense with yours that we're carving out time specifically for them in these strike teams where they are going to be the ones who figure out this process. Then once you have it successful, you train the rest of the people on this process and then it becomes process adoption. Yeah, we were talking about process earlier instead of AI adoption.

Speaker B: Yeah, uh, yes.

Speaker C: And so I think the more that people can recognize that these, how important these pilot leads are and I want people to think about this concept of like, can we actually carve this out for a role or half a roll maybe if we move this work to somebody else who has already optimized. I don't know how you execute that yet if everyone has a bunch of work to do, just like you're saying with your Brazil team. But I think it's really something to think about because that's what's going to allow, um, people to scale all that much faster. The sooner you can get those pilots proven and then rolled out to the rest of the team, the sooner everyone else gets to benefit from that new process.

Speaker B: It's funny because, uh, there's models that exist like that. When I used to be in professional services and we built something with like Apple or AIG or whatever, there'd be a ton of time where you talk about how it was going to implement it, what was the timeline, what is it for, how's it going to connect, what's the workflow going to be before the technology ever got in, that was months of planning and process workflow and connecting and to get everyone aligned and then the technology would drop. And usually that technology was the easy part. You would deploy something, rack a stack or you'd push out Uh, a build or whatever, and then, boom, you'd monitor it and then work through all that stuff. It feels like there needs to be something like that. I guess there's not that same culture in, let's say, marketing or operations or even support sometimes of, hey, what's that process? We need to get that right first before we drop this technology in. And if we had that sort of process and culture built into the way that we do things, there would be. There'd be lots of discussion of this is how it's going to flow. But a lot of times we're just like, well, when's this new campaign's got to go? We have to get this product launch out. This thing's changing, or this is switching over here. So we have to redo all this stuff. So we basically, we've set ourselves up for the situation, which we're doing so much stuff that we're not giving ourselves that workflow. Whereas in other cultures, let's say, whether it's IT Dev or others, they're like, no, we can't do anything unless we do the process. Like, we can't function. It doesn't matter. We can't just push stuff out. If we don't do the process, then things break or patients lose their records at hospitals or really bad stuff happens here. Like in marketing, they're like, okay, well, maybe the worst thing will happen. We'll have to fix a tweet, so we don't worry so much about that. So we just run fast. And then like you said, like, where's the time? Like, no, I'm a manager. I can't spare one of my people. So I like some of the ways in which you're addressing this, but I think the how is probably what a lot of people are interested in. How would you get people out of that. That hamster wheel?

Speaker A: That's so hard. Well, it's funny, I actually was going to say, I don't know, I kind of challenge what you said a little bit, because I wonder if. And this is a real lived thing right now for us. And also I think in the thought process around, where can you build in, it's not really risk tolerance. Like, where can you give autonomy to someone who also might be, again, more of entry roles, like, make them feel like they own something. So process for us. Like I was saying, we're so tightly tied to process that we found as we're trying to transform and introduce these things, if we give someone a process, it's actually like they wait just until they have everything they need and then they question it because it's like, well, your process didn't account for this edge case and so I only challenge it.

Speaker C: And like, because they weren't involved, right?

Speaker A: They weren't involved, they didn't get input. It doesn't flow into my. It, uh, doesn't work for tier four, it only works for tier three. And so I think it's a interesting double edged sword of like, I actually wonder, and I've been listening to a few podcasts, but I feel like AI in general disrupts so much that I think the thought process is like, do you just kind of throw like caution to the wind in some instances and sort of jump in? I don't know. And again, I think some of that's a luxury of certain roles that can do it because a lot of times you can't. So to your point of like how do you make the time? But one thing goes back to our practical like the how. So the how is that we did have a strategy. We do for implementation. I should probably reference it, gold star. But we actually developed it like early 2024. But I would say, you know, the interesting thing is it's been tougher than we thought. It's more challenging. It's not like we had as March 2024 and we had this thing, we had the tool, we had these fancy kind of comms and branding, but it didn't end up like now 2025. Am I where I thought we would be with the adoption of it and everything? No, but I think our strategy was, I do still stand by. I think it's a really good strategy. I think we just were transforming in so many other ways at the same time, to be honest. But we really tried to do it, build it in the flow of work, don't take someone out of their work. And so that was our huge thing that we still stand with, is that anything we want to create for our team, our teams tend to draft in Google versus in a CMS just because of the way that they work. I think because of the collaborative nature of what they're doing a lot of times and they have to do a lot of feedback has just worked better. And I think we've just grown up in Google culture for some reason. And so we instead of putting the AI solution asking, hey, go into the AI solution and do it, we're asking our engineering teams to build it into our content management system. We right off the bat built all of our solutions into Google and so we asked the AI partner to Actually just like build it into Google Sheets for something. So like, if you want to go create one piece, you go into this Google Sheet, you give it the context, it spits it out in there. Then if you want to do it in a Google Doc for a different draft, more of like an FAQ article, you do it in a Google Doc. And so it's all built into the flow of work there. And we actually had the models integrated and kind of that was like our prerequisite for building, which that was. The whole goal is that we wanted to meet people where they're at, not take them out of the flow of work, not even notice that you're using it. It's just right there. But it's not the standard version, it's the specialized version that has all of your special things in the backend. So maybe it's an interesting use case of, like I say, I still stand by it, but it's been challenging and we still haven't gotten to maybe that level that I would have thought. And that's where I question the people aspect. Like the people ask and you have a question. I love how you said it. When are the skills coming? I think we waited. Everyone was waiting for the skills to come. And I was like, I think we're waiting for everyone to use it. And I think it was a really good reflection. As a leader moment, I'm kind of like, I give myself like a D in recognizing and maybe like quicker reaction and pivoting because I think we're kind of in a stalemate almost of like, they want more process, more skills. And we were like, here's the thing. So it's a really interesting one to reflect on.

Speaker C: We look at the process the same way too, in terms of integration, driving adoption. The more that you have to pivot in applications and systems, the, uh, harder it is for people. Just like standard landing page, CTA marketing practices. Right?

Speaker B: Yeah.

Speaker C: So I think people forget or sometimes don't realize, and this is. Everyone wants to see what everyone else is doing, like how easy it is to build it into the systems you already have today, whether it's SharePoint or whether it is your Google Sheets or Google Drive, where you can build in those processes and then it's specific to the job. I think that's the big change too. It's not like you're logging into the system and having to think, like, what am I here to do today? You're doing your job in Excel, in Google Sheets and Google Docs. And then it's there because now it's part of the process and hopefully you've done training on that process documentation. But you bring up a good thing that the more that we can involve people in that process creation, the more ownership they feel over it. Which is where I keep thinking, hoping that those pilot leads are peers of that team.

Speaker A: Yeah.

Speaker C: So while not everyone was involved in the piloting, you have somebody who is a peer of theirs who does the same process and therefore can relate and really resonate with the rest of the the team and be an influencer from that sense. Did you see any of that?

Speaker A: Yeah, I think we do. And it's interesting we're seeing them kind of come out now as well as we're all talking about it more. And that's the idea. I actually was talking to some people like in all these roles, like, well, what would you change? Like, what would help adoption? That's what everyone said. Everyone's like, I think it's the show versus tell and it's that component of change management. It's not just a bunch of emails. So again, good reflection of your point on creating these little pods and these groups. I think it's giving them that autonomy to be those peer leaders and to be sharing and then also raising them up, like giving them that platform, platform to share and also like celebrating it. There's such a celebration culture, which is, I think it's super cool thing. It could maybe also make people feel kind of bad. It's like, you know, people are constantly sharing in Slack channels, lots of emojis. My team did this. Like we created a podcast with Notebook LLM and you're like, oh crap, I haven't created a podcast yet. I gotta get on this podcast thing. Which by the way, it's so easy. It's frightening.

Speaker C: Yes.

Speaker A: Yes.

Speaker B: Yeah, It's a one click.

Speaker C: Yeah. I'm so passionate about this aspect of it. Giving people the opportunity to step up and lead and then supporting and giving them visibility in that leadership.

Speaker A: Yeah. Yes.

Speaker C: And it's so important to culture and driving change because so many things you just touched on. You're giving people whatever kudos means to them in your organization. And it can and should mean different things based on the levers you can pull so that they feel good about the extra time. This is so career defining for so many people who are stepping up to lead no matter what job you have. And then you hope that it motivates other people around you so you're starting to get that influence and that extension of the culture factor to be like, oh, I never thought about that, or I should really try that, or actual showing people that it is working. Those culture quick wins, but also keeping that momentum. That's why slack cultures or teams are so important to driving all of that. Like having a team of people who are also responsible for keeping those online communities alive in terms of that sharing.

Speaker A: Really important. Yeah.

Speaker B: Oh, yeah, that's critical. Yeah. The people who step forward and really keep the effort going. So I think we've kind of slipped into the impact section of the show where we kind of wrap up. I don't think you're being too fair with yourself, Hadley, about giving yourself, like a D for, like, not recognizing, because you still find that. And Jessica, I'm sure, can attest to this because she sees tons of people. See how far you've gone just in your thinking, which is a huge component of this. And the cultural shifts that you've done and then actual efforts and things you've actually implemented and the workflows you've automated. Way ahead of so many companies. But it's like, tough, like you said, Hadley. It's like, okay, now we're this good. We gotta get even better. You know, it's like the whole saying, the better you get, the better you better get.

Speaker A: Yeah, totally.

Speaker B: And so you're caught in that. But just give us a couple sort of celebratory pieces of, hey, these are some of the big impacts that we saw.

Speaker A: Yeah, totally.

Speaker B: We shift the culture. We did these pieces. What were some of the big, wow, you step back and Hadley maybe even like, whoa, did we really do that?

Speaker A: Well, that was my one. I think when we were. We were having to share some updates for leaders recently, and I think you're like, so heads down. I think even though we talk about a lot, especially, I'm sure you all do. Like, you know, we're in the forest and the trees. That's one of our cultural values. Like, see the forest and the trees? And I think about it a lot because sometimes some of my team said, well, you don't want to take nap in the weeds. And I was like, I don't. I think I am. I was like, I think I'm napping over here.

Speaker B: What if they're really soft weeds, though?

Speaker A: That's what I said. I like the product weeds. It's really interesting. I've actually. I've had to find myself, like, pulling myself out. I really enjoy getting deep into these things. I think before it was deep in the customer's foresight. I can do tickets myself now. I'm like, wow, I want to get deep into one handwriting content and then understanding what great looks like and then translating that to AI and that's like a fun activity. And I'm like, this is amazing. And then also just seeing the impact I have. But I had to put my forest hat on and become a leader again and go back to. I had to share out an update on um, behalf of the team and this like really cross functional workgroup that we have enabling a lot of the big AI implemented pieces that we're building. And it was super cool. I think we've made so much progress. And that was the 500X. Like we really, a year ago we were kind of grappling with one. What does AI optimized knowledge look like? Like we had a vision. But I think at that time too the world was changing so quickly and as we were building things people were like, but is it that? Is it not so I think we all stuck to a vision and I do think it took us a little while to actually like operationalize. Maybe because of those back and forths. Everyone's in such a turmoil of what does this mean, what do we do? How do you go quicker? Who can go quickest? And then we just did it and we use AI to do all of it, which is amazing. And so that's that amazing thing. I got to share of. Like, remember when we were hand reviewing like these FAQs to try to get them ready and then we also were just putting back like the regular section or rewriting them into sort of a different format. No, no, no. Now we've taken every single piece of content we own and in one month's time through AI, we've taken all these things and we've distilled them from millions of pieces down to one core facts for these specific pieces that are then going to be right format, right structure, have the right, they're creating the taxonomy structure, all the right data to then um, power these AI experiences. And we did it super quick. And that's crazy. And you know, you needed product partnership to enable the readiness to like absorb that stuff. You needed our team to build the models to do the distillation. You needed legal to partner with you and say, how can we build like a legal risk flagger to think about things in a different way too and say some are really low risk. Like that's literally just you telling someone a fact. Don't need to review it, just put it in there. And so I think that was the biggest thing.

Speaker C: This is the two year to one month.

Speaker A: Oh, uh, gosh, I know.

Speaker C: Can you explain it to a five year old what you guys did? Are you allowed to share?

Speaker A: Well yeah, I mean we pretty much just took all of our content stores as they exist today. And to my point, on our vision of machine friendly best in class machine friendly content, we took it all, put it through an AI model and uh, out came our version and our definition of AI friendly content.

Speaker C: How many pieces of content?

Speaker A: It's a lot. It's just say like we're in like the millions of pieces and we came down to like the facts. Yeah. And the facts there are a lot less. There's a lot of duplication in what we do. There's a lot of things you write multiple times for multiple different systems. I think our fun one is we tell people like 1600 different versions of forgot password content.

Speaker C: So millions of pieces of content distilled into best in class your brand customer centric FAQs that are then being sent back into the product, right?

Speaker A: Yeah. And it's being used by the product. Yeah. And super cool. And then it's also the business piece where we take out that concept of kind of the tone of the brand because this is just that pared down fact that's going to use the machine to build back in. And then we work on the guardrails and we say well this is what good sounds like and this is what the guardrail should be of. How do you create the right Persona tone or something for a specific use case or early days on that one. But yeah, that's it. Then it goes back into the model or to the machine.

Speaker C: What about the team that's doing it by humans? Did you figure out what the reason was behind that and were you able to convince them otherwise? Good question.

Speaker A: I think maybe uh, we do have a follow, we have a chat. I think it's such a hard thing to crack. We have some really cool people on our team who just have the experience and I think have done some of these things before and are really trying to push the envelope. That helps so much. I think when you've got people have been like hey, I've been here, I've kind of done this before. But I think yeah, uh, what I gathered is someone. Some industries are more risk averse. We are risk averse. But I think internally in a tech we operate in a lot of spaces that have regulatory. But as long as we can really do the right thing truly and again have the right cross functional partners like legal and those involved, we'd Also take some risk in good ways. And so I think that's one thing. I think some industries just aren't. It's tougher probably to say, hey, I'm going to use AI to create all this new content. Like our product and tech partners are coming to alignment on what AI optimized content source is. Seems simple and it seems like should be it. But I think you probably get stuck. I think they're kind of stuck in the in between of does the knowledge expert own that or does the product expert. And is it like a JSON, uh, file or is it a different structure of my thing? You know, what, what is it? So I think it's that I'm sure there's so many people, we talked to a lot of other people in the same boat. They're really stuck kind of between even support from like leaders to do the transformation. I don't know. That sounds crazy. How's it going to go?

Speaker C: I think you've hit on the theme there that I like to touch on as well, is people. If you guys notice that, it's come up a few times in our conversation today.

Speaker A: Yeah.

Speaker C: And it really comes down to the talent on your team, the willingness and support of your leadership team, but also the people who are leading this. And I, I feel like we say this a lot, but tech doesn't drive transformations. People do. And so it's so important. You can have two teams doing the exact same thing and one team is going to get results and the other team isn't even going to want to start. And it comes back to people.

Speaker B: Yeah, that's huge. That's absolutely huge. All right, so now we've come to a lightning round section of the show, Hadley. This is where we just do a quick hit. Each person on the show shares one thing that either happened recently, news project, experience or something you want to rant about. A rant or a rave. I've got a rant today for sure. How about generative AI or AI in general? And do you have one top of mind, Hadley, that you want to talk about?

Speaker A: I don't know. I think AI is such an interesting thing. I feel like I live and breathe AI, even though I'm not like an AI professional. But one thing I think is trying to understand all sides of it. So the concept, yes, sustainability is everywhere. About a year ago, the discussion wasn't everywhere, which is interesting. So I know it's a hot topic now. It's something we've been kind of talking about a little bit since last year. Just More in forums. So I just had listened to a podcast recently that I thought was really interesting in terms of how can you see the good in the, uh, what can AI do to actually improve and think about the future of sustainability? And is it kind of like the medical side too? It's like, can AI help us to completely break the mold of how we're doing it today and actually get us to a more sustainable state, even though at the same time there's so many downsides to how it's having to be powered? So it's more of just that thought process of I think I'm truly grappling with. I feel like I'm so all in on enabling these things and I use it day in, day out and I'm building strategies too, but there's so many sides of it. So I'm trying to kind of develop and like look more, uh, in terms of the opinions that I have on that. So that's kind of my lightning route is I'm learning, which is a good thing. And I'm trying to understand more deeply and be smart and uh, what is it, you know, kind of like true responsible use.

Speaker C: Yeah, it's definitely. I can tell you a question coming up from a vendor standpoint, way more than it was a year ago, to your point. Yeah, and I love how Paul Raitzer recently put it, I can't remember which, if it was the state of AI or a podcast or what it was. I don't know about anyone else, but I listen to a lot of Paul Reitser from, um, the Marketing Institute. But it's the risk reward conversation. And it's the same thing from an IP standpoint. It's is the reward of this optimization, hitting our growth goals. What this is going to do, both for me personally and professionally for our business, is does that outweigh the risk right now? Even if you think about it, you can think about different things. And I love your point of using AI to think about this as a thought partner even too, but I don't drive to work anymore. There are a lot of different ways to justify how your life is being sustainable. And maybe you can't change anything with AI now, but you can make small changes that, uh, add up over time in other ways.

Speaker B: No, I think that's good. I think it's something that's good, that's top of mind as well, because that's long term. It's really, really important. So I think it's an important topic. Mine of course, is not nearly as existential. It's more of a rant. So, uh, vibe coding is the new thing. I think it's Andre Karpathi. I think he was the first one who kind of made it famous. Like, hey, look at English is going to be the most prolific programming language or the most popular programming language. So I started using Cursor like I was talking about made that game. And that's part of like you were talking about Hadley getting smarter, really pushing yourself to. I don't have time to code. Like I can vibe code. Absolutely. Let me do this and let me find use cases. The thing is though, there's some of the same sort of hallucination and issues though, when you have, you know, somebody writing code for you as well, where you write the code and you try and run the code. It says, well, I can't find this. You put it somewhere. So go find out where the thing was that you're supposed to run. It's like, I can't find this dependency. I'm like, well, you're the one who created the dependency. Why don't you go figure out where you lost it? And you're finding sort of the stumbling blocks. I thought it was going to take me like one prompt. I'm like, I'm going to do a one prompt, create a video game. But it took about 15 minutes, like a back and forth and like errors and errors and errors and okay, that looks terrible. And like, let's fix this and let's fix that. But at the same time too, a good part of this rant which is got me out of my comfort zone of like, no code can be a part of my job. Yes. And this is a fun way. Creating an eight bit video game, you know, where someone jumps around like Super Mario Brothers. That's fun. But then it shows you. Like you guys are talking about the quick win of, okay, I can do this. But at the same time too, you have to put energy into making it work. Right. Cause it's doesn't work right. Completely out of the box. But you have to change your behavior.

Speaker C: I need another podcast episode on how you're using coding for work, for marketing, because I need to hear more about that. But I think you're touching on another really important human trait, and that is tenacity. Because I think that the people who are succeeding right now with AI are able to continue to push through that barrier of it didn't work how I thought it would that first time. And the people who are, it sucked. It didn't work and not coming back. Everyone has had that experience, but the ones that you really care about getting right, and you keep going and pushing through that. Now you're seeing 10x whatever it is of the process. And so I, uh, think that the high risk tolerance, the tenacious trait among the other things of people and empathy in critical thinking that we talk about are just really important to people who are succeeding right now.

Speaker B: Is that your lightning round, Jessica? Tenacity.

Speaker C: You know what? For time, it is. Be more tenacious. Love it.

Speaker A: Love it.

Speaker B: Fabulous.

Speaker A: There you go. Yeah.

Speaker B: Instead of stay curious and stay tenacious, that'll be Jessica's there.

Speaker A: I'm gonna seal that one.

Speaker B: Fabulous. All right, well, this has been absolutely wonderful. Hadley, thank you so much for joining the AI Edge podcast.

Speaker A: Thank you.

Speaker C: Thank you, Hadley.