moderator  0:00  
Welcome to the scrum.org community Podcast, the podcast from the home of Scrum. In this podcast, agile experts, including professional scrum trainers and other industry thought leaders, share their stories and experiences. We also explore hot topics in our space with thought provoking, challenging, energetic discussions. We hope you enjoy this episode.

Dave West  0:25  
Hello and welcome to the scrum.org community podcast. I'm your host. Dave West, CEO here@scrum.org in today's podcast, we're talking with Tony Hinkley. Tony is the Chief Technology Officer in the UK for Avanade, for those who do not know Avanade, which there might be one or two out there. This is a 25 year old joint venture between Accenture and Microsoft, and they have a special place in scrum.org heart, because they were part of the founding group of companies that helped Ken Schwaber start this organization. So welcome to the podcast. Tony, thanks, Dave, always good to hang out for a bit. It's good. I'm excited to talk to you on air, as it were on recording, because we've had many conversations about the topic that I want to focus us on today. But before we do that, maybe you can tell our listeners a little bit about your role at Avanade,

Tony Hinkley  1:21  
yeah, absolutely. I affectionately say that I wear lots of hats at Avanade. My role is officially the Chief Technology Officer for the UK, and that means that I spend a lot of time thinking about where technology is going, how we build capability inside of our organization to make sure that we are ahead of our customers on those journeys, and to really build teams to work with our clients on making those technologies deliver business value. So that's very corporate. That's kind of the highlight message there. But my background is as an engineer, so I've been a programmer and architect and developer throughout my career, and a professional scrum trainer, as you know, for sort of over a decade now. So I still work quite a lot with delivery teams, with engineering teams, with product teams, helping them to work through how to get that technology to really serve them in delivering products for businesses. And yeah, that's that's where I spend my time, is thinking about what we should be doing and how we might do it, and then working with teams to make sure that that is actually how we go about delivering.

Dave West  2:38  
And the reason why you're on the podcast today. And no surprise, the world is changing quite, quite fundamentally. I'm faster than I've ever seen it. I've obviously been in this industry. Gosh, 35 years, something like that. Yeah, over 35 years. Oh my gosh, I'm feeling very old now, Tony, so I'm not going to say that ever again. But, and, you know, I've seen the introduction of PCs. I started on mainframes, and I've seen PCs. I've seen object oriented software development, service oriented architectures. I've seen the rise of the cloud. I've seen, you know, I've seen lots of things happen during my tenure in this in this industry, but in the last two years, I have seen more change in the way in which product delivery, and particularly software product delivery, is possible. So what are you seeing with regard to that, particularly around the adoption of AI inside your client base. What are you seeing?

Tony Hinkley  3:43  
Yeah, and again, like you, I've been around for sort of 30 years professionally being being paid to work in this industry. And I've seen mobile, the Internet, mobile, digital, when that became a thing, and this is absolutely one of the most disruptive periods that I can remember in the time that I've been working. I'm going to take two lenses on this, if you don't mind, because we talk a lot about what we do with clients, but I think it's worth remembering that Avanade itself has been massively disrupted about how we work using AI. And like I said earlier, we like to be ahead of our clients, so that when we go to them, we've got real experience. And so we've been going through adoption of AI internally and how we build products and services for longer than most of our clients. And so if we were to break that into a couple of waves, the first you said, two years, and that's about when chat GPT really came to market. Has been this era of generative AI. It's been around, how do we enable the creation and the analysis? Analysis of large amounts of data, whether that be code or documents or conversation, through these AI models, which have got better and better every several weeks for the last two years. And that's impacted a lot of business users, a lot of white collar workers, if you were to use the popular parlance in the press, are seeing the way that they do their job on a day to day basis fundamentally change. And as consultants, the whole industry professional services full stop, whether that be legal, or whether that be it, services have been probably hit hardest by the impact of AI because we are knowledge workers, and there is a perception, rightly or wrongly, that knowledge is now a commodity, as long as you can plug these AI agents into the right data sources, or they're trained on large enough volumes of data that the knowledge that sat in people's heads in professional services is no longer as valuable, no longer as important. And I think that's probably hit different people at different levels, by by different by different ways. But we're also a technology company, and so we've seen one of the earlier use cases of AI being in software engineering and software development, starting from basic IntelliSense, auto completion, type of type of tooling, and we've had that for years, but then into sort of prompt driven development where we can generate large amounts of code and define the standards for how that code is developed in much shorter periods of time. And the reason why I think we use the word disruptive to describe it is because people now feel that a core part of their job has been diluted or taken away by the fact that these tools have become ever more capable at doing some of those, some of those things. So I'll just sort of pause and reflect on that for a second, because this is the first time most of those individuals have had their work impacted in that way in the last 25 years of technical evolution, right back to desktop publishing versus traditional typesetting and printing was probably the last time a lot of office workers that were working in that kind of space have really been disrupted in this way, and I think that's why we we feel it so as Avanade, we have a consulting workforce, we have a technology workforce, we have a leadership team. Every single one of those audiences has been impacted, and that is the difference to some of the other changes of technology that we've lived through and experienced. I think if I then take it out to our customers, they are grappling with how to bring this technology into their company without the understanding of what the technology under the covers actually, actually does. So we have an advantage is that we have these business roles, these office worker roles, these knowledge worker roles, but they're supported by deep knowledge of technology. So we don't we don't feel quite as unaware of the ways of working under the covers. We don't feel quite as disrupted by it or displaced by it, but with our clients, they're across every industry. So this morning, I was talking to some people from central government. Yesterday, I was in meetings with an education client who do a lot of work with schools and teaching and everything in between oil and gas, energy, finance, all. They're not technology companies. So they see the outcome, they see the marketing message. They see some of the possibilities that are being touted on stage by big organizations like Microsoft, like anthropic etc, and they don't make sense of it in the same way as we do as Avanade, so part of my job is helping to dispel some of the myths talk about The realities of what the tools can do, how they achieve that, because if you don't give these tools good data, good guidance, good guard rails, and govern them, and governance is a dirty word for some people, then you can't deliver the kind of impact that maybe they're expecting. And none of that. Is a software problem or a product problem. This is a people and culture and change discussion before we even get anywhere near the tools. So I see quite a lot there, and I don't know if that kind of resonated with you.

Dave West  10:15  
Yeah, yeah. So let's just unpack that a little bit. So it was interesting that you led with the, I was surprised, actually, the Avanade thing, because you're right, it has been a huge disruptive technology to white collar, potentially disruptive technology to white collar people, of which, obviously, I hadn't even thought about that a consulting company is, by its very nature, a white collar organization. And so there's a lot of fear and angst, I'm sure, around that, and confusion engaging with clients. You know that so that there is that it's interesting. You You led with that. But are you? Are you seeing organizations including Avanade, really take advantage of the tech, not that that technology in a way that we potentially can. And what we're seeing with startups and smaller, smaller organizations. Are you actually seeing people start to see those productivity gains and and those opportunities,

Tony Hinkley  11:18  
absolutely and in maybe some some unexpected areas as well. Obviously, you and I both are heavily invested in helping product teams and product organizations deliver those those products and services to market. So if we start with engineering and Scrum teams, product development teams. We've seen the tooling become very mature very quickly, and the adoption of that tooling ramp up much more quickly than in any other part of of the business started earlier. It was being pitched at an audience that were receptive to it, which I think makes a massive difference with adoption and the benefits were really, really tangible. These tools are effectively generating content, generating knowledge in what is actually a really fuzzy space, like human language is a really fuzzy thing. Programming languages, however, are very prescriptive. The tools, the processes, the expectations, are really well defined. There's a huge body of of knowledge around how those how those things should be structured. And so the tools were much more effective in that space from a much earlier stage, and that's why I think we see product teams impacted so quickly and so heavily by the introduction of these tools. And now what we've learned from those teams is starting to bleed into how the rest of the organization use them, because as the developers, and by developers, I mean programmers, in this instance, not the way that we necessarily talk about them in the scrum guide, have become much more productive, and the output that they can produce has become significantly cheaper per line of code to produce. What it's done is it's revealed knock on effects elsewhere in the system, which you would expect. If you're a systems thinker. You know you exploit one constraint, you expose a load of others, and where it's exposed the product teams is in context and the quality of requirements and the ability to express our intent has become the new bottleneck, because these Gen AI tools will happily produce whatever you ask them for, including if you ask them for the wrong thing. And so we've started to see business analysts and architects and product owners and testers starting to adopt these tools to do a lot of tasks that traditionally they had no means of getting assistance with. And those those ripples are starting to head out into the rest of the organization. People are now seeing the tools and seeing what they can do with them, and starting to exploit them in lots of really interesting ways. And this is all just in the product space, before I even start to get into finance and HR and legal and other

Dave West  14:34  
areas, and before, before you do that. So I've also seen that sort of proliferation of capability across all of the developers of the product and and I think it's sort of some of its I dare I say, going back to its roots, like domain specific languages, well structured, defined requirements, gross. S descriptions, you know, flow diagrams or things like that. But the great thing about AI is it can help you do all of that. And so it used to be a burdensome task when, I don't know if you remember ssadm and things like that, but or building your ml, obviously, I was heavily involved with that because of my rep background, but that was a burdensome task that nobody could help you with and nobody really cared about at the end of the day. So you'd give it to your developers, and they go, great, what do you want it to do again, damn it. So, so I've seen a lot of that impact across development teams, you know, and I so I think the same as you What's interesting though is, and this is sort of disappointing element of this. Are you seeing a lot of smaller teams? Are you seeing people basically taking that decision and going, Oh, well, we used to need 10 people. Now we need three people. So are you actually seeing developers do more, but then, you know, the whole team just do more, or are you seeing it do the same, but with fewer people? Is that what most organizations are doing?

Tony Hinkley  16:15  
So I don't think that there is a CFO or a business leader out there that wouldn't love to reduce their costs and increase their output, right? So I think there's a natural tendency to look for cost reduction and optimizations, especially with I don't want to get political, because this is not the space for it, but the broader things that are happening in the world and the world around us, I think most business leaders would take that opportunity to reduce cost and consider it really seriously. But I've seen a couple of different sort of stereotypes of teams. There are teams who have historically not been able to deliver at pace the intent and the vision of their product leaders, and so they've got a groundswell of ideas and have never been able to capitalize on them. And for them, this is actually a top line conversation about we can suddenly deliver these features that we've always wanted to deliver with the capacity that we've got, and so we get much, much more growth. Yeah, and then there's other organizations that are finding actually they can't deal with the constraints that they have in their product leadership and and in that vision and intent, and so the developers are just not utilized. They're sitting idle. And why would you carry that cost forward if it's not driving any meaningful benefit? So I can deliver the same for less? Is one mindset, or I can deliver more with the same, is the other mindset. I would urge organizations to be trying to deliver more with the same. Maybe change the composition of the team. So talking to a client recently about, how do we bolster the left of the design and thinking more event storming, more ideation, more support for the product organization to come up with new ideas, rather than downsizing on the development capacity. And I do see people that are buying into that, and it's giving people a new lease of life. If you were a product owner historically, who was just swamped by taking care of your backlog and didn't have the time to be able to engage with stakeholders, work on the strategy, test the market. Those are things that are that are now much easier for you to do. So you can take analyst and researcher agents that are being provided by Microsoft in 365 Copilot to do market research, to analyze signals from the market. You can take things like GitHub, copilot or Claude and ask it to generate better requirements and user stories to hand on to your development organization. It's actually quite liberating to have these tools.

Dave West  19:17  
I think it's quite liberating for everybody. I mean, from that from up and down. So one challenge that I've seen a lot of organizations, though, is ultimately governance. The governance processes aren't designed to accept this level of capacity, this level of stuff coming out. And you know, meets every three months, or maybe they're using Scout, agile, or some other thing where there's this very strong governance process in place fair play, you know, it's, you know, particularly for traditional organizations, it's a big risk to release something right, and now suddenly it's. Three times as much stuff, or four times as much stuff in that bucket is

Tony Hinkley  20:07  
absolutely yeah.

Dave West  20:08  
So how the governance is that the next frontier for the adoption of llms to improve governance processes?

Tony Hinkley  20:16  
Well, so it's funny you should say that, and this isn't like maybe, strictly speaking, governance, but this is a real, tangible indication of what it could allow us to do. If you look at tools like GitHub, where you can now take things like pull requests, code reviews and that type of what was traditionally a manual task to do quality assurance, and you can delegate that to agents. Now you can have those government steps, governance steps, apologies that are required, but have them scale at a level that matches the input that's that's increasing so, so code reviews have historically been a thing that slows development teams down quite a lot. There's a little bit of hygiene and good practices there that can help. But actually, number of lines of code changed, and if you want me to review these things effectively, there is a certain amount of time that you need to take away from me as an engineer to do that those things are starting to be automated in a way that allows that governance step to happen much more quickly and much more autonomously, if you've got The good standards. And that is where people are now starting to fill gaps. No longer can we have a group of senior engineers with tribal knowledge of what good looks like. We need to document what good looks like. We need to codify that. We need to put it into a location where it can be used effectively by the automation, and then improve the governance to be more automated. So a clearly

Dave West  22:09  
articulated, transparent definition of Done. Yeah. I mean it just It blows my mind that you know people on one hand are saying, perhaps we don't need scrum anymore. And I'm going to get onto my box for a moment here, because AI is increased productivity and Scrum is all about teams. And you know, the teams are much smaller, therefore. But ultimately, those practices, what transparent means, may be different. Now, who does the work will definitely be different. How you execute a daily scrum might happen a lot more frequently because you've got agents that you're looking for bottlenecks, looking for opportunity to replan based on experience, etc, but all these principles are still very, very valid.

Tony Hinkley  22:57  
Yeah, it's one of the things that I've always liked about Scrum, which is why I continue to enjoy being part of the community, is that it's that framework that we can use. It's those principles to guide us, but it still relies on you as a practitioner to use them appropriately, to operate within those those guardrails of the framework to get the best outcome for your situation. So if you were a team that's historically been six plus or minus three, and the distribution of predominantly developers, rather than testers and other other roles in the team, two week cadences, you know all the stereotypical scenes that we see in in scrum adoption, you now need to have the knowledge to be able to pivot and refine and improve your implementation of Scrum to meet the new constraints. Yeah, and if you don't look at things like what is a definition of done now, where the person that is reviewing the definition of done is an AI agent versus what it looked like six months ago, when that was a team of people, then you're going to have a suboptimal implementation, and you're going to get very little improvement in your total flow through your team because of that.

Dave West  24:27  
So it's interesting. I was talking to an organization a few days ago, and they're really into Scrum. The guy I was talking to is really into Scrum, maybe a little bit too much, if you can be such a thing, but he'd got, he'd got his DOD. He'd call it dog, dog DoD dog that basically he'd created an agent that worked with other agents that, basically he'd got. So there was a code quality thing. There was a volume testing. Security Testing scale. They got all these like and he hadn't built all the agents to be fair. But what he did was he made sure his DoD dog, as he called it, was basically inspecting those things, and he continuously asked that DOD dog to and it was just really, it was kind of cool. It's different, yeah, very different from how you he worked before. But it was, it was a really interesting implementation of DOD actually sort of manifested it as a as an agent, and and all it did was inspect other agents and asked for their results, and he built a dashboard with it, and he did all sorts of things that were really, really interesting. I thought that was really, really cool. But you're right. It's I was always what. I met you about 10 years ago, Tony, and it's like personal thing here, and I was always shocked at somebody so techie, because you are a bit of a geek. I don't mean that in a actually. I mean that in a positive way, but you are very geeky. You always, you know, 15 laptops are on, piled on your desk, and you've got, like, you know, wires coming out of things, I don't know what, and I was always shocked by you loving Scrum, particularly at that point scrum was, there was lot of agile coaching and all of that stuff. But what you did through these 10 years is you've reminded me of the discipline and that and how that can really amplify the technical nurse of you by giving that sort of more disciplined approach to how you approach work and step back from it. I always think that was really refreshing. And I think we're at a stage now with AI that we need that I I'm using Claude code at the moment and playing with it, trying to build a little app for me and my son. And it's like being it's like being 12 again, using basic you know, I guess I was 13, but whatever age it's, it, I've never been but you know, what's funny is, without the discipline, I'm falling into some of the old traps. It goes horribly wrong. I've just deleted it. Ended up deleting something that I was surprised about. I had to get my son to do something again because I managed to destroy the entire thing that we'd done before. Discipline is really important,

Tony Hinkley  27:23  
yeah, and, and it just sort of to come back on on that discipline piece and the conversation we had earlier about reducing headcount versus increasing top line growth, one of the, one of the trends, I guess, that I've spotted is people who are specialists, who have experience, are getting significantly better results out of this AI tooling than generalists using that tooling or letting the tooling work without supervision. So as somebody who's been programming since I was about 12 years old, and has been paid for it for a lot longer than that. I found that the reason why I am so productive using these tools is because I set clear expectations of what I want the tool to be doing. I follow through on things like DoD dog is a great example. We've been asking teams to do exactly what you just described since 2001 but we've never been able to control people in the way that we can control technology. So all we're doing there is saying this has always been a good idea, but now I have a team of agents who might actually do what I've asked them to do in checking against the definition of done before they move on to the next stage of the life cycle. And so when you're looking at your teams, for people that are listening to this, if you're having to make difficult decisions around what your team structure is going to look like moving forward. Think about how useful and valuable that experience is probably more so now than it was in the past, because there's some of the other constraints around how many keystrokes and curly braces can one person produce in a day, the knowledge the ability to ground, the ability to mid, steer and correct an AI that's starting to hallucinate in one direction is going to be an important skill. And I would say that those skills are probably not universally available across your current team. So think about those, those things.

Dave West  29:43  
Yeah, and it's interesting that I do think skills are going to change what we think is important. I do believe, as you rightly said, that those, those first principles, the reason why you've been so successful as a software engineer is because you've always had. A focus on those first principles, you know, whether it's clean code, whether it's test driven, BDD, TDD, whether it's, you know, sort of incremental development, meaning, build little chunks, test it, grow it, etc. And yes, that might have been influenced because three and a half inch disks only could take so much data. But that doesn't matter. You know, you've, you've got those disciplines, I think. And I think it was, I think it was the guy that created Claude code actually was on Lenny's podcast, and he said when, when he's hiring, he doesn't really care about AI skills as much, but he really cares deeply about the those first principle skills, meaning, sort of like the ability to abstract the problem, the ability to sort of ask why, the ability to evaluate something can an appreciation for what good looks like, and the steps necessary to get to good. Because what's interesting, you know, with where the bottlenecks are, we're going to be less about coding and more about reviewing, and where that review abstraction lives as well is going to be different. We're going to be reviewing outcomes and out. We're not going to be reviewing outputs so much. We'll be reviewing outcomes. We'll be looking at making sure that the steps have been followed. It's a it's an interesting change to what we do in product delivery

Tony Hinkley  31:35  
and and I've been using a phrase recently, which is that effectively, context is the new currency that need to be able to bring to bear. Because effectively, without grounding these tools in clear intent, good standards and organizational contexts, because what's relevant for one company using at all is very different to another company, then you can't ever really exploit the true value of AI. And so when you look at what obviously my company is part owned by Microsoft, so I have a particular interest in the Microsoft tool chain, but they have really doubled down on what they're calling their intelligence layer, so the Microsoft IQ stack, and I won't go into what it is necessarily, but I'll talk about why it's important is because it's all about bringing the data and the context of your organization into a layer that AI can reach into and use to understand the boundaries that it has to work with and the data that it should be using, and that's a really important piece that's missing from some of the competition at the moment, especially in business use cases. So we talked about product development teams who are using these tools to build products. But the next step of that is teams that are building these capabilities into the products that they're then giving to their customers, and all of those tools require access to business data, access to knowledge bases, access to org structures and team profiles, And that information right now is going to be turning into a new currency to make those tools as impactful and effective as possible. Bigger context, windows, more tokens, smarter models. Is all really good, but actually what I've seen is context when you have that and it's and it's available and it's well controlled, we'll come on to governance and risk and stuff in a second, I think. But when you have all of that, you can use a cheaper, smaller, faster model and get equally good or better outcomes from these tools, but only if you give it the right data to work with. And that's where I've seen a lot of teams stumbling, is they use the latest version of anthropics model, you know, their Opus or their sonnet models, or they take the latest Codex and GPT five, three model this, this podcast will date so quickly that I've

Dave West  34:27  
okay

Tony Hinkley  34:29  
and not actually get better outcomes or move any faster, and they don't have the context.

Dave West  34:36  
I think, I think that's I was thinking that the other day. So I was on a we're recording another podcast with another guy, and he was talking about how he's helped his business organizations. They're in commercial, real estate, residential, letting commercial. And there's like a massive backlog, humans are, technology part of the organization, a massive backlog of things. Is that all of these, and it's across many countries, so you can imagine the backlog of, you know, legislation takes precedent to any nice features. So just something as simple as, you know, a Polish real estate company that has, like, I know, 70 properties, wants to look at their insurance risk profile, and they and it's all in systems they all have, and they couldn't do it historically. They just couldn't, you know, that was no feature that checked all of that stuff. And they rapidly did that. But what he highlighted was the fact that he had to start building systems that were open to llms so that they could do that. And he said the only way they could do that was with context, actually presenting not just almost a this is how we think people use this system process. You had to almost describe all of that and not not just data and metadata, but also process data flow, data usage, data, you know, that kind of stuff. And then when you could do that suddenly, then, then some somebody writing a quick prompt in Poland could get all of that information out. I thought that was really interesting. I hadn't really thought about that, but context is king, and it's a different kind of user, and it's a different interface to how we're going to be using these systems. We have to present that to these systems to be able to be used in that way. I don't know if I've described it well, but it sort of blew my mind a minute, for a minute.

Tony Hinkley  36:41  
Tony, yeah, I worked with a company. It's got to be going back pre pandemic now, and they held a huge amount of data around property, in fact, probably the largest data set of UK property data that existed, and even pre pandemic, before the tooling got to the point that it's at now, they were looking at ways to provide real estate insurance companies, etc, investment companies, with answers to questions about the risk in their portfolio, the value of their portfolio, the predicted performance of their portfolio, but that's because they really understood the data. Yeah, and that's where a lot of companies now are finding their investment is not going into AI tools that sit over the top of data. Those things are becoming more available, more commoditized, and there's a lot of options in the in the market, but looking internally at what data do they have? Have they got a good model around that data so that it's understandable? Can they describe that model in a way that makes it easy for these tools to consume? And again, we're back to this intelligence layer concept. If those things right, then actually the tools that you can then plug into it become much more powerful and potent.

Dave West  38:07  
Yeah, I was talking to another large organization, and they're talking about this idea of a context library, and you're sort of managing it, and talking about the product owners have to be context owners, because ultimately, it's not just access to the product, the system, or whatever it is that they're providing to their users. They're also having to provide that context to both the users of that system, for the execution that you know, that those Polish real estate people got, but also the developers of it. So you guys are like, oh, and on text changes as well. And managing it, we do, we shove it in GitHub. Do we or, you know, yeah, there's a whole new set of tools that are probably going to come out to manage context well.

Tony Hinkley  39:00  
And the thing that that then causes the knock on effect of that is, if that data is so critical, so valuable, how do we make sure that it's the right data, so quality, that it's secure, so that where we're exposing that value, we're only exposing it to the right people and the right consumers of it, and can we actually stop things like abuse data loss and hallucination by measuring how that data is being used and the answers that are being given off the back of it? So I say governance is a dirty word for some people, but actually that becomes a real challenge for lots of companies because they've got lots of data which they've never really structured, effectively, categorized, classified, segregated, secured, audited, all of these, all of these things that for. All under the big umbrella of governance and management that are now a massive intellectual asset and business critical asset for all of these new tools, garbage in, garbage out, has been floating around as a phrase since you know, I can remember, never has it been more true or more potentially damaging to put garbage in is going to, is going to drive a really bad outcome on the other side of it.

Dave West  40:28  
And we've always relied on the fact that there's a layer of disconnect from data, you know, process, techno software, apps and stuff, and there's a layer of people managing those apps that have sort of present hidden a lot of these frailties in our discipline around the use of data, information and context, we've relied on human beings doing that, which is, you know, been fine, and the fact that most of these Organizations rarely release software, really, I hate to say that, so we don't change that frequently as well. But you can imagine that if an organization, a big, traditional organization, with all of its huge, uncompetitive capabilities, meaning they own a lot of the market. They have access to massive amounts of capital. They if they could leverage that in a you could which brings me back to the laying off software engineers in these organizations. It's very disappointing, because do they not see the potential that they have, and you could do amazing things and potentially have a massive unfair competitive advantage. Yeah.

Tony Hinkley  41:46  
I mean, I'll put a stake in the ground and say that, you know, we are hiring despite all of the messaging in the marketplace around the space shrinking and the headcount need not needing to be as significant anymore. I look at other organizations who are what I would consider forward thinking, moving towards what Microsoft are describing as the frontier firm, in terms of their adoption of AI, and actually I see them hiring for skill upsizing some of their teams. And yes, there has been layoffs in the market, and I think that that can cause a lot of uncertainty with people that are doing similar roles. Those same companies that have issued all of those layoff notifications are also hiring. They're just recognizing that the composition of the workforce is going to be slightly different. The size might be equivalent, but the composition is different. And so there's people leaving, but there's also new people coming in.

Dave West  42:51  
Oh, I'm we're running out of time, Tony, and I could talk to you for days on this, and that might be a really good segue to sort of like the message to our listeners as we leave this you're, you know, the CTO of one of the most technical first organizations in the world. You're, you have an opportunity to see all sorts of things going on. What advice would you give our listeners who may be a feeling a little scared in with with AI, you know that there are layoffs, there is a big change happening. What advice would you give them as we leave this podcast today?

Tony Hinkley  43:29  
So I would suggest that people look at how they can embrace the technology, because there's a phrase that's been floating around for quite some time, is that AI will not take your job. Somebody else who knows how to use AI will take your job I talked about earlier, like the specialists with the tools perform better than generalists with those same tools. So help yourself to become more fluent in what the tools are and how they work and what you can achieve with them. And I think you'll be a massively valuable asset, if not to your current organization, to your next organization, and, and I think for leaders out there who are looking at their teams and thinking, How do I respond to this? I would say, look at how you have got new opportunities to strengthen certain areas of your organization that historically you haven't had to and focus your change effort there better context, better quality, better standards, better better guardrails, rather than looking at minimizing, necessarily The headcount around development. Because if you do those things, then you have a top line growth story that you can start to talk about, rather than just a cost reduction race to the bottom exercise. I've got one one final point, if I may. And I talked to my UK general manager about this. The other day, I think this is a really interesting time for mid market organizations, so not your big enterprises, your mid market companies, because these tools are really leveling the playing field around the products and services. You can deliver a substantially lower price point in terms of people and headcount capital, and now is a really great opportunity to start to disrupt if you're a mid market company who gets really good at using these tools effectively, you can punch well above your weight now in the products that you can deliver to market.

Dave West  45:41  
100% agree. I think it's an amplifier that we've never, we haven't seen since the printing press, probably, or maybe the PC, but probably the printing press, you know, and that that obviously revolutionized the world. You know, we had the Reformation. We have all sorts of Age of Enlightenment. We had all sorts of things. Even I wasn't around then. Tony, to be fair. Tony, thank you for taking the time today. I want you back in, you know, maybe in six months or maybe before, maybe before. Now we have to, because of the rate of change. There's so many more questions I have. I really do appreciate you sharing your insights with us, and I'm sure our listeners do as well. So thank you for taking the time today. Thanks for having me, Dave, that's great, okay, and thank you for listening today. Scrum, to all community podcast today, we were really, really lucky Tony Hinkley, CTO Avanade in the UK, came and joined us and talked a lot about how he's seeing both the impact of AI technology at Avanade, which is a white collar consulting company, and and the opportunity it's providing, and how it's changing where the bottlenecks are in the system. I really liked his use of the systems thinking metaphor and about bottlenecks and about and about constraints and how AI potentially can fundamentally change all of that for the future of product delivery, but the future of all potential knowledge work. If you liked what you heard, please subscribe, share with friends, and, of course, come back. Listen some more. I'm lucky enough to have a variety of guests talking about everything in the area of professional Scrum Product thinking AI and, Of course, agile. So thank you and Scrum on foreign you.

Transcribed by https://otter.ai