Ways of Working: Building World Class Product Teams that Deliver Innovation and Impact in the Age of AI

Jeff Chow at UXDX EMEA. Video: https://youtu.be/9aQ2KIoAKPc

Readable transcript: edited from the recording's captions for readability (fillers and false starts removed, punctuation and section headings added). Wording is the speaker's own. Timestamps are positions in the video. Names marked [?] could not be verified against the audio.

The waves that shaped great teams

[00:00:07] Jeff: Thank you. OK, hey everyone. I am Jeff, and I am the last thing in the way between you and cocktails. So look, I am going to be talking to you about a fun subject that I've talked to a lot of you about outside: ways of working in the age of AI. Now, did I use AI to make these slides? Absolutely not. I took out the matchsticks, the hot glue. I constructed these slides like our forefathers and foremothers wanted us to, with GIFs. And I am going to do this because that's the only way I know how to do it.

[00:00:50] I've worn many hats, so I know a thing or two about ways of working. I've led design teams, product teams, tech teams, all the teams. I've been a founder. I've worked at all sizes of organizations. I've worked at disruptors and, yes, the disrupted as well, so I've gone through that journey. And if you think I'm cocky enough to say that I'm Yoda, I'm bald, that's why I'm related. I'm also this guy right here. This is all about me going through this journey. I've homered into the bushes more often than I can say, because I've made my mistakes. I've suffered the pain. I see some of my teammates here; they could acknowledge that too.

[00:01:33] But along the way, the number one thing I've noticed is that it's all about building great teams. The magic doesn't happen from the smartest-man-in-the-room syndrome. It's not about founder mode. It's not about anything else. It's about cross-functional teams working together to make the magic happen, making great music. And what makes teams great has grown in, yes, waves. I swear to God, I made this presentation before everyone started talking about waves.

[00:02:04] Of course, the first wave was digital and mobile, and that's where we got agile processes and design sprints. How do we actually work together faster, moving on from those waterfall processes? How do we figure that out? Many people here spearheaded that and were at the forefront of establishing these new patterns for us. Then there was what I would call the pandemic. We were actually getting better at working well together, we were figuring this all out, and then it's like, wait, are we that strong? Is just being in the office a crutch? How do we work remote? What is this async and sync collaboration? How do we work even better? How do we have those emotional connections and relationships even when we're distributed? We're figuring this out, and these are all journeys. We're not there yet.

[00:02:53] And I think you know what's next. This next wave, and it's a big one, is AI. This is a significant wave. I will say, throughout my experience, going back to wiring servers into racks before the cloud, this is a big one. So how are we all feeling about this? Well, I can sense from everyone doing the cheer with AI, a little fatigued. But here's what I've seen. This is one person, a cohort of the group, and these are people here. I'm seeing a little bit more of this, to be honest with you. If we have the bingo card out here, if I say this, yell bingo if you finally finish the card: MCP. Bingo. Oh, there you go, they got it right.

[00:03:38] And then there's this guy: "I'm going to be a superhero." This is really great. This is the person that's hacking AI right now. It's fantastic. And they are also the same person posting to LinkedIn: "I used AI and I created this coloring book and I'm making $5,000 on it. I have a team of 50,000 agents. It's amazing." These are all great people, and I'm here to say that they're all right. You can be freaked the [ __ ] out, you can be really blasé about it, and you can be excited. This is exactly what happens in every disruptive moment that we've seen: mobile native, etc. So it's exciting. I'm not here to debate any of that.

An opportunity to fix our ways of working

[00:04:17] I'm actually here to talk about our ways of working. AI is an opportunity for all of us to uplevel how we work. This is a huge opportunity, and it's an opportunity because when there's disruption, there's appetite for change, and that's what's exciting about it. So what if we used this as the moment to solve the biggest challenges in our ways of working, to address the tiny isms we love to hate, and to open up the best of ourselves and our teams? What if we shifted the lens to think about that? Because I'll tell you what, our existing ways of working aren't all that.

[00:04:54] And it starts with this: our current ways of working center around the tools themselves. Every process that you work through today is there because we are using tools, and often different functions use different tools, and then the ops team creates processes to connect the dots between those different tools. So you've got your Figmas, your Asanas, your Jiras, all of them, and all these things are connecting the dots in that experience. And this is what it really feels like. Every time you get a little momentum: record scratch, got to go enter that into Jira. Every time you need a little help: record scratch. "Hey, I need a little help solving this." "File a Jira ticket." Every time you're trying to figure out some process, there's a little record scratch moment. And this is the challenge.

[00:05:43] Now, the real opportunity is: what if we had people at the center of all this? What if we were able to create a moment where teams could focus on the work itself more often than not? They're not thinking about the phases of their work. They're not thinking about the processes. And the tools are actually helping them go through the task itself. If that happens, it's not about the phase of the process. It's not about acronyms like PDLC or SDLC. It's actually helping teams figure out: I'm ideating, I have to synthesize that, I have to execute, I'm getting new user signal all the time, I'm getting business signal all the time. And as a team we're working the problem, figuring out what's better, the stronger elements in that.

What great teams do

[00:06:29] Today we have great teams who do this really well. We can see them through and through. And I did user test this outside: designers are really happy to be Adele [?], so I just want to make sure that that's clear. And UXR was actually really happy to be man [?]. So everyone's cool. OK, everyone's cool.

[00:06:47] But the thing is, these are just great teams, and the great teams have three key attributes that I've seen over and over again. They are constantly creative problem solving together. There's not one person saying, "Hey, you're the creative problem solver and I'll be the executor." They're doing it together. They are pros at co-creation. They will yes-and the [ __ ] out of each other, and they are helping each other and building momentum as they try these ideas. And they're rapid decision makers. They're not freezing each other by saying, "Did you validate that? Have you validated that people turn oxygen into carbon dioxide? Have you validated that?" They're actually pushing this forward and going nimble.

[00:07:27] Here's the key, though: it's the exception, not the rule. I think everyone has seen this team, and I bet most of you have been on this team. And I bet you're all looking into the distance longingly: "I remember that, that was fun." And here's the key. It's not there because we've actually built our ways of working not optimizing for the best teams in the world. We've built them maximizing the middle ground, and that is a challenge. And it's actually proven: people are saying that ways of working are not helping them make impact.

Omnipresent knowledge

[00:08:02] OK, this is where AI comes in, the boogie AI. The real value here and the real opportunity: I don't think it's deniable that AI is going to help uplevel impact. It's going to drive 10x, 100x drivers. It's going to create real impact. If we surround our people with the right tools, collaboratories, agents, and go through that experience, there's real value. And where there is value, there's opportunity to build a better place for us to work. That's the key. So let's skate to where the puck is going. What happens when you have omnipresent knowledge, democratized craft, faster automation? Well, let's go through it.

[00:08:53] So first, omnipresent knowledge. You have universal assistants, the copilots of the world, knowledge retrieval systems like Glean, agents who are automating the insights. Well, if you actually have that, you have information at your fingertips. You can make faster, informed hypotheses. Which user researcher here has not had that really sad moment where you see a product kickoff and somebody's like, "Yeah, we should probably run a study on that," and you're like, "Listen, dude, we've done that already. I have it." So this is actually getting the work that people have already done to where it's most important.

[00:09:36] That leads to less perfection in ideation, that scary moment where it's, "I don't think we can move forward until we have everything validated." Well, actually, the data is there, and you're able to go even faster. And what's truly powerful here from a cross-functional perspective is that we can start training people less about the scariness of how do I get the data, and more about how do I interpret the data. Quant everything. It's now our ability to think about how we go through these experiences in a better way and how we equip ourselves. Every discussion I heard today was less about the pixel pushing; it was about how do I leverage data to make better insights, better development, better steps in the process. This is at our fingertips.

Democratized craft

[00:10:17] Democratized craft. A little scarier. PMs will start concepting, designers will start coding, engineers will do a little bit of both. I'm here to say that this is really exciting and we have to get there. The yes-and culture: you are all very good at co-creating, but I bet every single one of you, when a product manager came to you with wireframes, rolled your eyes at a certain point. So here is the challenge: we have to yes-and each other. We have to create an environment where somebody shares an idea and you're like, "Oh my god, that's amazing, let's keep going." But now, what if the robots create that first version? The not-invented-here syndrome is thick. We have to be able to say, "Awesome, let me entertain that and let me build off of that." If we start building that muscle, we can jam better together, we can figure this out, and we can actually build more cohesion in early ideation cycles.

[00:11:13] The concept cycles will accelerate. This is exciting. If the concept cycles accelerate, your two user studies that were kind of opaque, where you don't know if you're sure but time is pressured, can turn into ten, with confidence: insight signal. And then there is this other thing: visual communication and execution are different things. There's never been a moment where I haven't seen a designer or a product manager or even an enterprising engineer share an idea visually instead of spending 20 hours talking about it. If it's visual, people get it. They get it right away. They start talking about it in a different way. That is the opportunity that we have, and it brings everyone who has a seat at the table more up front. That discovery phase becomes more energizing. We're solving problems, we're going through that experience. So this is where democratized craft has huge potential. Don't think about it as, "Wait, somebody's going to take my job." They're not. You're strengthening how we're going to be building the best solutions possible.

Faster automation

[00:12:16] And then lastly, faster automation: agents, automated workflows. I believe Anthony called it the cycle collapse [?]. I'm calling it a workflow stack something. We'll figure it out; we'll unify a term. But the key here is this means you're going to be constantly creatively problem solving. Think about a hundred-step process that's at L5. What is the value of humans and cross-functional teams here? Working the problem, figuring it out, taking every signal that you have and saying, can we do better? Can we do better? Can we do better? I've got news for you: that is actually what is exciting. That is where I see the most energized teams, when they're actually working together, working a problem. It's not when we go to our separate corners and figure something out. It's about those energetic moments.

[00:13:00] So that means nimble ideation becomes really critical. There are actually band-aids that we use here. We do design sprints because we don't really have time for that otherwise, so we're going to have this big moment, because now it's a design sprint. We're not innovative enough, so we have a big moment called a hackathon. But what if these were actually core muscles that we build? A workshop is a lightweight thing where it's just, "Yeah, that's obviously what we would do right now. We're working this, we're getting the right stakeholders, we're trying to figure out these ideas." What if a hackathon is just a very nimble proof of concept where somebody's trying to solve a problem in very extreme divergent ways? We can do this. This is exciting. And I haven't seen a single person come out of a design sprint, a workshop or a hackathon who wasn't like, "That was awesome, we should do more of that." We can now. This is the opportunity.

[00:13:50] So this is what we've all wanted: more time working, less time process pushing, deeper collaboration with people we enjoy working with, delivering outsized impact for your business. This is the opportunity.

Consistent rhythms at Miro

[00:14:06] OK, so a couple of things. We have not got this solved at Miro, but here are some things we're doing at Miro. The first thing that we're doing to actualize this is creating consistent rhythms. This is true; we're in it right now, talking ways of working. Every cross-functional product team, no excuses, is responsible for core collaborative skills, and those collaborative skills are problem solving, pace setting and empathetic communication. And this leads to the why, the what and the how of every process. Traditionally you might say, "Well, the product manager is probably in charge of that." No, we actually make the case that everyone should drive this.

[00:14:48] What this does is drive leaders to be able to communicate with their teams in a more effective way. They're able to communicate across teams, to other dependent teams. They're able to communicate up the chain, because communication is different depending on whichever audience you're talking to. Problem solving and pace setting: if you don't actually communicate things and you just set pace, you're just a micromanager. So all these things are really important, and the balance is really important, and even creating a nomenclature so that we can talk about it is important.

[00:15:19] We have a consistent way to navigate those discovery, definition and delivery processes: a simple terminology with a simple framework. These are not stage gates, but we have terms. So we say, "Oh, we should do a kickoff, a solutions review, a pre-release review, an impact review," and we always set a next date, so we set pace using that. If somebody says, "Yeah, the solutions review is in six months," and we're like, "Why?" it's, "Well, we want to go to the zoo to observe the marsupials to see if they want to do something." "Oh, that's interesting, I thought it was a notification bell." You can have these kinds of shorthand conversations to talk through what this decision is, and you can get alignment. It's a way to communicate more effectively.

[00:16:00] And the key here is that it's across the entire organization. I do this for major initiatives; teams do this for their initiatives. We even have this starting across the organization, in go-to-market and marketing as well, so that when you say, "Hey, I'd love for you to join a solutions review," they get it. It's not that you have to influence and figure things out. It's just a common muscle, and it's a lightweight muscle. And by the way, back to executive modeling, which I am one part of: we have this tracked too. If I'm curious about one project and what the status is, I can just easily go to our project tracker, see the next milestone and the next milestone date, done. I don't have to meddle. I'm like, "OK, cool, I can peace out for a week. I'm good. Figure that out."

Inclusivity: bad version and over-beers version

[00:16:47] The second thing is inclusivity. Really important. If you want to create this culture, people have to feel involved. I carried this through from our InVision days; this was called "bad version." Maybe it won't work for you, but the key here is that people are scared to offer an idea that's not fully baked, it's not there. So you just have to say ahead of it, "Hey, bad version: what about this?" It's communicating in a low-fi way. And what you see is people offer up more ideas. You also see that people don't litigate the idea if they say "bad version." It's not, "Well, have you thought of this?" So you have this moment where that yes-and momentum starts happening all of a sudden. Bad version.

[00:17:31] This is a new one we've done. Has anyone seen, when you're going through work and you're working a project and it's taken a while, you start saying things that a normal human wouldn't say, in the vernacular of your organization? It sounds really weird. Somebody's trying to craft an impact sentence for an OKR, and everyone's like, "What?" So we actually have these moments where we say, "You know what, let's take the heat off of this. If you were to explain this project over beers, or drinks, or coffee depending on your taste, what would you say?" It's not an elevator pitch. An elevator pitch is scary. It's like, if you're not done compelling us to say yes in 90 seconds, you're [ __ ]. But this is about saying, "Hey, casually, if you were talking about this..."

[00:18:14] And it's really transformed a lot of conversations, because somebody will say it and someone else will be like, "Why don't we just say that?" I'll be honest with you: at the leadership level, when we got completely contorted on our OKRs, we said, "Wait, what's the over-beers version?" We said it out loud and we were like, "Why don't we just write that? Everyone would get that, not the weird contortions that we've done." So this is another way to lower the heat and make it a little bit more accessible to all.

No one is perfect

[00:18:37] And then finally, no one is perfect. The process and the journey that we're going through for transformation is not perfect either. So this is all about how we embrace that. How do we embrace the fact that this process is messy? It's not consistent. We're making mistakes left and right. So we have two things. One is what we call inertia busting. If somebody is having an issue going through this experience, they say, "I need help inertia busting this." You dig in: "OK, I got it. Oh, that person's not listening, they're doing this other thing, they're telling you to file three Jira tickets, blah blah blah." Great. The other is fail nights. We actually just had one. People are sharing their fails; I actually shared one of my failures. It's really helping people understand, "Yeah, let's take more risks. Let's really push that along."

Dogfooding: Miro using Miro

[00:19:24] And then finally, dogfooding. I refuse to say "drink your own champagne." This is where it gets a little meta, but Miro is using Miro to transform Miro, to reinvent Miro. It really is one of those things, and maybe we get to do it because we're a collaboration platform. So we are collapsing the stack. We're accelerating synthesis between unstructured and structured work. We have our new formats in the experience, we have AI shortcuts. So we are taking all those brainstorms, and with one button that we've tuned for our purposes, it creates a document. Teams have gone from waiting a week for product managers to synthesize, to feeling like they are co-creating the definition, as it should be. The product manager is still accountable for the final version, but they are co-creating that definition.

[00:20:12] We are really mining groupthink. Groupthink is a team that's working well together, and they're brainstorming and high-fiving each other, but they've been thinking about the same ideas for a long time, so they're triangulating. So we have these AI sidekicks that will look at the brainstorm, yes-and it, and offer new ideas.

[00:20:29] We are taking our brainstorms and synthesizing them into prototypes. This is meant for ease of communication; this is about visual communication. So we're taking these experiences and making sure that we can share them. We are looking at this for product managers and designers. This is by no means anywhere close to code or even final design, but it is a helpful tool to help you understand how you articulate ideas, and we're going through these experiences, again, from unstructured to structured work.

[00:21:02] We're taking insights. We have various insights integrations, with Glean and others, and we're saying, how might we bring that to your fingertips? It should be where you're working. You shouldn't be like, "OK, pause, let me go over to the insight store, figure this out and bring it over." So you can grab that information. This is one of our own bespoke versions, to take product feedback and bring it in where the teams are ideating.

[00:21:27] And finally, we're trying to connect the entire end-to-end process itself: the processes that teams work through together to go through these various phases, these various formats, even third-party experiences like Figma references. So every designer here no longer has to keep sending the link over and over again on Slack to their partners, so that teams feel like they're all working on the same project together, not in fragmented ways.

[00:21:56] We've dogfooded this and it's accelerated things. Some days it feels like this: we're jamming, we're feeling good. And other days it definitely feels like this. There is no perfection here; we are not saying we have it all figured out. But the interesting thing is we are shipping impact faster. Our employee engagement scores are meaningfully going up. So we are on the path. And what I try to articulate to our team over and over again is that it is a journey. There is no end. And in my darker days I'm like, every company's a dumpster fire, so let's get over it. One of those works.

[00:22:38] So what is our opportunity here? And I'm saying ours as EPD. I said this in the morning and I mean it: EPD is the heart and soul of most organizations. If you are building the product, if you're building the systems, if you're building the tangible thing that people use, and everyone else is talking about it and selling it, you are the heart and soul of the organization. So if we know this is happening, let's use AI as that opportunity. Because I believe we can address how we work better together, and how we foster the best moments to make that the norm, because it is what we've always wanted all along. All right, thank you.

Q&A

[00:23:32] Host: That's a nice way to end the day, isn't it? A little bit of hope. I feel like there's a sub-theme woven into your talk, which is that these models, this technology, are going to keep evolving, but if we don't evolve along with it, the best parts of us, the inclusive parts, we're not going to make it. It's not going to work.

[00:23:50] Jeff: Yeah. Back to hopefulness: there's a lot of scary stuff out there, but I would say if we choose to assume the positive, you probably have more mental capacity to see how you might roll these into positive outcomes.

[00:24:04] Host: I'm sure, attending talks like this. I love the idea of, hey, what if every day was like a hackathon? What if it always felt like the design jam? That's a nice thing you can wrap your head around. Some of you might leave with that idea and tell others about it. Maybe we can all look at this with the beginner's eyes that we need right now. Do you have any other tips on how you can help other people understand that maybe it's us that has to change along with the world?

[00:24:27] Jeff: Yeah. I think the easiest thing to do, honestly: I haven't talked to a single person here, or anywhere I've talked to teams, who can't point to a single team that exhibits these behaviors. It's just that we tend to treat them as the exception, not the rule. Every organization has its unique isms, so I would never assume that what works for us would be what works for you. But they broke through somehow. The pattern I do see is utter transparency. A lot of our teams that are at the forefront of this record a Loom video of their work in progress every day. Every day there's real excitement where people are yes-anding that Loom video and talking about it. Every day they're creating some experiences. So there are a lot of areas where you're like, "Oh, well, that's a really interesting way of unlocking and getting that inertia a little bit unleashed."

[00:25:30] And then by the time they hit our solutions review, it's been so transparent that it's just a nice kumbaya moment. It's not even a "Well, gee, have you thought of this?" because everyone's been part of the journey in the most lightweight way possible. So how do you model that at scale? I will tell you right now, we're going through that process. And back to the kids' chaos GIF, I can tell you some of my team up front here are like, "Yep, not there yet." But we're trying.

[00:26:00] Host: We had a nice innovation here. Somebody asked, "What is EPD?" and then someone else answered, "Engineering, product and design."

[00:26:06] Jeff: Yes. I think we call them product teams, but then I get a little bit antsy in a UX area, where we say "product" so much. So yes, EP and D.

[00:26:17] Host: The top-voted one right now sounds more like a complaint than a question: "Feels like as a designer I need to do 100 more things in addition to my current workload, and babysit AI while doing all this. How can I survive?" Oh, we have applause. You can applaud if you like that, or if you don't like it, I'm not sure.

[00:26:38] Jeff: Yeah. I would actually say I would start by looking at the workload first, and litigate that, to be honest with you. We actually just had our leadership offsite and we were talking about workloads, and we haven't got it figured out. But you start realizing little things, the paper cuts that are happening. A design crit where you want to show every single click path, so the artboards look like a flip book. Engineering having already figured out that they should start on the back end, the endpoints aren't going to be changing, it's really some front-end changes, yet they're still asking, "Have final designs been approved before we start?" And that puts more heat on this experience.

[00:27:24] I would say there needs to be a little bit more inertia busting, and room in the system for us to operate nimbly. And by the way, this is still my first hypothesis: we should change that anyway, regardless of AI. There are little motions like that which make designers feel more like production pixel pushers, and you can take those off the table. But you have to make that change first. That frees up the bandwidth to start thinking more profoundly: "OK, my value here is actually the storytelling up front, solving the hard business problem through a customer experience," etc. So again, it's what we've always wanted. It's just how do you free up the space first, and then you can dig in.

[00:28:07] Host: Also, you showed this with your waves, and Pamela talked about this when she showed the graph: we've got two waves overlapping a little bit right now, post-2022 and also AI. Do you want to talk about the competition?

[00:28:27] Jeff: OK, competition. There's a lot of it. I wouldn't even just say it's FigJam; there are the free bundles, Microsoft Whiteboard and others. I saw a screenshot today where somebody was using that, and I'll talk to you later about it. I'll find you. I will hunt you down. But look, I think what we've found a lot in our competition is, again, standard competition is healthy, but we have to focus on what drives value. When we talk to our teams, it's about that cross-functional balance, about working through different experiences. One of our top-used areas is business process mapping and customer journey maps together, and how they translate. That's not happening in FigJam. Customer journey maps maybe, but not the other side. So we're trying to make sure that we are focusing on the use cases that we can solve. We also believe that FigJam is going deep into design and content, which is fantastic. They've got Make, they've got the other stuff, and that's cool. And we are going to be making sure that we are... what's so funny?

[00:29:37] Host: Someone discovered, I don't know, last year, that you can use this to troll.

[00:29:41] Jeff: Oh, got it, got it, got it. OK.

[00:29:42] Host: You can also open up incognito browsers and vote up your own ideas.

[00:29:46] Jeff: Oh good, I thought you were spinning us. But anyway, it's an interesting competition. We have a couple of tricks up our sleeves, including the democratization of design, not removing that from the designer, but that ideation up front is really important, and then growing that remit with others as well. So we think Figma and Miro can work like peanut butter and jelly. We do know there's a little bit of a turf war in the sandbox on visual collaboration. It's all good. It's a fun space to be in.

[00:30:16] Host: We have an incredibly neutral question about the validity of AI-generated information and vaporware. I don't know if you want to talk about it. Oh, we've got another one coming up. Let's see.

[00:30:27] Jeff: Got it. Well, I'll answer this from a Miro perspective: the hallucinations and the other stuff. In the discovery and ideation process, we never think ours is sticking the landing, ship it. This is about accelerating processes and steps, collapsing those steps. So we think we can tackle that and show real added value while teams are working through it. Some of this stuff is really important for us, and really important for some of our customers who are working through it. We are also a democratized platform, so we can use the keys of our customers' platforms. Big pharma, for example: they use OpenAI or others, and they actually use their instance that they've trained, they've figured out, they've narrowed down. So it's all about us adding value to the ROI of the systems that our customers are already using.

[00:31:18] Host: And I don't think this would be a full day if we didn't have something that trashes PMs here. So what if the PM goes off on their own, to hell with yes-and, it's "yes, thank you," and just walks away?

[00:31:33] Jeff: It'll never work. I mean, they'll do it once. No, it'll never work, for two reasons. One is that no one likes a cocky [ __ ]. I'm sorry. No one in an organization does. They lead to real toxicity. And by the time you want to unleash it, you've already made 20 enemies. And I believe profoundly that it won't even deliver value. Building something that you can prototype, that people can click? Yeah, sure, that's awesome, but let's go put it back through the process. And it's not just design. It's about bringing everyone through the journey: getting your marketing organization involved, product marketing, analytics, since it's probably not measured right, the SRE teams that are on the clock if it doesn't work and it burns the servers down, and they are the ones that get paged. It's a fool's errand. Prosumer, sure. Early solopreneurs, fantastic, have at it. But in organizations that have a business they're operating off of, it's not going to happen. Or when it does happen, you can be that person in the meme where the girl is smiling and the house is burning down in the background. You have permission to do that. So that's my belief.

[00:32:48] Host: Thank you very much, Jeff. All right, cocktails. Let's go. Cocktails.