전사 🔗
Anne: So welcome to episode twenty-two of Asynchronous and Unreliable, a weekly podcast where we discuss the latest ideas and concepts in tech. I'm your host Anne Currie co-author of Building a Green Software, the Cloud Native Attitude and author of the science fiction panopticon series. And today I have the great pleasure to welcome Matthew Skelton, co author of the award winning book Team Topologies And CEO Conflux, who wants to provide leaders with the levers and language to enhance decision making and create cultures that enable high performance without burnout. So welcome, Matthew.
Matthew Skelton: thanks very much.
Anne: it's really nice to see you. I mean obviously I've I'm a I'm a fan of multiple of your books. We met at a conference many, many years ago, which is where I met half the
Matthew Skelton: Long time ago.
Anne: people who are on this podcast and met at conferences. but I really love Team Topologies. But I just yesterday read internal tech conferences for the first time, and I really enjoyed that as well.
Matthew Skelton: Well, that book came out just before Team Topologies actually. for the for anyone watching the actual video feed, it's it's like a purple book just above my shoulder. anyone who's not watching the video feed, it's a purple book. and came published in twenty nineteen, I'm the co-author with Victoria Morgan Smith, who used to be at Financial Times in London. and we both had experiences of running internal conferences. well, we can maybe talk about that a little bit later, but but the way that Victoria and I both framed completely separately, but we we'd arrived at the same point and that's why we wrote the book, is there's a way of bringing together people for events like internal conferences in a way which is super, super valuable to the executive of the organization. We maybe didn't emphasize that so much in as much in that particular version of the book. There's a possibility for another version of the book coming out at some point soon where we're adding some some more stuff later. But it's an example of the kind of activity that I've been interested in for for not just interested in, but doing things in in this kind of space for like at least fifteen years, probably more than that, where we're thinking about organizational effectiveness, clarity of responsibility, bringing people together, Avoiding duplication, increasing alignment, a bunch of these kind things altogether. so it it's it fits in with a lot of the the themes that I've been focusing on for a long time. Even though the the most well known thing that I've done recently is is team topologies, which fair enough. but even that is really about alignment. Alignment, ways of working, effectiveness, clarity, responsibility boundaries, all this kind of thing. So it it comes from the same kind of place, if you like,
Anne: internal tech tech conferences studies kind of three case studies of really sophisticated internal tech conferences. One was the FT, which oddly enough, I also studied the FT because they they're they're one of the top companies in in adopting cloud native very very early
Matthew Skelton: Mm.
Anne: within their within their enterprise and i don't think that's a coincidence so oddly enough i i interviewed victoria your co-author for my book cloud native attitude and i also interviewed sarah wells who you work with quite a lot these days
Matthew Skelton: Sarah Wells, yeah. Mm-hmm. Good. Quite right. Because 'cause Sarah was also heavily involved in making the internal conferences at at Financial Times work really well. I know know, you've known her for a long time. and I have as well. And it's it's great that that you that you're able to acknowledge like her contribution to the Cloud Native success at at FT, because like that 100%. Like there there there was a there's a there is was whatever a group of a core group of people. Victoria's one and Sarah's another and there's a bunch of other people too. but yeah I I think I think the the comment you made there is is is very accurate. So the organization so if you think about the kind of cloud native period well
Anne: Mm.
Matthew Skelton: the cloud native period is never going to finish, right? We're all there's always going to be something that continuing but the the real the real act the real learning curve for a cloud native was what, 2008 to 2022? Two, twenty three, so in that for the for for the the early adopters, effectively. the one the organizations that really managed to adopt cloud native very effectively and rapidly were those that had a very clear approach to innovation, discovery, knowledge diffusion, internal alignment. And so it's not surprising that those organizations are the ones that many of those ran or still still do run internal conferences. Klarna, for example, is is one example, the the the the payment organization. they're one of the examples in the book, in in the internal tech conferences book. They are they're they're they're super, super ahead of the curve in terms of adoption of of new new approaches and ways of working and things. They their internal conferences have been running for ages. And that that raises a really important point. Now right so even with so with cloud native as you know, there was a real need for organizations to to deduplicate align adopt things very quickly, have a very a very in very intentional approach to to innovation and and diff and practices diffusion. with but you can sort of argue that cloud native is like only applicable to the IT department. I know it's it shouldn't be, but like you can sort of from the perspective of some execs then they don't really care about cloud because it's just an IT thing. Now with AI, it applies to the whole organization. So guess what? Suddenly now the entire organization, at least the entire knowledge work organization, needs approaches to deduplication, rapid discovery of innovation, sh sharing of practices, all the st all the stuff all
Load full transcript
the same stuff that these cloud native organizations have been running with and doing for a long time, for 15 years, 20 years, whatever. But now the the entire organization is feeling the need to do this. So I was talking recently to a CIO of a bank in the Middle East, and he said, Matthew. AI is forcing me to make my organization a learning organization. Now, the whole learning organization book came out, whatever it is, like 20 years ago or something. And like it's kind of weird, but you know, we'll take it that AI is forcing all every organization to to become a learning organization finally, because without it, you are gonna get so left behind and even just burn through so many tokens that that so mu so many, you know, LLM tokens that you won't achieve anything. So like It's forcing this this shift towards effective learning, discovery, standardization practices, things like this, in a way which is very engaging and very human centric, weirdly. At least that's the picture I'm painting. That's where I'd like things to get to.
Anne: Well, yeah, it's it's yes. And and you do need and and and internal conferences are, as you say, excellent for alignment and they're excellent for kind of engaging people in in getting involved and thinking and learning new things. I mean, something I don't know, you might not have realized the the the other company that you didn't mention that you did a case study on, MetaSwitch and Networks, although they they they're just bit next they got bought by Microsoft. I worked there for 10 years. and hard
Matthew Skelton: Did did did you experience some of the conferences then that that were ta talked about in the book?
Anne: Well yes, 'cause oddly enough I am obliquely mentioned in your book because I was a friend of the company that came in and spoke at the conferences.
Matthew Skelton: I didn't realise that, right? I'm looking through now. I need to go and read. Well, literally, I've got the book in my hand. I I will have a look later. I I I just hadn't clocked I hadn't clocked that at all that you that you were involved. Which is really cool, right? So you can actually speak from first hand about the the the kind of I guess the kind of practices and approaches that then make that that that make an internal conference actually valuable at the at the whole organizational level. Because it's it's not just about running internal conference, right? It it's it's what's the intention behind it. So when
Anne: Yeah.
Matthew Skelton: I, for example, when I ran or I I was le I guess I was leading the the the these conferences at a c a company based in London. This is back in like twenty twelve-ish. we had a deliberate approach where we wanted it was kind of technology centric but outward looking. So at looking out to the rest of the organization. So for example, one of the talks that we kind of curated was the head of the database team and the head of I think I think the sales team talking about like the Oracle database upgrade. Like on paper, that should be the most boring talk in the world. But because they came together and like did that talk in a way which was really funny and humorous. And then they had a live they had a live query showing the old versus new and a bunch of these other things. And literally it was a a a a technologist and a salesperson standing on stage at the same time, it had a really powerful effect. And we did similar kind of things a across across lots of different dimensions. to the point where the CEO, after we'd run like five of these every six months or whatever, the CEO stood up and said that these internal conferences have been the single biggest thing to align business and technology in this organization, which is amazing. Given an organization doing 150 million at the time or whatever it was, and
Anne: Yeah.
Matthew Skelton: we probably saved them five, ten million by running those conferences. Like genuinely CEO level value For running these types of knowledge diffusion type things, genuine business value. But it has to be very intentional, it has to be connected through to like what we're trying to achieve here, rather than just let's get some people together and nerd out about LLMs or or cloud or whatever it is. It's fine to nerd out, but what's the actual business outcome that we're really driving for here to to to make it stick? And so that there are definitely ways to run these kind of things which which which have that that impact.
Anne: Yeah, yeah. Yeah. And it's
Matthew Skelton: yeah.
Anne: it's also, I mean, not just good for the business, it's very good for the individual. As you point out in your book, very good for the individuals. Like nearly a a very large, I think it's like a third of the folk that are on this podcast were part of those. were part of the MetaSwitch conference. So say Jon Berger, he was a he he tuned all his material that it's in that's in his new book. on stage at that internal tech conference. Yan Qing Cheng, she was a very good on stage at that those conferences. Well I mean Liz Rice was at was at MetaSwitch is with me at the same time before we did all of that stuff, but the culture was there to start with. So yeah it's fantastically good for individuals to get used to public speaking. It's you know it helps really helps to shape your thoughts.
Matthew Skelton: even beyond that, so I agree with you totally, but I even beyond that, if you can get to the point where you've got individuals and and groups of people who can give a really good talk at an internal conference, they they they'll be good for giving a talk at a meetup group and
Anne: Yeah.
Matthew Skelton: start to spread out the practice that's useful for hiring and for and for ch sense checking whether what you're doing is is is is relevant at this point in the industry tech cycle and things. you know, there's no point spending lots of time and getting really excited about something which has already been solved and and you can just
buy as a service. but also if people are if people are very capable of giving a a presentation in an engaging way, particularly being able to change their message for different types of stakeholders, so they can dive down deeper for for people more specialized and they can keep it at a at a at a different level for people who are less specialized and maybe even signal that mid talk. So you might even say, hey, for the next three minutes I'm going to do a really deep dive for the technologist. But don't worry, I'll come back in three minutes. And you can wake up and we've got to continue. But that kind of that kind of ability to signal that stuff in in in in the middle of the talk then keep keeps people engaged and things. That skill is hugely valuable for like internal comms. If you're just doing a status meeting or you just do you trying to persuade someone to to to join you on a initiative or whatever it is, like that ability to talk about things in a in a in a clear way and articulate the benefits with the right kind of language, super, super valuable. Super valuable, not just individual level, but like for the whole organization. So there's a there's an outsized benefit from from helping from nurturing these kind of things and making them happen. And that outsized benefit is is then results in what you said about whether whether it's MetaSwitch or Financial Times, like being leaders in ways of working, adopting new new approaches and this kind of thing.
Anne: Yeah, yeah, yeah. I mean obviously I'm I'm calling you back to a book that you wrote a very long time ago by talking, but it's also relevant today, as you say, because AI is forcing very fast change on organisations and we need to to find quite new ways to s to to get everybody involved in the line.
Matthew Skelton: So
Anne: Ha ha.
Matthew Skelton: I I felt so let me let me ha let me do a confession here. So when I when I co-authored the Team Topologies book with Manuel Pais so we were writing it in twenty eighteen, published in twenty nineteen, the first version, and we got a second edition out in twenty twenty five, just about just just a bit less than a year ago now, at the point we're recording this. one of the assumptions I made when I was writing that book was that organizations would have A vibrant dynamic of people sharing things across different team boundaries and coming together and understanding all kinds of different perspectives. Turns out most organizations don't do any of that. They don't do any of it whatsoever. Now that's partly because I've been a little bit lucky in the organizations that I've worked in. Like they I've experienced that and I kind of just assumed, well, well, it's probably about the same everywhere. Most organizations are not most organizations would benefit from a substantially increased amount of cross. Team cross-department sharing, learning, innovation, diffusion, all this kind of stuff. So and that's particularly true in a context where you've got
organizational architecture that's been optimized for value flow. So that's effectively what team to podies is, right? It's organizational architecture for value flow. So it's the teams, but it's also the the systems and workflows that you're that you're that that and that support that value flow. so software services for sure and and digital infrastructure, but bunch of other things too, kind of workflows and and things like that. So all of that stuff is we're optimising for value flow. it was needed pre AI and that organizational architecture for value flow is absolutely needed with AI because that's how AI agents are actually able to navigate the the the the the different kind of domains and and different contexts for to to be able to orchestrate things and to to be able to carry out tasks and all that kind of stuff. So that's that kind of organizational architecture is absolutely needed.
Anne: stuff with point to construct things and track to update tests and all that kind of stuff so that's the track of architecture design together
Matthew Skelton: For pre-AI or without any AI at all. And then with some kind of AI, it's definitely needed
Anne: Yeah, but it's all coming back in W.
Matthew Skelton: as well. So like set up your organization for
Anne: For
Matthew Skelton: value flow by thinking about organizational architecture that works
Anne: Check that.
Matthew Skelton: in that context. The best starting point, I'm a little bit biased, but the best starting point, I think, is Team Topologies for that. All good, right? So you've got that in place. The danger is That those boundaries become too
Anne: I hope you speak up today.
Matthew Skelton: impermeable, is that the right word? They they become too watertight.
Anne: I don't know.
Matthew Skelton: So we need a way to share across
Anne: So we need to wait and see.
Matthew Skelton: the boundaries. And so I'm in the middle of writing a new book, which is all about exactly this. So we set up the organizational architecture for value flow. How do we make sure that the the information, the the knowledge and things doesn't get stuck inside these kind of flow-oriented boundaries? Well, let's make it intentional. With the organizations that succeeded in the cloud native era, like MetaSwitch, FT, the organization I worked for, and so on, these organizations were already very intentional in sharing knowledge.
Anne: Yeah, too confused and stuff.
Matthew Skelton: Actively diffusing this stuff, sharing it across boundaries, all of this kind of thing, running internal conferences and so on. So we already know the there's a correlation there at the very least. In fact, I I think it's a predictor for success, to
Anne: Yeah, I think it's a project of success.
Matthew Skelton: be honest. Like the organizations that do the sharing and the and the duplication detection of things are the ones that are going to succeed. But now with AI, it's like every organization needs to do this. So you've got your organizational
architecture in place, you've used Teams Prodigy to help you do it. What you're going to do to make sure you're intentionally sharing,
Anne: It told me.
Matthew Skelton: deduplicating, aligning, and so on. That it was a realization that most organizations don't do that intentional way of doing it. That that was a realization that made me think, well, let me write it down because I've been doing this for fifteen, twenty years. So the book is called Adapt Together, because it's how organizations adapt together. and the together part is clearly
Anne: that's a good time to try.
Matthew Skelton: about an intentional way of bringing people together to do this stuff. It's not
Anne: like intensive way of bringing people together to do this stuff.
Matthew Skelton: a we're we're not leaving it to individuals, effectively. We're we're we're thinking about it together. enables us to adapt new technologies, new compliance, new market conditions, whatever. And so it's a whole collection of patterns,
Anne: Is that whole can I can I
Matthew Skelton: case studies, examples, and so on of organizations doing this kind of things. And we're adding in some very intentional language, techniques, ways of using existing techniques to help achieve
Anne: technique warnings if you're using existing techniques to help change
Matthew Skelton: that. And it this whole approach of then being very intentional about deduplication, about looking for innovation, about sharing practices, creating shared standards and so on. This ends up being essential for an AI-based approach to doing stuff like value delivery,
Anne: sharing practices, previewing shared spends and so on. This ends up being essential for a based approach to doing stuff like that.
Matthew Skelton: creating products, that kind of thing. Because AI agents need context which is human generated. You can't have you can't you can't
Anne: Mm you can't look at you can't you can't you can't put in put back against the top of the phone. Like that that starts just doing the money collapse on a turn thing, it has to be acute intentionally.
Matthew Skelton: You can't put AI output back into the top of the funnel. Like that, that starts just to lead to model collapse and a bunch of other things. It has to be human AI intentional input. And that's what we're doing. We're not doing it in order to feed the AI machine necessarily. We're doing it because it's because it's how you have an effective organization. But a side effect is however you're using AI, you can provide additional context into the into the AI systems that you're using to make sure that the the quality of the information that is acting on is
Anne: in order to feed the AI machine necessarily doing it because it's the biggest the chair of an official organization. But the standard fact is however you're using AI, you can put on additional context into the things that AI systems are using to make sure that the the quality of the acting on is
Matthew Skelton: As high quality as possible.
Anne: Yeah. y yes, you you're not you're using it to align, to constantly realign, presumably, the the AI with what you actually need.
Matthew Skelton: Yeah, and s surfacing surfacing any any disagreements, surfacing any misunderstandings, surfacing anything where we've got I mean, where where the the terminology is not aligned, because if you've got the terminology that's misaligned, then AI is not gonna play well with that at all. so for th those of us who are a bit more technically minded on the listening in, so that's all that domain different design is is the new hotness, right? That again. going back to when it was two thousand four when the book was published, and it's twenty odd years later. I mean it's always been relevant, but it's it's it's absolutely essential now. so We've got we've got un underneath we've got the organizational architecture. On top, we're layering on this very deliberate intentional approach to creating context effectively by learning and sharing across the organization, by setting up the conditions so that leaders have the ability to shape what's happening without having to be in the detail of absolutely everything. So what kind of metrics and Other guardrails can we put in place that actually enable rapid value flow without there being a whole load of manual checks all the time. Again, it's been proven in IT delivery since 2020, 2010, certainly since the continuous delivery book came out. So we've proven the approaches that work in that kind of space. And now we're taking the same principles and applying
Anne: twenty two thousand eight twenty ten years continuously everyone came out. So we've proved from the approaches that work in that kind of space and then we're taking the same principles and apply.
Matthew Skelton: it to the entire knowledge work organization. I think that that's that's what we're talking about. And part a key part of that is.
Anne: The active approach to to thinking about this
Matthew Skelton: an active approach to diffusing awareness and knowledge
Anne: knowledge.
Matthew Skelton: across the organization.
Anne: it reminds me of conversation I had with set with Sarah Wells, who's
Matthew Skelton: Yeah.
Anne: in your book, that where context isn't clear, AI really, really can't handle that at all. And that the the examples were code languages. So Java is difficult because there's so much Java and it occurred in lots of different contexts. There was Java written 20 years ago, 30 years ago, when we were just doing things very differently, and we had different needs for the systems. And then there's modern Java, which is completely different and is written for the current context. The same with C and C. There's such a huge history.
Matthew Skelton: Mm.
Anne: But the AIs can't can't really handle that very well.
Matthew Skelton: Yeah, because
Anne: How do they do that?
Matthew Skelton: Unless you can somehow exactly unless you can codify yeah, well back then we use we use C or
Java in this way, but now we wouldn't dream of doing that. We use it in this way. Of course we do that. Can you and that's a really important point, to be honest, about the whole I adoption thing. Knowledge exists in the heads of humans and in the in the in the shared experience of humans. Like it's not just in human heads, it's a it's across groups of people.
Anne: is across groups of people.
Matthew Skelton: I mean there's actually a there's actually some slightly mind blowing effects that you see in groups of humans where where intelligence and knowledge is effectively shared between a group rather than just inside people's heads. And it's very difficult to get to to get you to get your head around it, as it were. But
Anne: Yeah, as it won't. But
Matthew Skelton: but we'll we won't talk too much about that today. But the i I think it's important to to To keep that in mind, knowledge and intelligence don't exist in the AI systems, not in terms of the LLM based systems, and obviously in a very in a very rarefied form or very, very specific form inside the non transformer based systems as well. Knowledge exists in the collective understanding and experience of people. And that's exactly
Anne: And that's what exactly.
Matthew Skelton: why approaches like conferences and things work really well. So you know very well the InfoQ queue.
Anne: website and they can keep conf conferences that come out of it and tell us to be involved in this radius and so on and so on. It's been great to facilitate some of the track by two months in the last few years. That's a really great example of bringing together that clearly well that's very great meaning that's very
Matthew Skelton: website and the QCon conferences that come out of it. I know Sarah Wells has been involved in this for ages and so on and so on. It's been great to facilitate some of the some of the tracks at at QCon London in the last few years. That's a really great example of we're bringing together that shared awareness. That's how we create meaning. That's how we generate context. That's where the context is stored. The context is not stored. Some of it's stored on the website. If you go to InfoQ
Anne: generate context, that's where the context is stored, the context is not stored. Some of it's stored on the website because in quite view
Matthew Skelton: And there's some great articles and and it's a it's an amazing treasure trove of of of things there. But a lot of the context is stored in in the shared experience of people who've been there and had that conversation. And that's ultimately a really, really important point. Difficult to quantify, difficult of some people to get their head around for sure, but that's the actual reality. That's the underlying reason
Anne: has some great articles and it's it's the main treasure trove of of of things there. But a lot of context is stored in MA because it's generally experience of who have been there and having a conversation. That's what's made a very important point. Difficult to quantify and difficult to some people to get their head around for sure. But that's the actual reality that's the underlying reason
Matthew Skelton: why approaches like internal conferences and things like that effectively making space for people to come together and share and learn from each other.
Anne: And then things like that eventually making space for people to come
Matthew Skelton: Whether in person or online, is part of the, if you like, secret source of organizational success. Because that's where the actual knowledge lives. It doesn't live in a wiki.
Anne: together and generally learn from each other, whether it wants more money. No, no, and in fact, oddly enough, I've had this conversation several times in this podcast previously as well, which is the value in a conference I always think is more in the hallway track than the speakers.
Matthew Skelton: You've got to the speakers and and those sessions there, 'cause that's what that's what draws people in. And you'll be able to have the recording of it and you'll be able to like talk about it afterwards and so on and so on. And and the fact that someone has if someone's going on stage, they have to be prepared enough to be able to codent condense their thoughts. And so there's there's value in that part. But you're right, a big chunk of the value is in is in the conversations that are sparked in and around that. So you need both. you can't just have a hallway track without the other stuff. But then y you need to look at h look at it holistically, I guess, is what i is is the way to way to do it. You've got an event, it brings people together, the the value emerges from the dynamics of that whole event, not just of one specific slice of it.
Anne: Yeah. I I had this conversation with Charles Humble, who I think I think you know Charles Brook pretty well. He used to be very involved with we he we he was the editor in Chief of InfoQ magazine. And he and he was saying that what you have to remember is that at a conference the the talks are not are not edited in the way that if something's on InfoQ people say something and then it goes to an editor and then it goes, I'm gonna fact check that, I'm gonna make sure that I'm gonna I'm gonna give you feedback and you're gonna have to rewrite it. In a conference, that task really falls to the listeners to talk amongst themselves and go, hmm, does did that hang together? Do we agree? Have we seen something similar? And you really do. You need I I I've I've seen that happen when I when I talk on something which is. An unusual thing that people aren't just already kind of it's not already in their heads that they'll go away and then it's not until the next break that people will come back then the break after the next one that people will come back and go, I liked your talk. I've gone off and we've discussed it all amongst ourselves and we've decided that it was okay.
Matthew Skelton: Nice. Yeah. Yeah, yeah. Yeah.
And and the the the so the reason why the the IT industry had particularly software had so many conferences, so really really valuable conferences like like QCon and and similar things is because the because of the pace of change.
Anne: Because change. Mm-hmm.
Matthew Skelton: and that's not gonna that's not gonna decrease any time soon. So th these kind of events and conferences will always be massively valuable. What's of course happening now is the pace of change for pretty much every other part of the organization is now
Anne: Mm.
Matthew Skelton: also catching up and being very, very rapid. and it's not comfortable for lots of people. Lots of people like things to be just the same as we've always been doing it. And for sure, for some for some professions things will pretty much carry on as they have done before. I don't know, like someone who's a professional massage person or someone who I don't know, is a sports teach sports coach or something. Maybe that stuff doesn't really change very much. but loads of other professions are are are being affected hugely by particularly by AI. So organizations that realize that that this is going to be a a a major problem. Like if if if they don't take an intentional approach, an active intentional approach to curating how learning happens and curating the direction that we're traveling in and a kind of shaping the kind of place we want to get to with different kind of technologies and things, they're they are just going to have they're going to have organ they're going to have whole departments of people Each one of which is pulling in a different direction. And if you think about a big container ship with a load of tugboats, if each one of those tugboats is pulling in a different direction, the ship's not gonna not gonna move at all. It's just gonna stay still or move in the wrong direction. that's why there needs to be an intentional we need to go over here. This is where we want to be with AI. We want to use it like this. These are our guiding principles. This is what we want to get. Forget your noodling around with Chat GPT, that's a complete waste of time. It was interesting for a couple of years, but it burned a lot of money, didn't deliver anything. What we're going do, like we want to achieve this. This is the specific or these three things are specific problems that we want to want to address using AI. If your use of AI doesn't, if you can't demonstrate an alignment to solving one of these problems with AI, don't bother using it. What's the point? Because we don't need more and more content, that's for sure. We don't need more and more reports. We don't need more and more code. We need things that solve this specific problem or these
Anne: Yeah.
Matthew Skelton: specific problems. And We've already seen this with with you know the the leading organizations that end up doing that. being way more intentional about about the either it's solving customer problems or it's solving internal like operational problems or what have you. But but there's still loads of organizations, unfortunately, which are just like, yeah, crack on and we're gonna measure you measure you by how many tokens you burn per day for your AI usage really? How much coal we're gonna shovel in the in the furnace in the in the in industrial revolution. You're gonna measure people based on that. Like it it's it's the direct equivalent. And it's it's easy to measure, but it really doesn't shift anything at all. so I it comes back to kind of that the the intentionality around what we're actually trying to do. What's the purpose of either this department, this organization, this initiative, the the you know, what problem are we solving for customers or what have you and And that that increased intentionality about Whether it's AI usage or not, or it doesn't really matter what it is, it it at some level, it's intentionality about how we're operating and what the what the future state should look like and therefore how we're gonna nudge people towards that future state.
Anne: Hm.
Matthew Skelton: That's ways of working, that's use of technology, that's that's how they work together, that's how what the team stewardship boundaries look like, a whole bunch of things. language that people are using to to to talk about solving problems, a whole load of things. How are we gonna nudge people towards that future state? This is the future state we look like. Let's nudge people towards it. Let's put in place just enough, if you like, ecosystem things like people running conferences, maybe l lunchtime talks, people thinking about common standards for doing things, but not clearly not in a way where where they're going to enforce it by saying, you will now do use this cut this way of working, because that never works, right? So like how do we set things up in a way that people find attractive or at least kind of fairly compelling to adopt a different way of doing something? How do they find out about it? If someone's found a better way of I don't know, peeling a banana, whatever. Like this is the much more efficient way of peeling a banana, much more effective. This will save us like a million pounds a year. Cool. How does anyone find out about that new way of peeling banana? Like what's your mechanism in the organization? What what what's your what's your information, your knowledge diffusion approach that means that most people in the organization within the space of six months have found this new way of peeling a banana? Like that that that's the nature of of the problem because if you don't solve I mean, obviously the the peeling banana thing is is a bit is a trivial thing, but the there are many of these, there are so many of these like things where we need to get people to to be aware of a new way of doing things. If it's not intentional, then the organization is burning millions of dollars a year on people doing things in a way which is by far the most effective way of doing it
and is not even pulling in the direction that we want. So the whole thing has to become more intentional, more active.
Anne: Yeah. I d it's a it's an interesting one. when I read Team Topologies, one of the things I liked most about Team Topologies was the idea of cognitive load. that you try not to distract people with too many things that are not germane to what they're what they're doing. And you people have an a an amount of cognitive load that they can they can take on. And you want to make sure that it is as germane to what they're doing as possible so that you if it's if it's if it's useless to cognitive load to them. And I was it's an interesting one that's like you have enabling teams which which kind of spread best practice in team topologies. and then to a certain extent, that is all about managing cognitive load, keep trying to keep all a lot keep people aligned with germane. knowledge and useful knowledge and not and and try and filter out some of the the less germane stuff. Conferences are kind of a bit like well you kind of always want to be on the edge of you may or may not it may or may not be germane. So how do you how do you control cognitive load so that people hear about things that on the surface might not be useful but might be useful. But don't get overloaded with a whole load of I mean, the the the the classic is chat you talk to ChatGPT and it'll give you three thousand ideas. and they're all interesting. And if you read them all and you try and take them all in, it's you know, your cognitive load is not ChatGPT's cognitive.
Matthew Skelton: Sure, exactly. So I mean ideas have always been cheap and they're they're they're effectively they're they're they're extremely cheap with things like Chat GPT. So in the new book, Adapt Together, we've got this concept of an innovation and practices enabling team called iPet. If you come from if you come from the northeast of England, it it sounds like you're referring to someone in in a in a in in a in a way which is endearing, like iPet. It's not that. I
Anne: Yeah.
Matthew Skelton: P E T, IPET. innovation practices enabling team. So it takes the idea of an enabling team from teams topologies but it it has a very specific focus. So you can have enabled enabling teams for a bunch of different things. You can have them for like ways of working, you can have a cloud enabling team, you can have an architecture enabling team, you can have a compliance enabling team, all sorts of different things focused on on on different solving different problems, and you've got different types of skills and things in there. That's fine. Inside this ipet innovation and practices enabling team, you've got people who've got a mix of people. They're looking for early versions, early emergence of good ways of working. Or they're looking for ways of working which are slightly better than the most common ways of working are now, and they're pointing in the direction that we want to go. So this I innovation
practices enabling team, they're they're people who definitely connected, very, very aware of what good should look like in twelve months, twenty four months, you know, five years, whatever. And they're actively looking for like the seeds, if you like, the the seedlings, just the the seeds sprouting that show early versions of of of a new a better way of working. It might be a better way for now. And we know that in twelve months' time we actually need to be looking for an an even better way of doing it. But for now we want to showcase this way of doing it because this will help to pull other people along. so it's a very intentional way of going across the organization looking for stuff. Well, we need to set up some incentives as well. Like, so what's the incentive for a team that's just done some new data cleansing approach? What's the incentive for them to share it with others? Don't know, but like that there's some ways of nudging behavior to to encourage them to share that. Innovation practices enabling team can be there to help the data team, for example, to talk about their innovation in a way which is really engaging and compelling. So they might literally work with them on How to write conference slides or how to write it how to how to write slides for a talk. We are not using 12 point font. Thank you very much. We are going to use a massive great 36 point font or a hundred point font in the middle of the slide with one word on it. And and then help them actually, they might be not very confident, they may never have given a talk, even to a small group of people. There's some s fairly straightforward like tips for giving a talk. Maybe that person is happy to give it try. Or maybe that person says, Do you know what? I'm never going to be confident standing up. I never want to do it. I want to run a mile away. Okay. So the the innovation practice enabling team can interview them instead. And maybe write an internal blog post. And so on and so on. But what we're doing is we're that that ipet is looking, it's being very intentional. They know what the future needs to look like. They're looking for emergencies of the start of that future across the organization. Because as soon as you've got, you know, more than about two hundred, two hundred and fifty people, you've got lots of stuff happening.
Anne: Mm-hmm.
Matthew Skelton: And it's all happening kind of somewhat isolated. So that that ipet is out there looking for that stuff, curating it, bringing it in, saying, look, we've got a lunchtime talk here, we've got some pizza, or we've got some sushi, or whatever the thing is, some nice like some nice salad, whatever people like to whatever's appealing for people to come come and and hear that stuff. or or we do a virtual version, like you get you get like a You get a little voucher in the post or in it by email saying like go and get yourself a a coffee at your local coffee shop and then bring it back and we'll all sit down together on the call. However you want to do it in in person or online. That ipet, that innovation practices enabling team has curated a couple of talks that week. Look, we've we've found this team, you know, team team this team over here has done this interesting data cleansing thing and we've helped them to talk about it. Here it is. And they've shared the code over here or they've shared this approach over here. at a wider scale, you know, it might be that the that the the HR department has started to learn how to use like a technique like Kanban for onboarding new new employees. And the HR department is talking about their collaboration with one of the IT teams who taught them how to use it, but the HR department's got a slightly different spin on it. So the HR department is now talking about. their experience of using Kanban with work in progress limits or something like this. And and their perspective is actually adding something new because it's completely outside of the world of IT and therefore is bringing like some ev the opportunity for other people to learn. Perhaps new hires in IT have never seen Kanban and work in progress limits and they learn about Kanban and work in progress limits from the HR department. There's opportunities to shape how this knowledge and awareness is shared across the organization very deliberately. And we go around that cycle all the time. Every week we're going around the same cycle. So I've literally done this. I've done this in multiple organizations and it's always seemed really obvious to me. But yeah, like I said before, it it's most organizations don't do it at all. we're going around the cycle often. We're spotting the stuff that doesn't work, we're spotting the problems. If we want, we can put in some additional uplift, like assessment approaches, like team based assessment approaches, things like this to to bring more awareness to the surface. But of course, if we've been running some of these either assessments or internal talks or something like this for quite a while, well, we've got the material for a conference or internal internal event. We run the run internal event. Does that then provide can that then provide the the people that are going to go out into the into the wider world and talk at a meetup group or talk at a conference? It really helps with recruitment. It really helps with actually validating that the way that we're doing things internally makes sense. Because
Anne: Mm.
Matthew Skelton: if your conference talk gets accepted somewhere, then it probably means you're at least what you've what you've got in your conference proposal looks pretty sound or interesting. you it might look sound interesting and it turns out when you give the talk, everyone the feedback from the conference is, my goodness, why are these people talking? This has already been done before. Here's here here's a here's existing offering that just solves that problem entirely. I mean it's a bit late to realize it, but hey,
better late than never to realize that what you this initiative that you're spending, you know, fifteen million on is actually completely solved by a a thousand thousand dollars a month subscription service, right? So Part of all this stuff is about building the organizational awareness to understand the context we're operating in, internally and externally, so that we're able to focus our limited resources on the stuff that really matters and that really starts to to to to solve the the problems that we're here to solve. That's ultimately what it is. It's like additional self-awareness and kind of organizational self-awareness. It's like a nervous system.
Anne: Mm.
Matthew Skelton: Like a nervous system for the organization. It that shouldn't surprise you, because like I I did a a master's degree in brain science. It was
Anne: Mm.
Matthew Skelton: called neuroscience that back then. Neuroscience now means something like it means something different, but effectively brain science. so I've definitely absorbed some of that stuff about like how the nervous system works in in in in in nature, and I think about an organization in those terms as like an organism and so like what's the nervous system look like? How well is it is the organization sensing its own internal environment and the external environment?
Anne: Yeah.
Matthew Skelton: So that book will fingers crossed be published first half of twenty twenty-seven. We'll see. I w n obviously, as you know, right, as an author, you ne you never commit to a published date, right? But hopefully first half of twenty twenty seven, if not the second half, obviously. and we're we're pulling together case studies as we go and publish them publishing a version of it online where we can. Not all of them will get published online, but but most will. And increasingly like the the work that that that I'm doing with my colleagues at co at Conflux is bringing together these these two different perspectives of team topologies as organizational architecture for value flow and adapt together as this cross-organizational, like very intentional way of doing building this continuous context effectively in order to improve alignment and deduplicate and detect innovation really quickly. And the combination of those two. is feeling very powerful basically. It's already been I already been proven out, like we've said, like organization have been do some organizations be doing this for 15, 20 years. But I guess the cross-organizational intentional sharing stuff and deduplication like hasn't always been codified. we I did I did a tiny version of it in the internal tech conferences book. But the adapt together thing is trying to codify a lot more of that. to make it easier for organizations to take it as a as a whole thing and go, Bam, let's just let's put this in place.
Anne: See it is interesting. you kind of alluded there to something that I think that the the FT, for example, did really well. In fact, all the companies that had done really well on adopt adopting Cloud Native that I included in Cloud Native Attitude, where I I went and talked to a whole load of companies who would who'd adopted very well, is is define at the highest level, get you know, get at this the C-suite level, they define what good looks like. So they they
Matthew Skelton: Hmm.
Anne: Make it clear so they at least have an idea in their mind about where they're going, where the company is going, so that you have something to be telling people in the internal conferences, something to be aligning against. Because as you say, alignment is very important, but the key thing in alignment is actually having something to align with, a vision of the future.
Matthew Skelton: Yes, I'd agree with that. And and actually you've helped me see something extra there, which is that if I think back to the organizations that have that have found some of this stuff quite difficult, I would say that they were finding it difficult partly because the leaders of these organizations did not The leads of these organizations often I think assume that the work is obvious and that the way of working is obvious. Or that it shouldn't be hard, it should be very straightforward and just get on with it. Why can't you just get on with it? Rather than starting with the assumption that actually we do need to define ways of working, we do need to codify ways of like language and terminology and and this kind of stuff in order to to get things done effectively. So the ones that succeed, I think you're right, are the ones that expect to have to define and codify ways of working and standards and a bunch of things. And the organizations that that always struggle are the ones where the leaders just think, how hard could it be? Just crack on and do it. And I think, I mean, it's a bit of a broad brush, but I I think there's there's there's something there about leadership sort of expectations and Leadership sort of disposition or posture toward the work itself.
Anne: Mm.
Matthew Skelton: is is the work really obvious? In some cases work is fairly obvious and we don't need to spend too much time sort of codifying it. Sometimes it looks like overkill. Other times it turns out that the work is not obvious at all. And so therefore, like articulating, as you said, what good looks like at multiple different levels. It's not good doesn't just like, we've increased profit by five percent. Yeah, but how are we going to manage this kind of data ingest? Because it turns out that like last year it was one gigabyte, now it's one terabyte, we can't do it in the same way. just just sort it out. It's really obvious. No, no, no, really. It's not obvious because it completely changes how we need to do this thing. So what's the right way of working? We need to define that, discover it. And we need the we need the the intelligence, human intelligence in the in the organization to be able to get to the point where there's enough understanding of what
good looks like to be able to effectively s specify that set that direction and continuously challenge it all the time. Because we've set a direction that might be valid for three months. Particularly in the in the context of AI, like in AI world, but in cloud and a bunch of other places too, what was relevant three months ago might not be relevant now. It's a continuous process of dis of rediscovery about what good should look like. And to make sure we've got a dynamic in the organization that can refresh what good looks like on a sufficiently short cycle. So that's what we're trying to do with the Adapt Together book. But whatever you use. your organization is going to have to do something like that, whether you choose to use a adapt together, whether you choose to go and buy the book or not, you're going to have to do something similar, which is like a continuous refresh of context and what does good look like. And and like I said, a dynamic of making sure that's always up to date, rather than just assuming that you hire people and then their skills at the hiring point, their skills and knowledge at the hiring point are enough for you for the next five years. That's just that's never been the case, but it's definitely not the case now.
Anne: Yeah, no.
Matthew Skelton: Thank you.
Anne: Things are very seldom obvious. They're it's like everything that's important, there's a trade-off associated with it, and people need to understand what they're prioritizing in that trade-off. You know, that the that that's that setting directions for people to align to is really a high skilled job.
Matthew Skelton: Yeah. Yeah yeah yeah. put it on t shirt perhaps. It's it's such a s important starting point for for effective leaders is the work is not obvious. That that almost almost like that that's the that's what's on the t shirt, right? It should say like I know the work is the work is not obvious or something like this. And so that's their starting point. So it's not obvious. Let's go and find out what it should look like for now, for the next three months, for the next six months, for for for a period of time which is suitable for that way of working, but we're also expecting to revisit it and reshape it. What was suitable six months ago for our data ingestion no longer is suitable. Sorry, we've actually got to change it. And and just be just accept that, be happy with it. That's a it's it's a price of A an organization existing in the ecosystem is is the need to continuously refresh how we do things.
Anne: Well that is an excellent final point, I think, for our Millia, the hour point on our discussion. And that is an excellent point. It's all about change and managing change and taking on board the difference and accepting that it's not obvious. The
Matthew Skelton: I think so.
Anne: Yeah. I internal conferences actually is another reason for doing internal conferences, isn't it? Because you You go out and you're forced to talk and you talk to people and you learn that actually it's not obvious because they say, did you mean bit this or this? And then say, No, it wasn't obvious because I could have meant this or I could have meant this. And and and it's like, Yeah, what do you want us to trade off for that? Do you want us to less of this? And also, no, it wasn't obvious because people don't know what to trade off for this. Unless you speak to people, you never find out.
Matthew Skelton: Exactly. Exactly.
Anne: so so thank you very much indeed for being on the podcast. And I wasn't expecting you to be announcing your new book. I'm not sure that you thought you were going to be announcing your your new book either.
Matthew Skelton: Exactly. There is something else that that I'm I'm working on as well, which is called Engineering the AI Native Organization. and it's a leadership program for taking non technical p people from a non engineering background through a journey of awareness about things like decoupling APIs, a bunch of like stuff that we know from engineering effectively. th like parallels to version control, things like this, things which which which we absolutely know we have to we have to asynchronous, things like this we have to work with. but then we're having to apply engineering principles to the entire organization now. So Team Topologies is you could say it was applying engineering principles to the kind of digital part of the organization. Now we're applying it to the rest of the organization. How do you explain decoupling to to head of HR? How do you explain APIs to head of head of finance? That is what that program will address, because it's going to help these other leaders to get enough of a handle on some of these core engineering concepts to be able to s have a better conversation with with other parts of the organization. Because we're going to need those kind of concepts in place if we're going to make effective use of AI in the organizations.
Anne: Absolutely.
Matthew Skelton: So that's why I'm working on it in the background.
Anne: well, yeah, so as I expected, lots of interesting things and I knew you'd be totally AI within organisations, because you know, that's the big the big thing at the moment. It is incredi it is the struggle that everybody is facing. And as you as you were saying earlier, it's it's amplified all the other difficulties within the organisation because it's accelerated things. There's nothing like
Matthew Skelton: Hm.
Anne: accelerating things to really expose all the choke points and the bottlenecks and there on all the problems that you can have to the challenges that you've got to overcome. So thank you very much for being for being on the podcast. Yeah. so
Matthew Skelton: Cool. Thanks for the chat. Loved it.
Anne: thank you very much to our listeners and our watchers of Asynchronous and Unreliable. I will catch you again on a future episode.