Plan Zapiera na nowoczesne zarządzanie projektami

Galen Low: In a time of job market turmoil, is it cruel for organizations to be focusing on AI led automation or is it actually maybe kind?

Wade Foster: To me, it is 1000% the kind thing to be doing. AI is very good at helping me with a huge variety of tasks. AI can help in the messy middle, but it doesn't take the job. The job is still the job.

Galen Low: Looking ahead to like, I dunno, 2030, what kinds of skills and credentials will tech companies like Zapier be seeking out for roles that are responsible for project delivery?

Wade Foster: I would say your ability to identify problems that are existing, have ideas that is exceptionally valuable. The second thing, then, you know what I would call like judgment, taste. And then the third one, how do you coordinate all these systems? It's the messy middle.

Galen Low: Do you believe that there is a threshold for agentic AI and intelligent automation that maybe we shouldn't cross?

Wade Foster: I'm not sure if there is an upper bound. Human's creativity is so far massively undertapped. The amount of creative output an individual person is able to do is enormous. I think the world is gonna be different, but I don't think this dystopian view is, it probably looks a lot like today, but just more.

Galen Low: Welcome to The Digital Project Manager podcast — the show that helps delivery leaders work smarter, deliver smoother, and lead their teams with confidence In the age of AI. I'm Galen, and every week we dive into real world strategies, emerging trends, proven frameworks, and the occasional war story from the project front lines. Whether you're steering massive transformation projects, wrangling AI workflows, or just trying to keep the chaos under control, you're in the right place. Let's get into it.

Alright, today we are talking about the role of AI powered automation in project delivery and whether it might be the key to reducing some of the bulky project management practices that makes lean teams want to avoid formal project management like the plague.

With me today is Wade Foster, the Co-founder and CEO of a little company you may have heard of called Zapier. Nearly 20 years ago, Wade quit his internship in financial services, learned to code, and together with a few friends, built a product that made it easier to get apps talking to one another and to build automated workflows between them.

Today, Zapier is the most connected AI orchestration platform. Connecting teams, tools, AI, and agents across more than 8,000 apps to automate critical workflows with flexibility, scale and control.

Wade, thanks for being with me here today.

Wade Foster: Thanks for having me, Galen.

Galen Low: First of all, Wade, I have to say it's such an honor to have you here and I need to thank you just for like being out there and sharing your stories and your insights so transparently so that we can all learn from them.

There's so much that I want to ask you, and I hope that we do a bit of zigging and zagging throughout the chat, but here's the roadmap that I've sketched out for us today. To start us off, I wanted to just set the stage by getting your hot take on one big juicy question that the folks in my community wanted your take on, but then I'd like to unpack that and talk about three things.

Firstly, I wanted to talk about what's different about automation in today's AI obsessed context compared to when Zapier first launched. Then I'd like to just pop the hood on what project management looks like at Zapier and how Zapier has achieved lightweight project management processes by eating its own dog food.

And lastly, I'd like to just get your perspective on the future of project management as a craft and how you see the role of our project manager transforming as we head deeper into AI powered automation. How does that sound to you?

Wade Foster: Love it. Let's do it.

Galen Low: Let's do it. Alright. I thought I'd start off with just one big hairy question.

I'm just gonna take a running start at it. In the circles that I travel in, there's like this love hate relationship with the idea of AI powered automation when it comes to project management. On the one hand, intelligent automation is taking away a lot of that boring admin from the job, and on the other hand, it's kind of threatening to take entire jobs and also creating this perception that project management can be fully automated.

So my big hairy question is this. In a time of job market turmoil where like mass layoffs and long difficult job hunts have become the norm, is it cruel for organizations to be focusing on AI led automation? Or is it actually maybe kind?

Wade Foster: To me, it is 1000% the kind thing to be doing. You know, if you put your hands on the keyboard and you work with the technology, right now there's this mysticism with AI.

But if you start using it, that mysticism goes away. And what's left is a much more pragmatic understanding of where is this technology incredible and time saving and creativity inspiring. And also here are the limitations and where it's not so good and where there's a human role. What I have found in the process is that AI is very good at helping me with a huge variety of tasks at any point in sort of like the delivery of a project, but it is still not taking over my job.

You know, my job is not the task that is being done. My job is understanding problems and figuring out what the right problems are to solve, and then evaluating the output of what we're making. Is the product good? Is the marketing good? Having taste around those things? AI can help in the messy middle of all that stuff.

In fact, it helps a ton in the messy middle of that, but it doesn't take the job. The job is still the job. You still have to go solve those things, and I think anyone that I've worked with who has really embraced AI quickly. Starts to produce more, do better things with it, just has more ideas for where it could go.

I understand humanity's you know, default reaction, which is to be one of fear and scarcity. I think it's only natural because it's easier for us to imagine, oh, what happens when this, but the reality of any technology is that it usually creates much more than it destroys. And I think the same will be true with AI.

Galen Low: I've gravitated towards, you know, some of the zeitgeists around developers and, you know, coders. If you think that the value that you create is how fast you can type, then you're probably missing the point of your role. And I think a lot of the same things are true in project management. I think a lot of folks, you know, there are anxieties around, okay, well someone's taking meeting notes like an AI note taker now.

Wow. Now the folks who thought that, that's all I do, they're gonna fire me. Not gonna lie. Okay. Like some of those things are true. That's how the role is seen sometimes. But I do like that expansionist perspective of actually it will sort of be creating more.

Wade Foster: Well, and was the job to sit there and take notes, right?

Like that was never the job. Like I agree. The job was to solve problems not to sit there and transcribe. And so, you know, I think if your job is truly that narrow, then okay, sure. I understand the fear a little bit more. But when I look around organizations. Most people, that's not their whole job. That is a task that they do in their job occasionally, and usually it's one they don't particularly like doing.

They don't do particularly well. And so having an AI assist you in that task is great because now it frees up your brain to think about the parts of your job that you were really there for.

Galen Low: Like coming back to something you said earlier, right? Like once you get people to actually start using it, the reality has become clear of what it's good at and what it's not good at.

And I do think there's a lot of sort of perceptions. At all layers of any organization where it's like, we haven't really tried it, but it says it does this, so I guess we can fire this whole team. Or I guess I'm gonna like be out of a job. But, you know, thematically, I think it's very true that you start using it.

You realize it's a partnership, like it's gonna take some stuff that it's good at. It's gonna leave some stuff that is not good at, it doesn't have to be like this wholesale AI transformation where suddenly the workforce is just AI, all agents just marching about. It's still a collaboration. It's still lifting us up.

Wade Foster: A hundred percent. I'll give you an example. Right now we've got our Zapier summit coming up, which is an event where we bring all of our employees together in person, you know, one time a year. We're typically a distributed team and as part of Summit. I usually give a kickoff keynote and you know, it's the one time I have the whole company in the same room at the same time.

And so, you know, I like to do a good job with it. In the past, you know, without AI, I'd have to sit there through and like think of the story I wanna tell, go collect anecdotes and data and examples and hearts of that task were fun. Like the creative element coming up with the story you write, et cetera.

Parts of it were like pretty tedious, where I'm like, I don't really have a great example of this. I don't have the data for this. I don't have these other things. And then also, you know, with any talk, there's always the blank page problem where you're just like, okay, I kind of have this like vague idea of what I want to talk about, but I'm not, I don't really have it, you know, in a storyboard yet.

And just in the last two days, I've sat down with Claude, I call 'em these brainstorm sessions where I'll say, Hey, you know, I'm gonna do an hour session with you. I want you to ask me questions sequentially. Here's kind of this messy thing I've got in my mind. You know, the end state is we're gonna have a keynote, help me just like work through this thing.

And the end result is, I'm actually putting in more time, more effort in making this keynote great. Because I have this amazing assistant going back and forth with me helping with this stuff and you know, and parts of the step are in the process. I can send it out and say, Hey. Go dig through our slack. Go dig through our coda.

Go dig through the internet to like pull quotes, pull examples, pull data. That helps me tell this point way better. I can go back and forth with on ideas and the end result. And this is maybe hard for people to believe. I will spend more time on designing this summit keynote than I will have ones in the past.

And so this is a place where AI is not actually reducing the time spent on it. It is actually increasing the time spent on it. But I find that the quality on the other side is gonna be so much better, and the actual act of writing it is way more enjoyable because I have this like just really eager thought partner going back and forth with me, where in the past I'd get to these parts in the process and I'd just be kind of stuck and then I'd go, you know what?

Like this is good enough, like I'll just call it and you know, hopefully nobody like spots the quality issues. So that's just one example. And you know, I think this is the kind of thing when you actually are really using AI, you start to just have a much broader understanding of like what it's good at, what it's not good at, what work you're gonna use it for.

The demands of these tools are just gonna be enormous because the creative potential of humanity is just way bigger than I think, you know, folks actually think it is.

Galen Low: I like those two angles, right? Like quality and joy. You get to focus on like, how can I make this as good as it can possibly be? Because I'm very much like that as well.

And I think a lot of folks who consider themselves creative are like that. Maybe everybody where it's like nothing's ever done. It's just so you can't look at it anymore and you're like, good enough. Walk away. You can kind of bypass that. You can have more fun doing it and you know, I think that's really interesting.

You didn't spend less time doing it. You spent more time doing it.

Wade Foster: No, in fact, I got to a part in the process where I had a couple threads going at once and. I started like very well organized, but then I'd gotten into the messy middle part. Claud had just finished a task and I had to stop and say, okay, we've gone and done like, you know, nine different things, like researching this talk.

I'm disoriented, so like we're gonna have to come back to the center. You know, I remember we did this. I remember we did this. I remember we pulled these data sources. I remember we have this like vague next step. Help me like reacclimate to this. Oh, and by the way, my session is done, so I need you to document this really well, because I am actually gonna come back to it tomorrow again.

And so leave me some breadcrumbs to help me get going tomorrow. And that's the kind of thing that like without the assistance of AI, you would just walk off and then you'd come back to it. And my first task tomorrow would be sitting there and. Having to just like figure out what the heck, what was I doing yesterday?

Where did I finish off? And now I'm gonna come back with like a nice index of breadcrumbs and all these things. And so it's just gonna be that much easier to jump right back in where I left off.

Galen Low: But I think a lot of people underestimate the thought partner workflow. I'm even, you know, I'm early days in my journey.

I'm like still at the keyboard having a conversation that kind of starts and ends and then I'm like, you know, I'm done. I'll feed it back. What happened? You know, I'm training it on some stuff, but it's not, I'm not like, see you tomorrow. You know? I really like that. It's like, can you go away and do some stuff?

Wade Foster: Yeah. I've started to get to like, thanks, good session. You know, I'll be back next time. I love that I started to get into that, like almost like a therapist.

Galen Low: You know, I think it's really good perspective on like how it's helping, how it's changing a role, any role really, and I think the joy and equality. I dunno, a lot of businesses are gonna struggle to like measure that, but I think it'll come out in other ways in terms of performance and value and what have you.

Wade Foster: Quality.

Galen Low: There you go. I wonder if I could zoom out a little bit and maybe we can dive into a bit of that world of like automation, not necessarily just the conversational thought partner stuff, but you know, getting into the agent stuff, getting into the intelligent workflows, getting into like MCP and stringing all these apps together, all this technology together.

There is no question in my mind. Yeah. What we're talking about is that AI is just changing the way that we work, but it's also changing like what is expected of us in terms of like our technical skills. And also in a way our strategic business acumen, right? Our ability to like zoom out and like elevate the conversation.

It includes project managers of course, but it goes way beyond that. And the way I see it, you know, Zapier's been quick to recognize that AI is changing things. And you've got products out there now like Zapier agents and Zapier, MCP, and it's kept Zapier, in my mind, in the forefront of like stitching apps together in a world where frankly we've just got like so many tools, right?

There's so many tools. There's a justification for all of it, but we're still in the zone of like. It's challenging to make them all play well together. Of course, automation in its purest form is not new. Like Zapier especially has been offering no-code automation as a service for like well over a decade.

So I wanted to ask you, what's a challenge that you faced throughout your past, like 15 years of no-code workflow automation that you see people still struggling with today? Is the way that you solved it different than how you'd recommend addressing it today?

Wade Foster: I think there's some enduring problems. You know, one that immediately comes to mind is inspiration.

You see this with AI where there's some people who are using it for every aspect of their life, their work, et cetera. And if you talk to them, you know, they'll tell you, this is revolutionary. This changes how I completely approach work. And then, you know, on the other end of the spectrum, you'll talk to somebody who's like, I mean, it's nice, you know, it helped me come up with a recipe tonight.

It told me a funny story for my kid, but like I don't see what the big deal is. Right. And then everywhere in between. And you know, when I spend time with those sort of two personas, and of course the folks sort of in the middle there, oftentimes the difference is the creative output of the folks that are really using it a lot is just a lot higher.

And they've often actually figured out how to use the AI to be a creative thought porter to help them come up with more ideas. Whereas the person who is just getting started. They're stuck. They're just like, well, I don't even know what to use it for. I don't know, like, what problems do I have in my day to day?

Where should I go and look? I think that's okay. I think most people are actually like that, but I think the big opportunity is how does Zapier, how does us as a community, how does us as entrepreneurs, how do we think through to help guide people along that spectrum? You know this is a problem pre AI.

Here's actually a funny story. Like you see this when you go talk to these leaders who are in charge of AI transformation in the organization. Oftentimes they all have this similar playbook where they'll go in and survey their workforce and say, what ideas do you have for AI? And they'll often come back with this like laundry list of ideas.

And the funny thing, is almost to a T, everyone will say nine times outta 10. What I get back doesn't require AI at all. It was something that we could have been doing all along. Traditional automation, traditional workflows, et cetera. And that just kind of goes to show like how stuck people get on like, what actually is possible here? What can I go build?

Galen Low: The art of the possible? For sure. Like that actually had been my stumbling block with Zapier to begin with, you know, pre AI stitching together APIs, right? And like I'm not very technical. I know what access points I might have in a tool. I know some of the workflows that I might wanna automate, but sort of understanding the art of the possible, I'm like, okay, well what can I sort of think through as a trigger?

How can I talk to this piece of software to like get this workflow done? But I think the inspiration thing is a really interesting piece because a lot of the folks I'm talking to in that, you know, category A, right? They're using the tools a lot. In a way, there's this sort of being okay with failure, right?

You kind of like you get inspired. Not failure as in it didn't work. Failure as in, okay, the idea I had that I wanted to try, I don't need it. It's not exactly what I wanted, but I'm gonna keep going. And they're like learning lessons as they go versus the category B, which I definitely relate to. Like the paralysis of like, what can I do with this?

Like I had. That app, if this and that on my phone for like 10 years and like I'm like, oh, this is great. I can like automate things in my house if it's raining outside. I can like set a reminder to like bring an umbrella, didn't set up a thing, didn't open it once. I don't even know if I created my account because of that thing that you were saying.

Right. I didn't have ideas or at least confidence in the ideas I had to like invest the time to build something with it. I think that's a hugely good point. Zapier has been game changing. A lot of the sort of like tools that are out there now, we have like AI platforms and LLMs, you know, connected together.

For the folks who are kind of like stuck because they're worried about where their data is flowing, what kind of thoughts do you have around like building guardrails for some of these workflows? For some of the folks who are just like, I don't really wanna plug all these things together. It sounds enticing, but I just don't know where my data is going.

How's Zapier tackling that?

Wade Foster: I think the nice thing about using a tool like Zapier is that you have the ultimate control of where your data is going. Like you hook up tools to authenticated data sources. You know, okay, I'm gonna take this from a spreadsheet and put it into Salesforce. I'm gonna take this from Salesforce, I'm gonna put it into an email newsletter tool like you are predefining, all of that stuff.

So that control is. The nice thing about a tool like Zapier, I think the second thing that is a smart practice that I see our sophisticated users doing is the more you're getting into letting the AI take control of those workflows. Oftentimes you're asking to make a decision or maybe you're asking to generate something, write an email, write a message, et cetera.

Oftentimes they'll have a human in the loop step where they'll say, Hey, I'm okay with you automating all this entire chain up until this point, and right here, this is when I want to come in, inspect the work and decide, okay, good enough, let's send it on its way. And so that's like the practice I see a lot of folks doing.

Now, where I think we're heading is I do think we're gonna get to a world where increasingly over time, if you're building these systems appropriately, you can start to take the human out of the loop. Now, I don't think you're gonna a hundred percent take the human outta the loop instead. I think what it's gonna look like is a lot like how manufacturing facilities run.

Manufacturing facilities have quality control bands and they use statistics to help them understand which, you know, if they're meeting that quality bar. And usually what they do is they'll sample widget's off of the run and you know, they'll take a look at one or two widgets. And if those two widgets look really good, they can statistically prove that, you know, the next hundred are gonna be within their acceptable quality bands.

And so I think you're gonna start to see more and more of that exist in these like production scale AI pipelines where you can start to take the human outta the loop and then use statistics to figure out how often does the human need to audit this to make sure that we're staying within this statistical quality bar that we think is important for us to fulfill our service obligations.

Galen Low: I'm really glad you took it there because like as I was sort of planning and prepping for this, the notion of automation, and I said it earlier, is not new, but it's really not new RPA and some of the things we've done in manufacturing, like we've been doing it for decades and I really like that sort of viewpoint that I think a lot of folks are just, you know, they're worried that it's gonna be all agents all the time.

AI running itself, that humans will be EF faced from that loop. Maybe that will happen, but like in the near term, I think that's like prudent the idea that actually this is a collaboration and then it will get to the next step, which is still an oversight kind of, you know, situation. It's not running free sky netting and taking over the world.

We can just like observe it work and probably best because it's working really fast. And if we were inspecting every widget off the line. It just wouldn't work.

Wade Foster: Yeah, you don't get the efficiency gains. It's not as fun that way.

Galen Low: I like this sort of thread of steel around this, like the fun bit, the joy of work, you know, and maybe I'm getting too poetic here, but I really like the perspective that, yeah, I think a lot of folks are like scared that it will automate the joy out of their work and automate them out of a job, whereas actually it could be the opposite.

You could automate the boring stuff. You could choose not to automate something that actually brings you joy. And to your point about the keynote, it might take you the same amount of time or longer to do that thing that you find fulfilling. And there might be a world where your relationship with your employer or your team, like that's worthwhile.

Right, like I can see it in the keynote. Why is it worthwhile for your keynote to be awesome or for you to enjoy it? Because when you stand in front of all those thousands of people at the summit, like you're inspiring them. You're delivering a message, you're telling a story that's gonna rally people together and that's gonna set the tone for the entire rest of the year.

Like that's important. And I think that's screwed value framing for, you know, how people can think about what they automate and I guess like what they don't automate. I wanted to zero in on that efficiency thing because you recently did an interview with my team about how Zapier is using AI to automate and streamline parts of the project management process at Zapier.

I'm just wondering, like, what are some examples of how your team has been eating its own dog food to achieve lightweight project management? Like what's been automated and what parts of project delivery have your teams maybe decided not to automate?

Wade Foster: Well, it's been maybe six months since we did that.

Galen Low: Yeah, I think so.

Wade Foster: Yeah. So, you know, I, I don't know if it's as up to date anymore. You know, I think the space is moving fast. A couple of principles that I think are important for us now, first, I would not necessarily put Zapier on a pedestal and say, Hey, we are masters of project management. Follow us.

We have conquered the mountain. But I do think we get some things right. So first. I think, and I believe that projects tend to go well. The smaller you can keep them, the smaller number of people that own the process end to end, the more likely you are not going to have big, you know, overhead to manage the project.

You can have a small team that can run free and get the thing done, less coordination, all that sort of stuff. So as best as possible, you wanna break down the work you're trying to do at a company in a way that allows these smaller teams to run free. And so we generally try to do that, and that means that there can be a lot of like lightweight project management tools and things like that to help folks get stuff done.

Now. It's not always the case that you can do that for every project. Some are just like big hairy cross-functional tasks. And in that case, one of the biggest ways that we use AI is to help with handoffs. So we wanna pass context from one team to the other. And so you can use that to, you know, summarize the outputs of one system.

So say like, product development team is working out of j. But the marketing team is working out of Asana. Okay, great. Let's figure out how to summarize like a key, you know, events when the status of this thing gets marked as closed or when the status of this thing changes to that, we wanna go take all this information that's like cataloged in this.

We wanna summarize it in a particular way that's relevant to this other team and make sure that's automating an update over into their system in a really good way. And so that helps with those like coordination, those handoffs, those overhead challenges. The second thing that we spent a lot of time doing, especially in these big cross-functional ones, is oftentimes there are tasks that are required of large groups of people who are only, you know, peripherally involved in the project.

Oh, now we need every single person in the company needs to go do this one, one little task over here. Right? Or, you know, this team needs to go do this. Or like, these nine teams need to all sort of, you know, come together and like get on this new paradigm. And almost inevitably part of the project manager's job is to go be the nag bot.

And nobody likes playing this role. That person doesn't have a fun time doing it. No one likes it when that person comes and knocks on their door. It's not exactly a glamorous part of project management. But it is a very easy one to automate where you can set up alerts and systems that go ping folks and let them know like, Hey, this deadline's coming up.

I noticed you're on the like incomplete list. I wanna remind you, you need to go do this. Can you check in? And you can also send those same alerts to their manager, so they can have checklist to go inspect. So in case you know, that manager needs to go reach out and say like, Hey, the automated system isn't alerting like the human.

I will now sort of come in and take a look at this system. That's another one that like is a pretty helpful, like very lightweight tool that can help in these projects that are just bigger and more cross-functional by nature.

Galen Low: It's so funny because at an agency I worked at, we had a nag bot for time sheets.

I mean, this is early days. It wasn't a workflow at all. It was just an automated reminder into teams to be like, Hey, time sheets. Do them. And you're right. I think it's like, here's, again, like that theme coming up. It's a part of the job that I'm sure there's project managers who love it, but most of the people in the circles that I travel in, they hate it.

It gives them that label of being like babysitters, cat herders, right? Like those things that kind of undervalue the role, it perpetuates some of the assumptions that like that's what they do, right? You take notes and you nag people about the thing, right? It's not the job. It's about problem solving.

Wade Foster: And here's an interesting thing.

A smart creative project manager will say, you know what's interesting? A thing that I can do with AI is I can actually do fun stuff with those nags. So instead of it just simply being, you know, a boring like status update, which is, Hey, you're supposed to complete your time sheet by this week. Thank you for your commitment to making our process run efficiently.

That's kind of boring. With AI, you can start to do more fun stuff. You can say, Hey, I would like you to write it in the style of a limerick. I would like you to add a funny picture to it. Like you can start to do things that now we all know that one project manager you worked with in the past that like somehow manages to do this just by, you know, force of personality.

That's not most of us. That's certainly not me. But you know, I would like to be that person. I just know that aspirational, probably not going to be, but with the help of AI, I could kind of mirror it. I can do a decent job of it. And so that to me is also, again, where AI gets to be pretty fun. It's not just the efficiency part of it, it's what's the quality you can put into these things.

It's just something that's so hard to do that at scale. You know, if you're a small team and a small project, sure you can do it. But if you need to like send a customized limber eyes note to 120 people, it ain't gonna happen.

Galen Low: I do like that because I think some of the people who do it naturally, right, like that is their mo.

The mo is, I'm gonna write it in such a way that this person's kind of gonna, you know, smirk or enjoy it. But it could also be the reason why someone's like, but I don't wanna automate that 'cause everyone's different. It's how I build relationships. I do like that mini thought partnership. Hey, I do this Limerick thing.

Or we have like a sarcastic British standup comedian persona that we use sometimes for internal comms. Please take on this persona and like, you know, switch it up a bit and maybe learn some of the preferences of the team so that this becomes a fun thing that is kind of like a thing that we can talk about that brings the joy back and doesn't destroy that, you know, relationship building of please fill out your timecard thank you management, right?

These memos that you get that honestly, I end up tuning them out, right? The ones that are very automated. They don't cajole me into action the same way that a human does, but I do like that persona. That's very cool. Just in terms of like small teams, you, we mentioned earlier your teams are mostly distributed.

I think that sort of knowledge transfer is a really important bit, right? So selling information between teams. What about teams that are like not distributed? Are they at a disadvantage because they haven't digitized some of their comms? Is there maybe even no such thing as a organization in the digital space that's not sort of using tools and I am, and stuff like that?

Wade Foster: Well, I think you brought up the important point of this, which is are they at a disadvantage by not digitizing now remote, not remote, that's kind of whatever you want. The digitized thing, if you really wanna go take advantage of AI, you do need to get it into text somewhere. Maybe not necessarily text.

You could probably use audio and things like that and then transcribe it. But ultimately. These are large language models. They work great with text. And so if you have a practice as a team of just documenting your work, getting things out into public, it's gonna help these tools go ingest those.

This is, you know, one of the reasons we've had this value of default to transparency internally, pre LLM. We always just thought of it as, Hey we're here to assist humans. Like it's great when you're putting this stuff in public channels in Slack because now when the new employee joins six months from now and they're trying to figure out what was this project?

Why did it matter, et cetera, they can go trace back those digital breadcrumbs, and so we always believed in that process. It turns out that when you put an LLM on top of this stuff, it's like jet fuel. Now I can send an agent to go crawl through all of this and say. You know, give me the context on every customer who has ever had this particular problem inside of Zapier and return it to me in this particular format where, you know, in the past, like, okay, I technically could do that, even if it was, you know, well documented, but it still is somewhat tedious now.

I just like prompt it away, go work on another task, and then like back the answer comes. And so I do think that teams who really just work hard to create exhaust of their work that is documented somehow. And when I say documented, I don't mean like someone sit down and wrote like very proper documentation.

This could just simply mean, record your meetings, make sure you have a transcript of your meetings, and that's good enough documentation for an LLM to take it and say, great, I know what to do with this now. So that's kind of what I'm getting at, is it just needs to like exist in some format so it can go do work for you.

Galen Low: The D word is a scary one, right? Documentation sounds tedious in and of itself, but really what we're saying is just like having it transcribed into text so that somebody or something like an LLM can crawl through this. I'm curious actually, because I've seen some organizations try to do that, have conversations in public and record stuff and blah, blah, blah.

I've seen them try and fail, and I think it's because people still don't feel safe to do that. You're talking about Zapier, like pre AI, but still just wanting to have information out in the open. Like how did you build that culture where people were like, yeah, okay, I'll record all my meetings and put them in a public channel and that's fine.

You know, I won't back channel somebody and be like, oh gosh, let's only say the smart things out loud and then DM me with all the bad news.

Wade Foster: Well, you know, a lot of this we try and do with cultural rituals and standards and things like this. We've had the default to transparency value for a long time. And so as a leader, you know, I try and have a habit when someone dms me, I'll be like, Hey, can we move this to this channel?

Can we pull this over here? Et cetera. And just like those little nudges and stuff like that over the time, and, you know, that tries to reinforce like the way in which we wanna work. Now a credit to my EA Courtney. She figured out last year that Slack actually gives you analytics on what percentage of messages go into a DM versus what go into a channel.

And so she said, Hey, I know another way we can sort of help keep people accountable to this, but also make it kind of fun. Is I'm gonna create a leaderboard. And so for the exec team, she now publishes once a month the leaderboard for how people did in a given month. And so, you know, just to give you a sense, you know, and our executive team, you know, we had the top three people were north of 60%, and then every single person on the executive team were above 50% in terms of, you know, talking in public channels.

And so we look at that and we're like, that's decent, but we've had months where, you know, our top folks are getting like almost pushing to 80%.

Galen Low: Oh wow.

Wade Foster: It really is possible to get like most of your stuff happening in public. Like I think of Zapier as like very transparent, but I've heard from organizations that are even more extreme with them.

I would be keen to see their metrics because it would not shock me if the upper limit is actually more like 90% for, you know, sort of public transparency on this. Maybe even 95% if you really wanted to push it. But if you're kind of just like normal transparency people, I think you should try it for more than 50%.

Just like get more than half of your messages out in public.

Galen Low: I like that sort of baseline and maybe even the upper limit, right? Where humans were social creatures, but we're like not all that social Part of it is sometimes you need to send a message that is not public and that's just the reality like a hundred percent won't happen.

Wade Foster: Yeah. That's why it's default to transparency. You always are gonna have something confidential that needs to be discussed. Maybe it's a. Some sort of like employee relations issue where there's like, you know, performance problems going on, you're not gonna put that in public. You know, maybe your company's working on a confidential acquisition or something like that, you know?

Okay. That can't be out in public, right? There's always topics and things like that probably have to be kept a little bit quieter to a smaller group of people. But I find that more things generally can benefit from being in public than most people think. And once you build a habit of it again, that fear goes away.

You know, I can talk about bad stuff in public and you know what happens when everyone's just talking about it? It's kind of normal. Like it's just how you go solve problems.

Galen Low: It strikes me that the fear, right? The folks at Zapier, they've managed to sort of experience it in such a way they're not like, oh, I'm not gonna do that again.

Like, oh, yeah, Wade just busted into my public chat and like, chatted me publicly, you know, told me I was an idiot. Like, I'm imagining that's not happening. It's like, feel safe actually, to be transparent.

Wade Foster: I mean, we have tough discussions in public, like I will give, you know, public feedback and critique and things like that, but the vast majority of messages don't have that experience because like, I can't be everywhere.

Like, so I'm not gonna be critiquing everything. To me, it kind of boils down to, like, your therapist might tell you this, like when you're obsessing about going to an event, you're like trying to figure out, you know, what's the right thing to wear? Should I wear this or that? Am I looking too dressed up or not dressed up enough?

Or you know, you're an event and you tell a joke and it doesn't really land well and you're just like, you're so nervous about like how you're showing up. Did I do it right? Did I not do it right? And generally what the therapist will say is like, people are just not thinking about you that much.

Galen Low: That's fair. That's a really good perspective.

Wade Foster: And so, you know, I think humans just like, we're just so narcissistic. We're just so thinking about our own world that like it actually is holding us back. From putting more information out in public and realizing the benefits that come with that, you're holding back all this goodness that you have to share because you're worried that one tiny thing might not be perfect.

And in some ways that's kind of the selfish way to go. Like if you actually put it out there, you would be doing so much more of an advantage to everyone around you. So you kind of just have to get over that like tiny little fear hump so that you can go experience the promised land.

Galen Low: That's a wonderful way of looking at it, honestly.

It's like, yes, get over that little thing so you can make you know things better out there. It's probably not gonna come under as much scrutiny as your like egotistical, internal anxiety is making it and it'll probably benefit somebody and that's worth it. I wanted to circle back because I am that guy, the project manager guy, but I'm like, well, I've got you, right?

You have project managers, program managers, you know, producers and what have you. At Zapier, I've got a community of folks who, you know, are project managers by title mostly, and yeah, they're diving into automation. They're diving into AI, and it's kind of like raising this question about like what. Will be valued about the role of delivering a project in three to five years time.

So I thought I'd ask you, like, looking ahead 2030, what kinds of skills and credentials will tech companies like Zapier be seeking out for roles that are responsible for project delivery? And actually, will it even be a PM titled role?

Wade Foster: So, a good question. You know, I think what I am experiencing is there tends to be.

Two, maybe three skills that are growing in value at the moment. So first I would say is your ability to identify problems that are existing or identify opportunities that you might be able to go solve, have ideas that is exceptionally valuable because these AI systems, they're inert, they need prompting.

You've gotta kick it off somehow. And so the folks that are like saying, oh, this is something we should work on, or, I noticed this was a problem, or like, how come we can never solve this? Et cetera, and they're figuring out, okay, I want to go do something about that. That Kickstarter, the firestarter role, like growing in value.

The second thing then that is really important is. You know what I would call like judgment, taste, you know, these AI tools are able to create so much work now and you know, we've all heard of the phenomenon of AI slop. And AI slop is, you know, somebody who is just generating a bunch of stuff with kind of not a lot of thought, you know, a fair amount of carelessness going into that.

So the value of somebody who actually knows what is great, you know how to make a great movie, how to make a great software product, how to write great music, what is great art, what is great images? Someone who really has good taste and good judgment around that. That also not going out of style. That is super duper important.

And then the third one, and this is the one that's probably like changing the most, that has the most evolution going on with it, but I still think is very important, is how do you coordinate all these systems? It's the messy middle. It's the chaining of the prompt and using these tools. But this one is just very fast paced.

It requires folks who are very good systems thinkers, they're really good at understanding these, breaking down the problems, but they're often also like, you know, staying up with the trends, figuring out how to push the envelope, et cetera. Whereas I feel like the first two more like fundamental skills, like those feel, like things that are like, you know.

Steven Spielberg is always gonna just be a great director because he just has such good taste in like how to do stuff like, you know, you could put him in the stone Age or you could put him a thousand years from now and he's still great. Like, I don't think that's changing at the end of the day.

Galen Low: I like it and I agree with you because it's kinda hard to put on your resume, right?

Oh, I'm great at spotting ideas and I have got great taste, but I do like that balancing act, prove. Yeah, exactly. Yeah. We don't all have the same kind of privilege as Steven Spielberg of having our work out there and like, you know, critiqued. But I do like that balance, right? Because I think a lot of folks right now, project managers and otherwise, they're worried that either they have to become a systems architect, systems engineer, like a builder, or they have to like double down on, you know, human skills and somehow demonstrate their value in the workforce based on that.

But I like that the picture you painted is actually both. Be able to see problems and be creative and have ideas about how to solve those problems, right? It's that inspiration thing that we talked about earlier. There's that taste, right? It's like, okay, well quality, right? Well, we were talking about that earlier as well, and like, what is quality and like, how do you achieve it?

I, you know, I'm thinking about your keynote, right? Where you're like, you could spend less time and make it AI slop if that was your goal, but it wasn't. Sometimes it will take you longer to do that thing. It'll be more enjoyable and the quality will be higher, and that's worth something, you know, in the world of work, in the world of tech and otherwise.

I love that there is a quote unquote technical component of just keeping up because the world's turning really fast. AI is embedding itself everywhere. Automation has been around for a long time. It's not worth it to ignore it. You kind of have to be savvy enough about it. And fairness. You could have fun solving some of the problems yourself, right?

You identify problems, you can solve them. You can have the taste to understand what is a quality solution. I wonder if I could round out just by getting a little bit philosophical, you know, we've been talking about automation, we've been talking about AI. We talked a bit about agentic AI. We've talked a little bit about the sort of upper limit.

I'm just wondering, Zapier, as a company and you yourself as one of the founders, do you believe that there is a threshold for agentic AI and intelligent automation that maybe we shouldn't cross? Like how much automation is too much automation, and how will we know when we get there?

Wade Foster: Gosh, that is a theoretical question.

I'm not sure if there is an upper bound. You know, I think humans, creativity is so far massively undertapped. All of us are capable of far more than we even believe ourselves to be capable of. And oftentimes our individual tools and our collective tools. Are holding us back and I think in this sort of first chapter of AI, we're starting to feel that where the amount of creative output an individual person is able to do is enormous.

You look at the people who are really adopting this stuff at staggering rates, their productivity is massive. And to me, I look at that and I'm like, well, why couldn't all of us do that? Like, why couldn't all of us be able to sort of deliver that kind of stuff? And I don't mean it from a purely just like maximize economic dollars out to the system, you know?

Yes. Some of this is like powering. Industry, but some of it's, you know, making games, making art, making entertainment, things that humans love and just doing more of it in different ways and unique ways. And I think that like the more we can get humans to like unleash their creative output, the better that is for all of us.

So I, yeah, I don't know what the, like upper bound of that is. And I guess I don't share the worries of folks that say, oh, well, you know, we'll take over all the tasks from humans. I'm like I just don't see that, like, you know, you think about chess, humans are clearly inferior to machines at chess, and they have been for a long time now.

They just are, we're just not as good. And yet chess is more popular than ever. We love watching humans play chess. Even though there is clearly machines that are just far better than humans, you know, you look at stuff like that and you're like, I think the world is clearly gonna be different. But I think this dystopian view is, I don't know, it's probably not that, probably not the utopian view.

It probably looks a lot like today, but just more.

Galen Low: Yeah, we tend to think of extremes, but you know, I really like that position that there might not be an upper limit on automation and AI because there's not an upper limit on human creativity. Yeah. We're not gonna like run out of things to think about or do.

Wade, I feel like that's a good place to end it. Thank you so much for spending the time with me today. Again, I could pick your brain for days, but I wanna be kind to your time and I don't wanna hold you prisoner here. But thank you so much again for coming on the show, for sharing your perspectives and your experiences and your stories.

I love what you're doing at Zapier. Is there anything sort of hot and new coming down the pipe at Zapier that folks should pay attention to?

Wade Foster: Folks should check out agents. These are a new way to build automation. You can come in and describe it in human language and you can run these agentic systems in a purely automated way.

If you're using tools like Claude and Cursor and Claude Code and even ChatGPT, you should check out MCP. These AI systems are so much more valuable when you can pull your actual context in, when you can say, Hey, take stuff from my email, my calendar, my Slack, my. Meeting recordings and use that as context to help me solve this problem.

You can do so much more. And I find that the folks that start to use MCP are often the folks that experience that first major unlock in terms of going from like basic AI usage to more transformative AI usage.

Galen Low: Wade, I'll include some links to Zapier MCP, Zapier Agents. I'll link to your profile as well. Hopefully people won't have you too much, but I wanted to say thanks again. Thanks for coming on the show.

Wade Foster: Thanks for having me.

Galen Low: Alright folks, that's it for today's episode of The Digital Project Manager Podcast. If you enjoyed this conversation, make sure to subscribe wherever you're listening. And if you want even more tactical insights, case studies and playbooks, head on over to thedigitalprojectmanager.com.

Until next time, thanks for listening.