Warum KI die Teamzusammenarbeit beeinträchtigt (und wie man das behebt)

Galen Low: Brainstorming with a declarative, sycophantic, know-it-all AI can be frustrating. Doing it as a team sounds like a nightmare. But if we were to set aside our biases, leave behind our psychoses, and truly understand who's better at what, AI could potentially become one of the strongest enablers of collaborative creativity since the invention of, I don't know, poster size sticky note pads.

So how do we get from wanting to throttle our LLMs in private conversations to having it in the room with our teams, adding value throughout the collaboration process? To explore that, I brought in a former agency head of AI and creative leader who is on a mission to solve the AI collaboration puzzle.

We'll be talking about how to address our team's attitudes towards AI-supported work so that collaboration flows more freely, what it could look like to have AI in the room with us when we're ideating as a team, and how we can start using AI more collaboratively right now. Hope you enjoy the episode.

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.

And if you've been liking what you've been hearing from us lately, please consider following us wherever you're finding us, and maybe even leaving us a review. All right, let's get into it.

With me today is Luke Alexander, founder of Supervenient and former chief digital and AI officer at Four Agency. Luke has spent his career as a passionate digital generalist, obsessing over the intersection between people, technology, and creativity. At Four, he led the agency's adoption of AI across hundreds of employees and helped explore how generative AI could create value not only through efficiency, but through new ways of thinking and working.

Now, with Supervenient, he's exploring what comes next, moving AI away from the solitary chatbot use cases towards something that feels much more like a shared and democratized experience.

Luke, thanks for hanging out with me today.

Luke Alexander: Galen, thanks so much for having me.

Galen Low: I'm super excited to have you on.

We've got a mutual friend, Harv Nagra, who has an, an amazing podcast, "The Handbook." If you haven't checked it out, check it out. I'll include a link in the show notes. But he patched us together, and honestly, when I was going through your background, I was like, "I need to talk to Luke. I need to have Luke on the show.

I need to maybe have Luke just in the back of my head at all times." You are someone who, I hope you don't mind me sort of characterizing you this way, but I see you as someone who is a creative who dived into AI pretty hard. And I think that's kind of a nice, you know, combination, not to mention with agencies like Four and other work experience, you've been dealing with some big brands on this topic.

And I'm like, "Okay, this... You know. You've, you've probably got some stories, and I wanna hear them." First of all, I just wanted to start by hitting you with a bit of a spicy question that my listeners would love to get your take on. But then I'd like to use that as a lens and just look at three things.

First, I wanted to talk about your vision for hybrid AI human collaboration and what it unlocks. Then I was hoping we could get a peek at the rough edges that come about when AI and human egos sort of collide. And lastly, I'd love to explore things that, like anyone, can start doing right away to make AI a bit more collaborative instead of just using it in isolation.

How does that sound to you?

Luke Alexander: Yeah. That sounds really good.

Galen Low: For a lot of folks in my community and the teams that they're working with, their LLMs have become their main collaborator. In fact, sometimes they would rather get the answer from AI than reach out to their colleagues, and that often makes sense when it comes to getting facts or gathering information.

But in situations where we're coming up with new ideas, most chatbot conversations I've seen feel like they're kinda living in a bit of an echo chamber. So I thought I'd ask you, as someone who is a creative person and an AI advocate, is our current application of AI making us less creative?

Luke Alexander: Yeah. So it's a great question, and you can, you can absolutely see why it's such a hot topic.

It is a spicy topic, right? So, on one hand, you know, let's go back a year or so when you had you know, people producing photos with people with six fingers in, all that kind of stuff, and thinking that they were like, you know, at the, at the forefront of creativity. It's, I think it's quite easy to look back at that time and go, you know, "Oh my God," you know, "AI bashed into creativity and created this sort of nuclear blast."

But actually, you know, the, the positive side for me has always been that what we saw there actually was people who don't have necessarily particular skills starting to be able to express their creativity in different ways, right? And there's an interesting, quite recent review that was done by a couple people at, at MIT, where they looked at some recent studies around creativity.

It's called The Hidden Cost of aiAI-Assisted Creativity, if you wanna look it up. It's really good. And they found ... They looked at these four studies, and they found that there's, there's quite a lot of evidence that AI sort of enhances individual creativity. So if you give people a creative task, and you let people talk through it with AI, you let people sort of brainstorm, you let people, perhaps people who aren't any good at sketching, start to sketch out their ideas and think visually where they haven't been able to do that before, you get some really interesting effects.

So that's definitely something I've seen. On the other hand, 100%, as soon as that leaves an individual person, as soon as it enters like a workplace or a project where creativity is not there for its own sake, but it's there to act as an input or act as a part of a project or deliverable, you get some horrible effects.

So you get the work slop. Work slop is, you know, the, the ... I'm sure all your listeners will have, will have heard this term. So it's when you're like, "Hi, I need one idea for a brainstorm," or one idea for a, a, a project, creative project of some kind. And instead of that person giving you one idea, they're like, "Well, here's 700 that I've generated with my AI," and "Here's 20 pages for each one that are like visuals."

That's lovely. In, in the previous work, we'd be like, "Wow, you put so much effort in." What it actually shows us, I think in this case, is people are just not spending the time and applying the creative filters that you need to apply, right? So whatever your creative management process is, most people, if they're managing creative processes, will have some kind of filter built in, whether it's like the double diamond or, or something like that, right, framework.

That goes out the window just because there's so much velocity. There's so much ability to generate. So there's that. There's also the thing that you alluded to earlier, which we- I've seen, as you say, in my previous life as an AI consultant, when I've gone into organizations, I've seen this a lot. The most senior people often are prone to this, which is a kind of AI psychosis, because they're talking to an AI.

The AI is feeding back, "This is amazing. This is clever. This is incredible. Keep going. Keep going. Keep going." And they're going down rabbit holes, and I think they know individually that they're not doing their best work necessarily. But it's, it's addictive. And what happens then when they bring it to a group is that sense of unease, the unhappiness they feel that they are...

They know they're not doing it properly. They bring it to a group, and then they become super defensive. So you get this terrible effect where people take something away, do a job, and actually go off on a real tangent, maybe do something they're only 50% sure of. But when they come back to the group, they're so keen to make people know that this is still their work that they overcompensate.

So you get this terrible effect where, you know, unless you are building spaces for collaboration that actually take into account the way people work with AI, all you end up doing is, is breaking teams. And I have seen, sadly, you know, a couple of occasions, genuine occasions, teams that worked really, really well together that have been like, it's like a divorce.

Like the AI has broken their creative process. And so that's why, you know, I'm going to talk about it maybe that's the sort of what we're trying to solve with, with Amble, which is the product that we're building is to just reset that and say "We need to start thinking in different ways of collaborating and this, this can work well."

And just on this last point, I know I've gone on for ages, but on the last point, another study which is, it's called the Cybernetic Teammate, and I, I think a lot of people have s- have seen it, so it's this one was I think a Harvard Business School survey. This was they, they basically went to Procter & Gamble and said, "Can we study your teams?

And we'll study some people individually without AI, some people individually with AI, some teams without AI, and some teams with AI." And they found that the individual performance, so individuals with AI suddenly bumped up on like knowledge tasks, so some creative, but mostly just sort of knowledge tasks.

And they found that the teams with AI elements, they were the most variable, and they were the ones that, that you saw the, the biggest improvements at like the top 10% tier of quality of work. So there's definitely some evidence here that if you get it right, where you've got teams working with AI in the right way, you can produce like some, you know, work where the it's greater than sum of its parts, basically.

It's just a question of how you do it, and you have to do it in a quite considered, deliberate way.

Galen Low: I really like that. Honestly, there's so much I wanna unpack there. The first thing, though, is, that I love is like that sort of theme of democratization is, is sort of coming up and, and the reasons why we need to put some effort into it.

Like I, I really like this notion that, you know, now you have the opportunity to be, to be creative. And from a democratization lens, you know, you see this in sports, right? You're like, "Okay, well, you know, folks who got good at hockey first had to be able to afford hockey gear." You know what I mean?

And so like there's this sort of, "Okay, well, now we can all play hockey," sort of thing. And then the, the, the natural thing, but we're not all good at hockey immediately. So we need to still figure it out. You know, we still need to develop, you know, our, our taste and our ability to curate what comes out when we can suddenly play hockey, you know, at the NHL level or whatever, right?

It's it's a whole different thing.

Luke Alexander: You're totally right. And for me, so the doctorization, I'm, I'm so passionate about AI as an enabling technology. And you know, I have, like a lot of people, I suspect listening to this podcast, if you're in a creative space, you have probably some people around you who really, really do not like the idea of AI.

So I always have to be like very aware of that. And, and my big, my big piece, and I really genuinely believe this, is that it changes, it levels the playing field in a lot of different ways. But, but... And the hockey analogy is perfect. The thing that I always look out for is it's not whether someone was necessarily gonna be great at a particular thing, but it's the point in the process where someone says "Let me show you what I mean."

And that point is where if you are someone who come, is very naturally is good at drawing, good at putting this together, good at putting words together, you have this ability to basically persuade people that your idea is better than it is. Kind of, right? You're, you're a very persuasive person, or you can draw beautifully, and so the s- the scamps look amazing, and maybe it wasn't the best creative idea, but it was, like, the most beautifully expressed one.

And, and that's fine, and that's good, and there's a place for that, and, and part, that is part of creativity, and I'm not knocking that. But it means that someone who couldn't, who couldn't do that, who was sitting there with an amazing idea but couldn't put the words together. Or maybe we're talking about someone who's like neurodivergent in some way, someone who's, is, you know, speaking a second language, like barriers of all kinds.

That's the point where you can say, "Look, this levels the playing field." It means that if you get 10 people in the room and you give it a chance, you're more likely overall to get a great idea out of that room.

Galen Low: And I think that's the optimistic opportunity here, taking your other point, that we still need to get there, right?

We still need to get to this point where, you know, someone using AI doesn't, doesn't make them hyper defensive, doesn't make their, you know, team get into like a divorce situation you know, isn't turning people off and is a way that we can actually collaborate instead of this like murky I don't know, if I say I used AI, then people are gonna think it's bad and then I need to defend it.

Or people love that I have 700 ideas, but you know, it, we're, we're not further ahead. This is just more stuff. And I honestly, I see it a lot even where I'm working right now, right? Where there is tension. There's tension around the like you know, even, even the people who have trained their LLMs to sound like them, to think like them, you know, they're putting information into it.

They're, they're putting in a lot of energy to make sure that their ideas can come out faithfully through an LLM, but then everyone looks at it and goes, "Well, that looks like, you know... This person would never write full paragraphs, and now I'm getting full paragraphs that sound like them, but it's clearly AI, so now I'm annoyed because now I feel like this person didn't wanna take the time to give me feedback or to write their feedback to me, and I'm insulted."

Meanwhile, the thing you're saying, which is maybe that person wasn't really good at formulating their feedback, and now this is helping. This could actually be good. Not just, you know, not sort of getting past that sort of sting that, "Oh, they just threw my deliverable into an LLM and said, 'I don't know.

What do you think?' And it came back, and then they copied and pasted it and sent it," is how it feels. But when you think about the, the level of effort and, you know, have some empathy for the person who is doing this, yeah, maybe they're really time poor. And it's like you got some feedback instead of none.

Or maybe they're just, you know, really self-conscious about the way they communicate and they feel like this is better. Even if it's not quite them, they were like, "This is an improvement. People should like this." And either way, the, the intention there is actually, it's quite good. It's like I do care enough to provide feedback, and I wanna get it in a, in a format that people can understand in action.

And me saying, "Make better, please. Make it pop," is probably, you know, not what they wanna hear either. So I, I think that's it's such an interesting stage and like of the, the human dynamic because of AI, not the AI dynamic necessarily, which, I mean, maybe is a good segue. You brought up Amble and this is clearly something you are, you're very passionate about, like taking the AI experience from the sort of like localized individual thing that supercharges the individual, like you said in the sort of HBR report- And taking it into sort of a team level where teams are working together, AI's in the room and no one's, you know, feeling necessarily self-conscious about using AI 'cause it's there.

It's standing there. It's, it's helping out. You touched on this a bit, but I'm just wondering what was the thing that took you down this path to build an entire platform that is, you know, collaborative AI, and what's different about it you know, compared to having our own sort of private conversations with AI and now having, I guess, kind of like, you know, a group chat with AI, a meeting with AI?

Luke Alexander: We're building Ample, so, the company that we, we started a few months ago, and, and we is me and a couple of, of, of old friends, so people I've worked with for 20 years or so. The more we talked, we realized we were seeing the same thing, and it, it is this sort of AI psychosis thing, this idea that people who are really good are kind of falling into bad habits, but also the, the democratization that we thought would happen in the workplace where everyone gets to contribute maybe a little more wasn't happening.

So we started to look at that, and we thought, "Well, what's, what's a, a way that would work for us?" And we, we played around with a few things, actually, and one of the things we found really powerful was, was putting the AI kind of in the middle, so putting it where we could s- all see it.

If you can see some- If you sat next to someone, and I used to do this quite a lot, if you sat next to someone and watched them use a normal you know, ChatGPT or Claude or something, you can quite quickly see that the people who are very good at their jobs use it in a more advanced way.

I mean, that's obvious. Okay, so your prompt writing, the things you ask about and care about are gonna make a big difference. And that's why, you know, in, in software engineering AI has got so good at coding. Writing the code is to some extent, this, this will get me in trouble, to some extent a solved problem.

You know? You know, it, it, it is no longer the case that you need to sit there and hand write these lines of code. But the difference between someone who's never, ever, ever coded and someone who's spent 20 years building software, they will just write totally different things. They, they are wr- They are essentially using the same tool in two s- such two completely different ways, they might as well be different tools.

And that's also true of creativity. It's also true of, of well, lots of places where people use AI. So if you put AI in the middle, you get the benefits of that because you've got people who can use... who bring different things to the table, and they are collaborating together. They are basically, you know, this, this is oversimplifying it, but they're kind of co-writing prompts or they're, one of them says, "Here's something I want you to do," and the other person goes, "Oh, actually, hold on a second.

Make sure you do this," whatever. That automatically raises the bar. But also because everyone's seen it happen no one gets defensive because everyone saw them use the AI and everyone's happy with that because that's, that's what it's there for. And equally, in some cases, you might want to say, "I don't think the AI's appropriate for this."

Like we're writing, I wanna write something in my voice. I don't think the AI's appropriate for this. And that's a decision that you can state in a really useful, collaborative way, rather than just like being, being, being obstinate, you know. So that definitely formed it. I think when we started looking at it as well, we were seeing this tendency towards like meat proxying, which is a horrible term, sorry, but the idea that people just sort of like act as, as sort of messengers between LLMs, and we're like, "That is silly."

You know, that is that's a really silly way to build organizations where you've got this sort of wall of, you know, of human, and all it does is pass things forward. And you can get rid of that just by not needing it because you're putting the AI in the middle. So there's that. We then thought, okay, what...

Well, that, that's great. That's a starting point. But does that solve it? And no, it doesn't. And there's a few things we then went on to think about, and one of those things is it needs to be proactive. It needs to know what the project is. It needs to know things. It needs to know what everyone should know, and it needs to be able to say, "Hold on a second.

You could spend the next six months learning how to use AI for this project, or I can just give you some things that might be useful." So you've done the last three projects. You always leave writing the web copy to three days before it's needed.

This has always caused a problem with knock on blah, blah, blah, blah. How about I do a draft, I, the AI, do you a draft now six months before it's needed, and you can just feed it in as you go. So rather than you having to be like, "I'm gonna use AI for this," it's able to say, "Hold on, I'm... I know all this stuff about project management and all this stuff about how to build the things you're trying to build or do the things you're trying to do.

I'm not gonna run this for you, and that's not the right route, but I can absolutely give you some, some helpful suggestions during the process." So we looked at that. And then the other part of it is just, you know, that kind of coding the, the democratization of coding experience, which, which we, we really loved, and we're like, "Is that something useful?"

We're like, "It definitely is something useful." And I can talk about a little bit. This is not super relevant, but being able to generate interfaces that teams can use without them having to buy software, suddenly that totally boosted everyone because you can have an interface where people don't have to learn how to use, say, Excel or whatever it is, or Photoshop, but you can just boil it down to what the team needs, put that in front of everyone, and then it becomes collaborative.

'Cause again, you're breaking down those boundaries and you're getting, you're democratizing the process.

Galen Low: I love that. That's so cool, and it's blowing my mind. And, you know, I, I, I'm not gonna lie, when you started, I was picturing this like, you know, the Polycom in the conference room in the middle, right? You know, AI's in the middle, it's just standing on the table, and it will pipe up sometimes.

You know, we'll ask it things, but you know, fundamentally it is, it is your... It's just a tool w- that's listening. And then as you went, I was like, "Wait a minute, this is, this is different." A, I really like that sort of pair programming mindset. It's like, well, if we are prompting together or asking together or, you know, creating together, we're learning from one another, e- each other as humans.

And also, you know, taking away this sort of layer of trying to have to figure out AI alone, I guess, is the other thing. I, I, I had this thing in my head where I'm like, oh, y- you know, sometimes I say, I'm like, "AI is, you know, the big thing right now, but eventually it'll be as commonplace as typing."

And then I'm like Oh, what if I'm a slow typist in the room? Am I-- What if I'm, like, showing that I'm bad at prompting as, as feeding into that psychosis. But I, I won't go there right now. But I- what I really like is that notion that it's it's not even just like a teammate necessarily.

It's definitely not just a tool. But the interface thing you said, you know, it blows my mind because, you know, how many times have I spent 15 minutes looking through the menus of whatever.

I do some light video editing. I use DaVinci Resolve which is could do much heavier duty editing, and I'm looking for just that one feature to whatever, crop a circle, right? You know, I'm like, gosh, if only this interface was just dumbed down to what I need, then I would just get on with it. And I, I love that idea that that could be that thing. I was just gonna ask you if you could walk us through what the experience is like of having AI participate in a collaborative session with the team.

You painted a picture of how it could sort of, be helpful or intervene but I'm just wondering, you know, what's a scenario where it can help the ideation flow instead of I don't know, getting, getting in the way.

Luke Alexander: It... And it can get in the way. Let-- You know, just be, be, be honest about it as well.

You know, there's no part of this which is let's put AI in there and it will just magically make things better. And I think it's interesting, if you look at things like OpenClaw and Hermes and Claude Tag, which are sort of... They're, they are a similar principle. Here's an AI that we can put somewhere. In raw forms, they can actually really...

Well, nothing worse than they can distract people. And particularly, you know, when we did some initial tests with, I'll talk about brainstorms because it's quite an interesting one. But if, if you, if you do a brainstorm and you say, "Let's bring an AI into a brainstorm," and you just say "Come be a brainstorm partner," what happens when people start sort of using it is they sort of defer to it, which is just what you don't want in a brainstorm, right?

And part of that is the automation bias part, is the people sort of, you know, demurring to the, the AI. But also it's, it's just bad user experience. AI has still, when it's voice-based, a kind of a lag, and there's always a sort of... You'll see this when you u- anyone using AI, they're like, "Am I using it right?

Is this correct? Is this... Am I doing it? Oh, oh wait. Oh, is it, is it gonna... Has it got it?" You know, all that kind of stuff. So anyway, so let me take you through, like, how it could work then. So, we very early on we were like, "Okay let's... It's all about eliminating friction, colla-collaboration," blah, blah, blah.

And we started talking about this and we went, "Hold on a second. That's just wrong, isn't it?" It's not actually about c- eliminating friction. There's loads of friction you do wanna remove you know, finding a time in people's diaries, sure, whatever. But actually a lot of creativity comes from friction of some kind.

People disagreeing and then the disagreement leads to a really good idea, or people getting something wrong and the wrong thing turns out to be actually what, what you want in the end. All of these things are like, you know... The, the, the sort of using AI just to smooth things, so coordination friction, it turned out to be the wrong route.

So we were like, "Okay, when we go through collaborating, we need to make sure that when it's, when it needs to reduce friction, it can, and when it needs to up the friction or turn the friction into something useful, it can do that as well." So let's take brainstorms as a good example of this. So we built a few brainstorming tools.

We did some experiments and we discovered loads. We discovered, like I mentioned earlier, that people like defer to it we discovered that people generally have bad ideas what brainstorms are and what they're used for. So there's kind of four like learnings, I guess, that we came from. One was like, you know, there's plenty of evidence anyway that the brainstorms are not great.

And actually let people come up with their own ideas separately creates better results, but then coming together to share them. So we're like, okay, let's make it so that if the AI is seeing what everyone's doing for the last few weeks or months, whatever, it should be able to pull some good ideas out of that and put them in the brainstorm to get the conversation started.

And then we were like, okay, no bad ideas in a brainstorm. Let's interrogate that. Well, it turns out actually that's a really bad way to run a brainstorm because all you do is you get people just agreeing at the start, and it's really interesting. I was at a talk, a lady called Cathy Peach at Nesta in, in the UK, and she was saying that when she runs...

She's about the collaborative intelligence. When she runs sessions, she asks room-splitting questions at the beginning. And those are questions where she, she knows that the room will split. And basically what they do is they ask these questions. If they get consensus, they just get rid of the question.

I'm oversimplifying. Sorry, Cathy, if you're listening. But they got get rid of, because there's no point asking a room of people to agree on something. You only really want to ask them where you're gonna get diversity of thought. So getting to a point where you're splitting the room is quite important. So that was another idea, like getting the bad ideas out.

And one of the ways that we came up with for AI to help with that is we already know a lot about the project. We know about the audiences for the project. So if you're designing a label for a, a new oat milk or something, you know who your audiences are. So why not bring them in? Bring them in as like digital twins or personas, whatever, and have them there just gently voting on ideas.

So if all three of your personas don't love what you're talking about, you know, the AI doesn't shut you down or press a big buzzer, but you do... You can see like it's getting down votes. And because it's the AI doing it, you don't take it as personally. And that's a, a big finding that we found is that if you get just even the three of us talking about what to name our company, things like that, actually having the AI as a, not a judge, but as a sort of a, a facilitator to keep track of things means that you end up offending each other less basically.

Or not offending each other. It's fine to offend each other, but you end up being able to sort of de-risk the conversation. So this, this, this whole idea of like psychological safety is something that we've been trying to harness basically. And it's interesting, there's something called the Habermas Machine, which i- if you've not seen it, you should have a look at it.

So it's just a, an experiment that some people put together, which is basically when you're doing negotiations or you've got like a contentious concept, rather than have someone facilitate, the AI sits there and kind of rewrites the same statement, you know, to allow for all the different people in the room, and it's much more effective at producing a consensus than a human is.

It, it's just human nature, I guess, just psychological, but, but that's a good way that you can use the kind of AI psychologist idea to, to, to get there basically. So there's loads, I mean- There, there's a whole bunch of things. So the, the brainstorming tool we built builds on all of these and, and you join it.

As you chat, it just, it doesn't ever talk to you. It just puts Post-its up as you say things, puts them into categories, does what a, like a, a someone leading the, the, the process would have to do normally. Keeps everyone focused on the actual task. Maybe comes up and says that you've only got three ideas, gotta move on, that sort of thing.

And that sort of facilitating approach has been r- really effective and really fun as well. People love it.

Galen Low: I like that. And it's, you know, it is the sort of paradox of a human facilitator, right? You're trying to play that role of someone neutral and, you know, quasi objective and making people feel heard, but also, you know, making sure that not every, you know, idea makes it to the final cut.

And it, it is, it's political. It's so funny when you're talking about these sort of layers of... I'm gonna frame it as layers of psychosis, right? We're talking about the psychosis of okay, well, am I using AI well enough? The psychosis of, well, AI just told me that these are great ideas, all 700 of them.

I, you know, it said it in such a way that maybe I should believe them. There's the like, oh, us working together and, you know, trying not to, you know, hurt each other's feelings in an unproductive way but still introduce enough friction that we're not just nodding at everything and that, you know, we're not making decisions.

And I really like that idea that, okay, well, actually, you know, even scientifically so far or at least the, the research is showing and your research is showing that in, as a fac- as a facilitator role, you know, AI can be quite effective at just kind of keeping us on task and formulating our ideas without necessarily removing friction, and ideally, without necessarily adding more friction than necessary Outside of our own sort of individual or cultural psychosis and attitudes towards AI.

Luke Alexander: Yeah, I think that's it exactly. I mean, and worth saying, I think, you know, there is very much still a role for human facilitators in this process. But the reality is in most day-to-day workplaces, you don't have trained facilitators who are gonna be effective at moving this stuff on. So it's another, it's another way you can kind of bring some of the benefits of that, I guess, using technology rather than having to re-engineer your whole processes in a way that will never really happen.

Galen Low: Yeah, and I, I love that for my audience as well because, you know, in my community of, you know, digital project managers I've been saying and others have been saying, we're like, "Well, you know, facilitation skills are gonna be really important right now." You know? You know, instead of the things that are getting automated this is a very human thing that we can do that helps drive towards goals.

It is in some ways management of a, of a project. It's leading us through. But we're meant to be q- quasi-neutral as well, right? We're not like shot callers necessarily at the project level. We've got stakeholders in the room, and we should be good at that. But I also like the flip of but if someone's taking care of that, then we can be contributing in a completely different way.

And I think that's true of anyone in any craft. If you're, if you're not taking notes, you can participate. And then the question is, okay, well, what are you bringing to the table when you're participating if you're not taking notes? And I think that's it's an interesting question that gets into how a, a team collaborates which allows everyone to be more present.

I think it's, yeah, I think it's, I think it's fantastic. You had mentioned in its current form, at least your brainstorm sort of, platform, the AI doesn't speak. Earlier we were talking about maybe it will come in with suggestions and, you know. And you had mentioned, right, quite rightly that the tone, right, the, the sort of ethos of a lot of these LLMs is convincing or authoritative enough that we kind of defer to it.

We're a bit demure. We're like, "Okay, well, you clearly have trained on everything available on the internet, and I have only read, you know, seven books in the last, you know, as many years. So I guess AI, you're right." But I'm wondering as you've been sort of creating these experiences and, and when AI is talking out loud has this introduced, like, more friction?

As you've been building this out, has anyone sort of like gotten in a fight with AI? And if so, like, how did it go down? And, and who won?

Luke Alexander: I mean, 100%, yeah. So we have so for the brainstorm tool in its current format that we're, that we're playing with, the AI doesn't speak for those other reason.

We have a separate, a different tool that we, within our product that we've been working on, which is all around briefings. So your audience will know full well the importance of a, a good briefing, right, in a creative process. So we, we thought, "Okay, this is really important. We've, we've worked in these industries for years and years and years.

We know how important briefing is. People often get it wrong. Let's just have the AI take your briefing." So we set it up and we said, "Okay, you're gonna chat to a person, you're gonna ask them these questions, and while you're doing it, you're gonna draw a mind map so they can easily see where you've got to, and they can correct you."

So that kind of worked okay But oh my gosh, the AI did, there was something about the way it was asking that was just incredibly annoying. So when any... And it, and I, I don't know what it is. It wasn't being aggressive, obviously, 'cause it's AI. Maybe it was being too unaggressive, but it was, it was sort of fi- fixating on the points that didn't matter, and it was going, "Okay, what about this person?

What are they doing?" You're like, "Well, I don't, I don't, I don't wanna tell you about that." People felt they were being sort of led by it, which it should be the opposite way around, right? And you, I mean, you'll know, right? You, you know, the, the skill of taking good briefing is a skill. So yeah, we did have a few occasions on that, on that front where people were, you know, not sort of shouting at it, but getting pretty, pretty pissed off with the AI just sort of wasting their time, and I think that's the main thing.

People are so quick to move from, you know, the first two or three seconds using any of these tools if people haven't encountered them before, they're like, "This is magic." But it really doesn't take long before they're starting to go like, "Well, why, why can't it do this? Why is it, why is it asking me this three times?

Why doesn't it, why does it keep misspelling this person's name?" Whatever it might be. So yeah, no, we definitely have had that, that situation. And it, you know, it's interesting to watch people use, particularly our tool, but, but obviously other tools. And when I was, when I was doing implementation, you would have people who...

And actually I used to sit, so in one of the offices I used to work with, work in, I used to sit at a desk here, and three desks away there was a, a CFO and three desks away there was one of one of the directors of this company. Catherine and Patrick hopefully they won't mind me using their real names.

But Catherine was all about being nice to the AI. So she would always say, "Oh, thank you. Okay." And she would have a full-on conversation with it before she asked her first question. She would say, "How are you doing? How's it going? What are you..." You know, and discuss things with it. And then she would say, "All right.

The thing I'd like to talk to you about today, copilot, is this." Patrick would be the entire opposite. He'd be like, "You stupid machine. You've got this wrong again." And I was sitting there. I was like, "This is a really good test. This is a really interesting case," 'cause who's getting the best answers? And actually, interestingly, it varied.

There were definitely points where a really sort of quite aggressive approach meant you were being direct, you were being clear, you were stating your point, you were expressing why you were in control, and that worked really well. There were some points when being friendly and nice and, and, and gentle with it led to a more human response for copywriting.

So it's actually interesting. I think there actually, oddly enough, may be an argument for occasionally getting into a fight with your AI. That's not necessarily a bad thing, if you know what I mean.

Galen Low: I think that's f- I mean, honestly, that's fair as, as I think about it. So I mean my, my biggest corner turn as I was sort of starting to use AI, like LLMs specifically in my day-to-day, was like, "Okay, just pretend you're in Slack chatting with your teammate."

Not, I don't, you know, not to the extent that Catherine was you know, having a conversation you know, I, I wouldn't ask it about the weather. But I would sort of address it like a teammate and like any, you know, good teammate. The collaboration element of being a human, right, is, is up and down, right?

It's not like we're always gonna have good conversations where we're pleasant to one another, and sometimes I do need to be more direct with one of my employees, right? To, to kind of make sure I'm, I'm, you know, they're, they're staying on, on course, right? Not getting distracted. And these things that we do to sort of adjust the way we work together.

And yeah, they... You know, in, in reality it's not always one or the other. Definitely people have styles, right? Someone might have a more direct style, but doesn't mean that every conversation is going to feel like you're being yelled at. As I say that I'm like, "Okay, I know a few people who, you know, maybe that's not so true."

It always feels like you're getting yelled at by them. But you know, I think there's nuance to the human experience and, and how we experience other humans that, yeah, sometimes don't show up in the same way when we're training AI. Because AI's like, "Okay, I'm learning about you and I've put you in a box.

I think I've figured you out now, and I'm only gonna give you that." But i- with the human experience like it, it might be the case. Yes, I've seen, you know, what you said briefing is an art, and yes, I've definitely seen that with humans and AI. I've gotten frustrated with AI because it's like asking me these verbose questions, like going down this other rabbit hole that I don't wanna go down right now, like you said.

And also I've seen humans do that too, right? And they're like, a, a, a good human will be able to read the room and be like, "Okay, I'm going the wrong direction. We'll circle back." You know, that's what the parking lot is for. All those other, you know, semantic things we say at work to be like, "Ooh, okay, well, we'll ta- we'll talk about that later because clearly you don't wanna talk about it now."

And it's interesting that you know, we think of it as so binary for AI. Not getting down the rabbit hole with the anthropic, you know, is AI conscious and does it have feelings thing, but yeah, it's like it's, it's also nuanced.

Luke Alexander: That, that's really interesting. That's a really, really good point and I, I like that.

Being able to take a breath is something humans can do and, and AI can't as well. So if you go to a colleague and you say "I think this," and they disagree, you can both take a step back, reevaluate your thoughts and come back and, and have a... You can agree with each other or disagree or whatever or come up with a third route.

The AI doesn't do that, but it acts like it might. So I think it can be quite dangerous I guess to have a situation where you've got you know, a conversation with an AI where you're o- you're overly convinced it's a human. It's very interesting. I think something that's come up a lot with the latest anthropic models, which are super sophisticated obviously, is that previously if you sort of said, particularly when you're coding, and I've had this.

When you're coding and you say, "No, I don't wanna do that. Let's take a different route and we'll do this instead," less sophisticated models would say, "Okay, this is fine. I take this. We will then do what you've said." Latest kind of, particularly the, the latest papal and, and 8+ models, I know plenty of people who have had the situation where they, they've sort of said, "No, we're not gonna do this," and the AI's kind of- Agreed but done it anyway, or agreed but said, "You're kind of wrong, so I'm gonna, I'm gonna take a third route that we haven't agreed."

I think the, the model of treat it like a, used to be treat it like a, an intern or treat it like a, a colleague who does, just onboarding onto a project, whatever that is, I don't think that works as well as it used to anymore. I think we need to develop a new way of talking to our AIs and, and, and understanding what kind of things they're gonna listen to and what kind of things they need more convincing about.

Galen Low: It's a hugely good point that this is a moving target, right? We, we... It takes us so long to be like, "This is the way we should use AI," and by the time we get there, the models are, like, seven steps ahead and it doesn't apply anymore, and we have to kind of, you know... In the process of us, you know, we're talking about this sort of psychosis, right?

And everyone kind of needs to, to get there culturally you know, within their team environment of how we actually feel with AI, about AI, and how we work with it. Meanwhile, it's still moving forward, right? So by the time we all agree, and we're like, "Yeah, okay, this is how we're gonna use it. We've written it down.

We've got our sort of, you know, our, our model or our charter." Oh yeah, it's, it's, it's already different. And we need, we need to have that in our minds that it's, you know, it is going to continue to be different. But also, you know, to your point arguably getting more human characteristics.

Like, how many times have, have I asked someone to do something, we're like, "No, no, no, we're gonna go this direction," and they're like, "Okay," and then they'll do the, the third option that we hadn't talked about because, you know, they have conviction. They think it's the right thing to do. And so might our tools now, and it's this interesting quandary where it's like, is it a tool?

Is it a teammate? Is it an intern? Is it an expert? Is it, you know... And then it, it, it's very easy to fall into that hole of "Okay, well, I guess we better just believe it because, you know, maybe it knows better."

Luke Alexander: It, it might do. It might know better. That's the thing. You know, it, in... Again, in software engineering, which I think is where it's, it's most obvious at the moment, a lot of software engineering is about Understanding trade-offs between capability and risk.

Should I open this surface up it, which enables all these capabilities, but obviously gives it another surface for attacks or whatever it might be. Your AI will, will, will advise you, and you see this with a lot of people who don't normally develop software starting to vibe code. The AI will, will talk them into this incredibly sophisticated security setup, for example, which for something that may never leave their computer.

And on one hand, that's clearly ridiculous, blah, blah, blah, and it wastes huge of time. On the other hand, the moment they do then want to put it live in some form, the AI was right. So it's very hard to you, you don't wanna totally close yourself off and say, "We're in charge." Because one of the benefits to these systems, they do have all this embodied knowledge, and that's genuinely useful for people.

Galen Low: I honestly, I think it's like the context judgment, I guess. I don't know how else to phrase it, but, you know, and in some cases I'm like, "Yeah, yeah, AI, but like you don't know all the context. Like I know this person. I've worked with them for 10 years. You know, I'm not gonna train you on that. It's gonna take too long."

I'm like, "This is, this is just, you know, from my context, I, I think I have more context on this to make the call than you do," versus, "Yeah, I'm not a cybersecurity expert, so actually, you know what? That's probably fair." And we do this with humans too, right? It's like, okay, well, you know, listen, like like you and I have probably got it a lot, right?

"Oh, you're the creative. You're just the creative. You're just the project manager. You don't have all the context here." And, and maybe we don't, right? And again, we have to sort of be able to sort of, you know, take a step back as well based on how much knowledge and, and, and the quality of the judgment of the person who is making the call.

We won't always agree with it and it could still go sideways, but like that's, I think how at least how I look at it is okay, well yeah, you've, you've seen this a lot and yeah, I trust your judgment because you've seen this a lot. And AI has seen a lot because it's trained on a lot. But you know, it's trained on a lot of information.

So security is a great one versus stakeholder politics. They might not be able to make that call. But I, I mean, that all being said, I do love that idea of like having, you know, this sort of synthetic audience or like a, you know, the avatars of people who, you know, are stakeholders and what they might say, 'cause I think that's that's just good customer experience, you know?

Like user-led thinking, you know, not just my idea's great and who cares what the end recipient of this, you know, cares about. But it's it, yeah, it's a really interesting one to sort of navigate that. I, I wonder, I mean- This is sort of related, but I did wanna get in there because we were talking about AI sort of facilitating, it's a good facilitator, and we were talking about, well, sometimes a human might still facilitate.

And it stands to reason that if a human's still facilitating, sometimes AI will be a participant. And I was curious, you know, as you've been sort of designing this i- is it different? Do, do facilitators need different skills? Do human facilitators need different skills to facilitate a session where AI is a participant?

Luke Alexander: Yeah, I think so. Well, I mean, just the most obvious one, and I, I, I'd be very interested to know whether, you know, you and your, your listeners have encountered this. Any situation now where people are sort of in a group session where they have access to their laptops, nothing else, people are a lot more distracted by the fact that they can do things in the room or ask questions in the room.

I mean, people always used to sit there Googling stuff, that's fine. But I think at the point where people are going off and really engaging another participant in the room who, who isn't physically there, where the AI is like, "And I jump to them," where you end up with the defensiveness what I talked about before, where people are swearing blind that they're, you know, it's their idea and it isn't.

It's like a, something that Claude's come up with. The facilitator's job there is really hard 'cause you, A, you don't want to diminish anyone. B, there might be some politics in the room where you can't speak against people who are more senior or whatever. But you do want to try and try and get across the idea that we're all here in this very valuable time, physically in one place.

Maybe you're all on a Zoom call, but it's all valuable time. Is... If there's something here that can be done outside of the, the room, that should be done outside the room, and what we're here, we're here to do is to have this conversation first and foremost. But also accepting sometimes there may be really useful points.

You know, we've had sessions recently where you start Claude off at the start of the session looking at some things, and halfway through it comes back and it changes the, the tone of the conversation, changes the... And that can be really, really valuable. It's just building it in and building in a sort of understanding that you have access to these intelligences but the intelligences are there to do particular things, I guess.

I think it's a really open, open problem. I mean, I've definitely seen more recently people sort of, sit in some kind of creative session, and really they'll start, like they'll have gone, they'll have heard the idea they like, and they will have then gone past it and started what would've been a process they couldn't start until they were like back at their desks or whatever.

And they'll be like, "All right, let's build it. Let's design it. Let's..." And that, that seems lovely, but it destroys the point of having a process, the point of having these gates, the point of having these abilities for people to feed in, and that's really powerful. We think we took a, a website project which is a classic example where you have Typically, traditionally, a kind of creative or designer or someone who's gonna care deeply about the user experience, whatever.

And then you've got developers who care deeply about how it's built. And the fact that those two people can now be the same to some extent, you can have someone who's never done any formal training in design or UX sit there and say, "Okay, so I think what we do is this," and you can have the UX design person say, "Okay, so I turn that into a website."

You very quickly get to the point where unless there is a very effective facilitator in that process, people are stepping on each other's toes, they're annoying each other, but also they're not, they're not taking, they're not taking advantage of having a team. At that point, just set those two people up on two different projects if they're that amazing, right?

So actually it's, it's making... It's part of, I think, a facilitator's approach in this new era is, is being champions for the human. So what, what bit of this human that I've physically got here is gonna be really valuable to this project? I mean, there's loads of industries that we're talking to at the moment where the apparent value of someone was the 300-page report they produced.

But of course, that was never their value. Their value was what they knew to go into the report, or the questions they asked or whatever. So at the point where you can say that you don't need to worry about this big chunk of work, and we can just have your Present, like your, your brain fully engaged in this for five minutes that can end up being more valuable, right?

So, I think, I think it's a really interesting time to, to be a, a project manager, but, but also just to be working in spaces with humans who are at very different stages of their use of AI and, and trying to navigate that, I think.

Galen Low: I love that. Actually, you just blew my mind because that idea that it's if we get people's competency to a certain level where they don't need to be working together on the project team, we can just put them onto other projects.

And I know some people listening are like, "Wait a minute, are you just, are you getting rid of jobs?" And I'm like, "Well, no, not really." We get... We're, we're... Like you said about the pair programming, you could cross-train. And I think you said earlier, it's when you're working with someone who's coded for a decade, and you're, like, new to coding, and you both have access to the same tool, but we're gonna be using it different ways.

But if someone was to say, "Oh, you know what I would do with that? I would ask this, I would include this detail, here's why." And then suddenly we're getting more of that context from somebody who's been, you know, thinking about coding or developing applications or back-end infrastructure for a decade, and we're getting ourselves to a certain level where, yeah, it might actually be, as a staffing model, I think that's really interesting to say okay, well, at a certain point this UX person could, you know, build the app on this other project, and this is something that we, we could explore, and they might enjoy it, and they might do a really good job.

Luke Alexander: Yeah. And you know, there's, for, for agencies in particular so for anyone who has a, you know, has clients, you know, you've always wanted to get the, the people who understand the client the best, those who are closest to the client, as close to the work as possible, and that's now possible without really having to...

That can be the same person. And I'm not saying get, fire developers at all quite the opposite. But like here's a good example. Most marketing agencies, and I work with a lot of marketing agencies, cannot for the life of them seem to manage building their own website. It takes them years and years and years.

They never get around to it. It's never because it's a particularly complex problem. If it was a client, they'd solve it immediately. It's because of a priority thing, but it's also about this kind of it's ours, I wanna make it perfect, I wanna use all the skills in the company. And actually, we did something with...

It wasn't an agency I worked at, but it was a suggestion I gave to another agency where I said, "Give the website projects," this was a couple of years ago, so when it was possible to start building websites. "Give it to the most junior employee and tell them that you're not gonna say anything mean about it when it comes.

That you just, you just wanna see what they can do." And you know what? They came up with a pretty decent three-page website for this agency that several... And it turns out it's not that hard after all. But you can definitely get in your own way. So actually some things there is a, a real benefit to having a unity of, of thought approach on it.

To come back to collaboration though, which brings it kind of full circle, that's where, that's actually a bit of an artifact of, like, where collaboration could go wrong in a pre-AI world. And that's when people are just like over-- they're overvaluing all the inputs and they're undervaluing the outputs of a process.

So again, something where AI can, can do a good job 'cause it can make the outputs free or very cheap So someone come along and go, "Well, let's just have a look what happens if we do build the agency website a bit like this," sketch it out, blah, blah, blah, something like this, and then you can see it rather than feeling like you've wasted everyone's time going around in circles, so.

Galen Low: Yeah, no, I love that. And especially, you know, as someone who I won't notice a typo in something until I've printed it, you know what I mean? It's once it's the real McCoy, then you can ha- really feed back on it. Even wireframes is a bit tough. You're like, "I guess so, kind of." In your head, you're like, "Well, we'll change that later.

I'm not gonna let that fly," right? Whereas if, if you can get f- you know, further faster to build a thing and feed back on that, then yeah, there's ... it's more of a conversation. That's that's really interesting. You know, the, the Amble experience seems amazing. You know, I'm really looking forward to getting my hands on it.

For teams who, you know, for whatever reason, aren't able to have a look at Amble or use it for their, you know, collaboration, what's something that they can do right away to start using AI a bit more collaboratively instead of privately?

Luke Alexander: So one thing I would say there is i- I think a lot of people are tempted to sort of do the sort of put the AI in Slack thing.

My advice generally isn't, and is until everyone who's kind of using it and on the same page understands it, actually, that doesn't seem to work. So I would say rather than take that route, try this exercise, which is something I do with quite a lot of people, is literally you get on a call and you share your screen.

No one is allowed to edit, say, a Word document or PowerPoint, whatever it might be, some kind of artifact that you're creating. All you can do is feed into the AI, and someone has to type it in. That's fine, but everyone, you can only prompt, because what you end up doing then is talking about the artifact as something that you all care about and want to do well, rather than getting annoyed that someone's typing slowly or getting annoyed that something's not phrased how you would phrase it, you know?

Once everyone agrees that it's just not a ... It, it ... And it can be a draft. You, you can say "No one has to care about this at the end. We can go revise it, and we'll give it to Bob or whoever is gonna take it away." But that kind of collaborative principle, you'll find that the conversations you have around it, which previously you wouldn't necessarily have been able to, if you think you do a half-hour call, there was no way you'd be able to write a 30-page document.

But nowadays you can have that being written, and you can just feed into, into ... It's, because it's on screen, you can see what's happening, what's changing, and you'll get people say, "Ah, you know what? That's not quite right. This risk register is it's missing these parts," or, you know, "We haven't really catered for what the client wanted here," or, "No one's mentioned security," or whatever it might be.

So I would start with that. I think that's a really easy way to get started with, with collaborative AI in its most useful case. I wouldn't go down the sort of, you know, trying to necessarily put Claude in the middle of things and, and brief it, because without the framework around it of the project, it will give you very weird results, I think, and you'll end up with one person driving, essentially.

while everyone else kind of sits back and goes, "Well, I don't really know what's going on."

Galen Low: I love that because you know, we wouldn't do that to a human either. I mean, maybe an intern or a co-op student at an agency, but you know, not on purpose, just out of necessity. But normally you'd bring someone in, you'd, you'd give them context around, you know, what their specialization is, and, you know, you'd onboard them, and then, you know, they'd- you'd feed them context as you go.

And what I like about that model is that instead of, you know, everyone's own way of using AI privately A, everyone's kind of getting on the same page. They're getting used to not holding the pen necessarily, and they're also sharing context that the AI is presumably learning from as well, having just been born yesterday.

Not really, but you know what I mean. Just showing up on a project, been like, "Cool," "you're gonna help us write this brief. But also I'm gonna take the pen from you every now and again, and I'm not gonna tell you why I made a change." It's never gonna actually figure it out.

Luke Alexander: The context point is so important, and it's interesting, like if you look at how humans interact when they're working on a shared project, the context negotiation part, the point where they're deciding what matters and what doesn't, always happens.

It's, it's crucial. People do it without realizing. They set the, you know, tone of the meeting or whatever. They say, "We're doing this, we're not doing that." They bring people up to speed. That's really natural, but they're very quick. Humans are very quick at doing this, really quick at identifying what in our vast pool of knowledge in our brains and what's happened in the last 30 years on this project is relevant to this exact moment.

So that's what-- that's a gift we can give the AIs, right? They're, they're bad at that. You know, if you just, you know, in Copilot, for example, which you know, Microsoft's AI tool it has the ability to look at everything on a project that you've ever had and any, any email you send and stuff. And most people say when they use Copilot that, that they spend so much of their time basically just telling it, "That's not relevant.

That's not relevant. That budget is not for this client," this, you know. So, so yeah, use all the human ability to, to do that context filtering up front.

Galen Low: Luke, thanks so much for this. Just for fun, do you have a question you wanna ask me?

Luke Alexander: Yeah, there is. I mean, so I'm fascinated by project management and particularly digital project management and how it's, it's changed, right?

So I've been on the sidelines of it. I've never been a project manager myself. One thing that I've always hit against, and it's something that really feeds into Ample and so I really would like an answer, is if you have a team that does not want to be managed, has AI helped with that at all? I'm thinking, for example, like Asana is one of the, you know, is, is seen as a really, really good experience, a beautifully built product.

I've seen it used in a couple agencies, but al- only ever seen people bounce off it, and it ends up being someone's job. So my question is, has anyone solved this? Has anyone solved the managing a team that doesn't want to be managed problem with AI?

Galen Low: I mean, arguably, we've just shifted the problem. I mean, I do- know that like the thing I see the most is, right, like on the value prop for a project manager would be like, "Hey, listen, just like a, like your agent is gonna do the follow-ups for you. You don't have to be the person who nags about time sheets or, you know, is this thing on track for, for Friday or what have you," which, you know, in, in, in its- in itself is like, it's more complex than that.

Sometimes when I'm checking in with somebody, I'm not just managing them, I'm also trying to get information back from them so I have more context to, you know, act in different ways and make decisions. But I have seen it take the edge off of that like, "Oh, my project manager is now nagging me. It's Wednesday, it's not due till Friday, leave me alone."

These like nudges that come through the system, I think are a- an interesting way to solve the problem. I, it's funny because like I was at an Agile conference a, a couple years back. You know, Scrum, for example, right? This is our methodology where we're gonna move fast and we don't need a project manager.

We're a self-managing team. We're just gonna get rid of that role, product owner, Scrum master, you know, development team, let's go. And even, even in that context there was this whole separate movement against that conference to be like, "Don't bring me to a conference. Leave me alone and let me do my work."

And it was this sort of oh, not only did the system where we didn't have a project manager not take away that problem of feeling managed and just wanting to get on with your work. You know, people are now like, "Okay, well I don't wanna show up at standup because that also doesn't let me work."

You know, at a certain level, I mean, like humans who don't wanna be managed don't wanna be managed. Like I don't know if there's a solution to that. And definitely I, I've seen it go the other way, right? Where it's like all the nags from the AI agents get ignored or actually are more annoying and some of the context isn't shared as a result.

So I would say it's not a problem that's been solved yet. And also I, I kind of wonder, you know, you, you mentioned about the friction in the creative process. I wonder if sometimes the friction in the being managed process, as someone who myself does not like to be managed, I'm trying to understand the value of it, and I think it is one of those things where, yeah, there's gonna be some friction because we don't all want the same things.

We don't all agree. We don't work the same way, and we have different priorities. So, you know, there's just gonna, gonna be this tension of sort of feeling nagged and feeling questioned and feeling, you know, under pressure that maybe is just part of the process, which is- Yeah ... yeah, maybe just giving my audience license to nag.

Luke Alexander: No, I love that answer though. That's really interesting, and it's a really... I was listening to a, a negotiation coach a while ago talking about how you deal with a room full of people who obviously like are never, essentially never gonna agree on a solution, and you, you find something else to agree on. So you find like you, you agree on the process instead, or you agree on what doesn't get included, and that, you know, produces an environment where people also feel like there's progress being made.

And I quite like the idea of like the unmanageable team. It's, it's finding something that you can all come together, even if that is how annoying it is to be managed or something. I don't know.

Galen Low: Yeah, totally. It's hard. I know I'm annoying, but we're gonna get through this. We can all agree on that. Luke, thanks so much for spending the time with me today.

I've had so much fun. I've learned a lot.

Luke Alexander: Yeah, me too.

Galen Low: Where can people learn more about you and Supervenient and Amble?

Luke Alexander: Yeah, so if you go to www.supervenient.ai that's our website, and you can find out more information. It's in a really ... it's in the MVP stage. It's on a waiting list. If anyone listens to this and would like to find out more, we are recruiting a very small number of alpha users.

You get to use it for free for a few months, give us feedback, tell us how useful it is. We pay for your tokens, if that's so you get some freebies there. And follow me on LinkedIn and, and find out more as it goes.

Galen Low: Luke, thanks again.

Luke Alexander: Thank you so much, Galen. Really good chat to you.

Galen Low: All right, folks. That's it for this episode of 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, create a free account with us at thedigitalprojectmanager.com. Until next time, thanks for listening.