Mural CPO on How to Put AI in Multi-Player Mode for Your Team; Why AI Made Work Lonelier, Not Better | Elaina O'Mahoney

0:00What's interesting is visual collaboration is actually absent in the AI era. Ideas are really cheap and you can go from an idea to an actual working prototype in hours or you can write a doc or you can create a slide deck in minutes. That's actually really compelling.

0:16Carlos: Elaina O 'Mahoney, CPO at Mural. I'm curious to know, what is the real competitive landscape for you? The competitive landscape is really interesting because when you're a product leader, a lot of your just skill set of being very good is about future casting. Kind of like in the past when like some say people would report fake calls on their CRM to show productivity. Now they're like creating whatever it is on AI to say, hey, I burned X amount of tokens, right? So how do you ensure that in a way, like some of this usage is actually driving the right business outcomes? Just because you can with AI, and I'm probably going to say the controversial thing doesn't mean you need to do with AI.

1:24Carlos: inside big enterprises. While your best forward deployed engineer might sit in legal, ask what did you learn, not what did you ship. Let's get into it. Welcome to the product podcast, Elena. I'm so excited to be here. Thanks, Carlos. So you are the chief product officer at Mural, which is a product that I've used for a very long time. And I know that when it started, the whole pitch was, hey, this is a better version of a PowerPoint, basically. But now this whole industry of visual collaboration has evolved quite a bit. love to get your take on. Absolutely. So it was meant really to collaborate about ideas to brainstorm.

2:00Our co-creators from Argentina were looking at how they build a game and how they come up with ideas really fast when people are across the globe or want to collaborate async. And that's where the idea of Mural was born. And I think what's really interesting is when you work for an organization or when you're a team and you're trying to come up with a multitude of ideas, you're trying to brainstorm, converge, diverge, and then ultimately decide on a path forward, that was Murl's bread and butter. It was why we took off so fast during the pandemic when everybody's like, we have to find a totally different way of working and Murl's it.

2:42Carlos: No, you reminded me of Zoom's initial value prop. It was, it just works. That seemed to be enough because the authorities were just so bad, right? And now it's obviously expectations way, way higher. That's not enough. And so curious to know what real visual collaboration looks like today in the AI era. Yeah, what's interesting is visual collaboration is actually absent in the AI era. We are all in our tools, all by ourself, with words or voice. I don't know if you dictate to yourself often or speak to your LLM of choice often. I probably do that more than type these days. But it's very lonely and it's not visual.

3:23And when I go to collaborate with my coworkers, I am now sharing words in a doc rather than understanding through visual collaboration and design something more easily digestible to get to decision making faster.

3:37Carlos: And you're bringing up the point on multiplayer. And I know we've been talking about this for a long time. I think at the beginning, it was more of an idea. Everybody who's used AI in one shape or form, they realize that at some point you get stuck and you are not able to fully collaborate on the same context. but with a visual collaboration tool on paper, you should be able to truly collaborate better and get stuff done in a way that you can't when you are just using a little limb of choice on the site. But for some reason, I don't see enough of those use cases yet. You know, like I see a lot of people still using visual collaboration tools or PowerPoints or Google Docs kind of in the same way, just putting comments on the site.

4:15Carlos: So I'm curious to know what like really like visual collaboration would look like for a product team that is trying to build together. Yeah, it's really interesting because now ideas are really cheap and you can go from an idea to an actual working prototype in hours or you can write a doc or you can create a slide deck in minutes. That's actually really compelling. But when you go to share any sort of artifact, it goes into both back to these silo moments. So for us, Mural, and when we're thinking about our product process, we have to almost forget the process part and we have to create a system that really enables faster decision making now that ideas are cheap.

5:01So how do you take all of these different artifacts, spatially connect them together so people have the same context? Because it's really that's what the game is about now is am I playing the same game as you are? And do I actually know where you started? And can I catch up to you as fast as you are so that we can actually just make decisions faster? Because you have so many more decisions to make.

5:27Carlos: Yes. And I see that the same way as in the past, a lot of these SaaS companies were trying to replace the spreadsheet and they would position themselves in the middle. And then they would say, and you can integrate with everybody else. Now, I'm curious to know, what is the real competitive landscape for you? And how are you thinking about integrations with other players? The competitive landscape is really interesting because when you're a product leader, a lot of your just skill set of being very good is about future casting. You would have to create these strategies that are like two to five years forward.

6:00But it's almost a in the moment feedback loop that you now need to be able to create in a system level thinking. So for us, it's really just about having the right information in the right moment in time. So, for example, if I start with a document and then I want to actually spatially present that idea to multiple people, can I quickly create that document into a prototype, into slides? But can I connect them spatially together so everybody has the same context to make the right decision moving forward?

6:38Carlos: I see now companies like Slack or Teams, right? Trying to be that space for people to collaborate. And you also have the issue tracking tools like Jira or Linear also trying to communicate, hey, this is the place where work gets done. But the reality is that it's still unclear where work gets done, where agents are being deployed. Like as companies are trying to truly deploy this in a way that drives adoption and it's not just a few people that are creating their own context. Where does work happen and how does the right flow of information has to go? Yeah, it's an interesting concept where every tool right now in the stack is almost converging in a circular area.

7:20And everybody is also converging on like a chat widget UI where work happens. How many different companies have kind of the same tagline of where work happens? And yes, OKRs can be in a certain area and it'll say what work is planned. And then you can measure against what work has been planned. Slack is very good for communication in the moment where everybody is talking all within this one area. And then you'll have notions, confluence, doc, or JIRA or linear where you're still planning work or decisions that have been made. And the space that really is still ambiguous is how are decisions getting made and the context of why that decision is made.

8:07And that really is a blank canvas area that we think we're converging on. that, yes, meetings, we talk about stuff, but the meeting notes are after. It captures a summary. You still don't have the context of why the decision was being made. So how do you bring some of this desperate information together and really get to the why behind decision-making in the moment?

8:31Carlos: You're making me think about how some of these products that try to become the center of where work happens started implementing AI. And it was usually like a button that said AI or this star. you click and somehow summarizes information or does something magical, but now it feels like AI is not just a battle, right? And I mean, it's kind of happens across the board and maybe some of the things don't need to be called AI at all, right? So as you think about the user that hasn't set up the context, right? The one that shows up and is suddenly in a canvas where they can start playing around with different pieces of information, like how for a product team specifically, How do you see that happening today when they are maybe not super technical?

9:13Carlos: Maybe they're not a designer, but they still need to have that ability to ship stuff, you know, to not just depend on the next person in line to slow them down. Yeah, we have this arm of our business we call Luma. It's one of our services offering that we have. Now being part of this organization, really going through Luma practitioner training, I think something we try to break away from is people thinking they're not a designer. If you've put together a slide deck, if you've done a drawing and everybody drew when they were a kid, you're a designer at heart. So everybody actually needs to embrace the fact that you are a designer.

9:52You can be a designer. You may not do that by trade and you may not do it to the proficiency of those in the design community, but everybody can design. And there are tools out there that can help you design. But I think the point is that visual pictures, whether it's a customer journey map, whether it's just a diagram, whether it is a even a data analytics diagram, you can actually absorb information faster than you can if I were to read a document. And it's that visual element of being able to contextualize information that much quicker that really speeds up your ideation process, that speeds up how you actually make decisions, what you decide to take to market, what net new information you may not have to make a high quality decision.

10:48But doing that in a visual way, and now that I work at Mural and I do that more frequently than I did previously, I can see how we can actually get to maybe the disagreements faster. How do we get to where the challenges and the bottlenecks are so much faster when it's visual?

11:08Carlos: And I think that what you just said about you also leading by example sets the tone. Like I've hosted product leaders from Anthropic, Marcel and other AI native companies, like really AI native companies. And they're all shipping from top to bottom, right to left. Everybody's a builder in a way. So I'm curious to know for you personally, as the chief product officer, how do you actually set that tone and shape? Yeah, absolutely. I do think everybody should be playing and shipping. Now, should everybody be shipping in production all the time? Maybe is the lens where you can ship in safe environments.

11:45You can actually solve some of your core bottlenecks as the CPO where you inherently would make decisions and choices. you can actually democratize some of that by just shipping workflows, shipping agents, shipping research agents. There are still some things that are quite scarce that you can work on internally. Data science is still scarce. Really good taste is still scarce. Synthesizing analytics is still scarce. So you can actually decide as a CPO where you play in that space to actually make your team 10x faster.

12:25Carlos: I just had this recent example. Today, I shipped the new email signature for the company and I didn't really ship into production. It was like that 80 percent. But I felt unstoppable, you know, because I could express myself in a much more visual way enough. So now we can skip a couple of steps. And ultimately, there is an expert that is still going to validate and another expert who is going to put it into production. But I think that, first of all, it feels great. It feels great. I think it's not like a major, major release. But I like your point around shipping doesn't need to be something major into production.

13:01Carlos: Shipping a workflow is still like a great way to demo that you can do it. Yeah, it's also something that we went through here at Mural, which hopefully is relevant to other product leaders and just, you know, engineering design and product teams, where usually a product leader would write this very, very big product strategy doc. And sometimes it's five pages, sometimes it's four pages, it makes the rounds, people comment, and maybe it starts to slowly die. What we can actually do faster as leaders is actually create a prototype of something that we're envisioning. And I'm not saying it will ship like this prototype, but it gives a conveyance of some of the principles that you actually want the teams to explore in a North Star direction.

13:47And you as the product leader can do that so much faster. Or you as the PM, or even the engineer could be like, hey, what about this concept and idea? And it actually moves the entire team faster in their ideation cycle or to their decision-making cycle just with a visual.

14:03Carlos: Your product, I mean, it has both angles, right? You have the consumer or the employees and smaller businesses as well as large enterprises. So I can imagine that in small-scale business, you have no choice, that you have to do a lot of different things. But when you start talking about very, very large companies, over a thousand employees and whatnot, maybe they use their product in a non-AI way for some brainstormings and stuff like that. Like, how do you go about highlighting the new horizon and like making sure that people who are not super technical yet feel comfortable and get that feeling of magic?

14:42Yeah, I think it's actually lowering the bar for somebody to learn in a lower risk environment. What AI has created are like almost the halves where you almost have to virtue signal that you're always ahead of the game or you're on the frontier. And those that you're like, I think I can write my emails and do the summaries. And I can even write a Slack bot that summarizes the things that I need to do in the morning. But how do you give them the space to then showcase things that they're doing and saying, am I on the right track? Or somebody that's doing something very frontier and proficient. are you creating the space for them to just try?

15:22And because AI is so solo player mode, it's really hard to see people work out in public to get the ideas of how to move forward. And so when you're working for Fortune 500 companies, how do you create a safe space that gets more people to try things but fail at first at trying it? It takes you, I mean, Carlos, I'm sure you tried to write many different things over and over again. And you had to like learn by failing with a lot of these tools. And how do you create that same environment for employees?

16:00Carlos: Totally. And our clients are also, a lot of them are large enterprises and we actually try to bring right tools for them. And one approach that I've seen work is when we pick smaller teams, even though the company is big, like first to get those first champions because I haven't been successful at doing these major rollouts? There's some things in this moment in time where you almost have to rely on some things that were true in previous transformations, where there are large organizations that have innovation centers for a reason. Because if everybody had to do something very dramatic all at once, it's really hard to roll that out at scale.

16:40So how do you get teams to try something differently? But you have to almost break a process. You can't have a process in the very beginning. You have to set up a system for them because the process is going to change as they learn. And I think that's the hardest struggle for very large enterprise organizations is they're keeping their process, but then they're trying to also adapt to a new way of learning by keeping a process. So you have to set up a system and system thinking.

17:09Carlos: And for that type of system, do you require some type of forward deployed engineers at some point? Or is it something that you think can be done a little more organically? I think forward deploy engineers are really, really helpful because there are usually the ones that can use not only the tools that are most best for the job today, but then it also gives people the insight as long as they're sharing to say, oh, this is how that person did it. This is how they automated that workflow. This is how they did the discovery. This is how they used Claude or Codex or name your tool of choice. I now can do a bit of what they did.

17:53So whether it's a forward deployed engineer or whether it's somebody in legal that's automated their process, it doesn't have to necessarily be an engineer. It can be somebody in any department that has already adapted to the tools, that has already automated some of the way they work and embedding them into a different team or a different environment that hasn't yet.

18:16Carlos: Imagine an attorney playing forward deployed engineer. That's really cool. And how do you go about identifying those type of early champions to give them the stage so they can also highlight what they're doing? Yeah, I think it's about being able to create a forum for people to showcase their work. And I hate to keep saying Mural, it's a collaborative tool, you can see what people are doing. And I use the legal example because we have a person that sits in our legal department that we have been trying to steal to the product and engineering side for a while. And our head of legal is like, don't touch him.

18:50He is so good here. He's automated 50 % of our processes already. And when I think back, that's actually probably better that he's there automating a different part of our business than me stealing him for us. But he's able to showcase his work in Mural. He shares it with the entire organization and it gives people like these light bulb moments in every department.

19:14Carlos: One of the things that I think is unique about your product is the community aspect. And some companies have decided to open source the product or tap into communities. easy to say, so hard to crack, because this is not just a discussion forum where people would propose ideas, right? And you've been doing this for a long time. So I'm also curious to know how you are tapping into that open community of people using Mural to then bring some of those best practices to the way you build a product or the way other companies should be using your product. Yeah, I think we have actually introduced a newer community within Mural, and we've got a subset of our customers that are first invited to this community.

19:53And I think that almost ties in back to the previous point that I was making. It's about sharing how you're working, how you're deciding, how you're thinking. It's the same reason, Carlos, you exist. You've created this community where people can share ideas about how they work, and it sparks another idea in another company from another person. And I think the thing I think was most surprised about when I joined Mural a little more than a year and a half ago is how many people are so passionate about human-centered design, visual collaboration, ensuring people can all feel like designers, and ideas can come from everywhere.

20:32Carlos: I mean, I asked you that question because that's very an important piece of my business, how I think about it. And, you know, like when I started thinking and building community for our company, I didn't have any expectation in return. I knew I had to find a way to measure ROI. And at some point I couldn't. And I was like, you know what? I don't know how to measure ROI, but we're still going to do it. And that I think sparked something. Because sure, it would be great to say, and we reduce the amount of customer support or we increase revenue. And I'm sure there's some derivatives to that. But like the actual giving value is critical because there's so many other options out there, right?

21:10Carlos: Like how, why would someone dedicate time to build and share what they're doing with your product? Yeah, I think it can also serve multiple purposes. A community that inspired me, I think, earlier on was Linear's community. If you are just trying out the tool, they'll invite you to your Slack community. And it's there where everything is about adapting to change, doing something brand new. And when you're doing something brand new and there isn't a playbook to follow, that's when community is most important. It's creating this trusting and open environment where people can say, I've tested this out.

21:49I've done it this way. Or you can ask a more vulnerable question. And then people are there to almost say, I know an answer. I've tried it this way. Or I can connect you with somebody else. Do you want me to show you? It's the value of that loop that's actually quicker because you're getting it from people just not inside your echo chamber.

22:11Carlos: Yep. And I want to also connect this to how you're thinking about pricing, right? Because at some point, if you can really drive adoption within your business and you build the right integrations, you're unstoppable, right? Like people need that product to survive. That's the heart of how they build. So it gets to the point now with like token usage and whatnot. not. So how are you thinking about pricing? This is the thing that actually keeps me up at night. And I wonder how many organizations that are, to use the very, very vibe name of AI Native, we weren't born when token and LLMs were there.

22:51So our pricing model reflects a lot of the seat based pricing that everybody now attaches to legacy SaaS. But there's almost a double-edged soared with usage. So when we speak to some of our very larger customers, there's almost a black market for tokens. If they use their own tokens, they'll like have to go to three other employees to say, I've got 80 % of the way there with my prototype and like my tokens got cut off. Susie, can I siphon some of your tokens? And now it's like it disincentivized the wrong behavior for this. So it is a decision that we are taking with the utmost care. And it actually goes back to we serve totally different types of businesses from a person that's just bought us for themselves for maybe their three person company or their even solo consulting gig all the way up to a fortune 500 10 ,000 company.

23:47We have to think of pricing models throughout that same thing. And And maybe the pricing models are completely different. But we know the value of our product in this very near future has to be felt with AI at its core. But we also don't want to disincentivize folks from using it and have to keep an accounting ledger of tokens.

24:07Carlos: We are at the early innings of this. I mean, obviously, you see the examples of companies that are burning tokens and they are very proud of it. I also know companies that send the right mandate. Like, you have to burn X amount of tokens. I know I'm not going to mention those companies, but they gain their own system. Kind of like in the past when like some say people would report fake calls on their CRM to show productivity. Now they're like creating whatever it is on AI to say, hey, I burned X amount of tokens, right? So how do you ensure that in a way like some of this usage is actually driving the right business outcomes?

24:40I think the first thing does come with measurement, which was actually the first inclination of most organizations of here's a tool, just use it. And usage showed adoption. And then everybody's like, holy crap, this is very expensive. Everybody, no, come pull back, pull back, pull back. So it's really about starting with a measurement layer and then measuring how they're using it. Just because you can with AI, and I'm probably going to say the controversial thing, doesn't mean you need to do with AI. Maybe spellcheck with AI is quite expensive. There's other things that can do spellcheck that don't have AI.

25:20So it's can we apply it to the right things? Can we at least measure our costs? And everything is going to come back to learning loops at the moment that give us the flexibility of the decision making. and not going too far down the future casting of decision making because it's really hard to pull back pricing models. Once pricing models are in the air, people are really going to fight tooth and nail to not change it. And the last thing that we're trying to make sure is predictability. Any business has to understand that cost for their business when things are so uncertain. So our cost of our tool can't wildly swing from month to month at all.

26:05There has to be predictability to it so that they can continue investing, that it can incentivize the right usage behavior, and people can get value out of it freely.

26:14Carlos: Can you give me a sense for the current team that you run? Yeah, the current team that I run, we really focus on both broad capabilities and very specific personas because we are a very broad based tool. We do have some traditional triad teams where we have designers, engineers, product folks, and then we have some product engineers, ones that are playing dual role. We also have some designer product folks that are playing hybrid roles. So it completely depends on the area of the business that you work. So if you're within our core product that serves all of our users, we're going to have more of a traditional triad because you can't just frontier test and try with, say, banking and finance organizations that have to go through a lot of administrative and security.

27:10But if we're in a newer area of our business that we're trying things out, we can go to hugely untraditional models where you have different hybrid teams where you can ship to production very, very fast. And change management isn't really part of that issue. It's high experimentation. So we really run that gamut of the teams working completely differently depending on the outcome they're trying to drive.

27:36Carlos: Ultimately, you oversee the different teams, right? Zero to one, the more core product. Absolutely. I take it from that model of maybe you're a city planner, so you have to break down freeways. You have a working city, so the way that you change it has to be more thought through if you're going to break down a freeway and put a new one or you're going to make a toll bridge versus you are a frontier team. You're just back in the day before America was America, you're sending out West and then you just have to backpack. You have to kill your own food. You have to forage. And the way that those two teams actually work, completely oppositely different, but they're serving the same business.

28:19Totally.

28:20Carlos: And I was having a similar conversation with the CPO at ServiceNow and public company, over$100 billion market cap, but they are down like at least 25 % year to date. So when you have that type of market pressure, I was trying to understand, okay, how much effort do you put in the future, knowing that it might not pay off next quarter. So in your case, as a private company, I'm also curious to know, how do you allocate those resources, like new beds or things that are less mature versus things that need to kind of keep the lights on? Yeah, it's the innovator's dilemma, I think, is also always the phrase where you have a business that you need to serve today that's really reinforcing your revenue stream.

29:02But you know that you want to actually evolve your product and service for possibly a totally different audience. So you have to really think of the allocation in bets, as you said. And maybe we do a certain percentage for that bets that we know could be self-fulfilling from not forgetting our core business and why people buy us today. So it's really about spreading that allocation of bets for that team and letting the team within that bet area help us understand what's working and what's not and do we need to reallocate those dollars. And that decision making is much, much, much faster. But again, I'm paying attention to a lot of that future bet of what's working, how do we make something that's maybe completely different than what we do today, but stays core to collaboration, bringing people together, and visual.

30:00And it's really that allocation, but you have to look at revenue.

30:04Carlos: Exactly. So traditionally, there would be big QBRs or maybe annual planning. And so I'm curious to know, how often do you reprioritize now? Oh, that's probably once a month. In the future bet casting, that feedback loop is once a month. We, our CTO and myself, are meeting with the teams to say, what did we learn? We're not just saying, what did you ship is, what did you learn? Because you need to actually change your decision making faster. If we're not changing our decision making or we're still thinking the same way we did a month ago, we didn't gather the right information. It has to be something net new.

30:45And so meeting with our future bet teams once a month helps reinforce that feedback loop that helps us make, oh, maybe we need more resources in this future bet because they're on to something. It's almost running like three mini startups and a business that's been around for over 10 years.

31:05Carlos: Over 10 years now. Wow. I remember when I met your founder early, early days, he was telling me a story about the video game that turned now into Mural and now look at where you are today. Yeah, he's still there. Mariano's still around and he still is definitely an idea person. Yeah. I haven't talked to him recently since I'm from Spain and we just won the workup, but I hope he still likes me. Oh, I know. I was also there during that time. There's a lot of Argentinians in Spain. There was a lot of shouting for them. Elena, it's been awesome to have you on the pod. Thank you so much for your time.

31:38It's been wonderful.