BCR 137 - Paul Glover and Automated Agile

Lee
Today I'm talking with Paul Glover, from a company called Brando in the UK. And we're discussing automated agile support. Can you give us the elevator pitch about automated agile plays?

Paul
Yeah. So automated agile is really a way of thinking. And all it really is, is on its understanding the AI can utilize you can utilize AI throughout the product delivery lifecycle. And in doing so, add productivity. But there isn't really a strong understanding of how to do that all the way through. There's individual use cases which are important. But it's about leveraging that rich context to produce, you know, great outputs. So what automated agile does is think about the whole product delivery lifecycle, and how you can get the most out of it utilizing AI.

00:01:00:17 - 00:01:26:13
Unknown
So then if we look at the origin story of, automated agile, what challenges did you have bringing it to where it is now? Well, I think it's it's about, understanding for people. I think because it moves so quickly and that's overwhelming. So, what automated agile doesn't try and do is they use this tool, use this process, use this way of thinking.

00:01:26:15 - 00:01:48:02
Unknown
What it does do is come up with, a way of thinking about that, to understand what drives the best performance out of each of these steps, and how do we create a structure that allows you to put new tools and new processes in? And I think as a concept, that's something that, you know, people, are kind of thinking about that in a use case specific, you know, sense.

00:01:48:02 - 00:02:06:21
Unknown
Now, you see a lot of tools out there that will solve individual problems. But in, in terms of a whole workflow, you know, there are limited people out there really thinking about that and helping businesses. And that's really what this is automated agile about, rather than being a product is more around how do we help businesses understand what the next step forward is?

00:02:06:21 - 00:02:31:20
Unknown
Because the Florida Neath them's changing at rapid pace. Yeah. Isn't it? Yeah. Star Trek and I'm thinking the holodeck in particular is there any similarity between the holodeck and automated agile. Yeah. Well, the holodeck theory is something I like to talk about a lot, you know? And and it probably rings true with, like, nerds and and people who watched Star Trek, but ultimately, it's all about, you know, you have an idea.

00:02:32:12 - 00:02:57:05
Unknown
And then you have your product, and in between those two, there's a lot of friction, that, you know, is generally what I do for a living. And then that's, you know, what the product of your life cycle is? Where? It's in a in a holodeck in Star Trek. You walk into a room and you say, you know, give me, you know, 1920s London and it will and I that's really how software should be, you know, you should be able to speak to a tool and say, this is the product I want, and it shows that you straight away and then, well, I'll change that.

00:02:57:07 - 00:03:20:12
Unknown
I'll update it. So in our tests that we've done, utilize in a, you know, a process with automated agile at its heart, we've managed to create prototypes, you know, instead of 90 minutes, you know, 75 minutes for one type of prototype. And what that does is put something in front of a customer and give them the opportunity to, see whether or not they want to make any changes to it.

00:03:20:14 - 00:03:40:22
Unknown
And that's what we try to do. That's why we call it the holodeck theater. It's that how do we close down that friction as much as humanly possible to make the process, of customer feedback, and customer improvement, you know, as, as, as, immediate as possible because, you know, the, the primary measure progress is working software.

00:03:40:22 - 00:04:10:14
Unknown
Right? So it's it's about how do we, keep those phases small and keep the a question that we're asking AI to complete as simple as possible. But yet make that forward progress more rapidly than we would be able to, using traditional agile methodologies. It reminds me then, because of this, you know, talking to the, the engine rather than, prompting it by typing, it's, do you remember you probably do a couple of years ago.

00:04:10:16 - 00:04:31:08
Unknown
Yeah. Up until maybe last year, there was a big push by all of the senior tech people, you know, in Silicon Valley saying, you know, everyone will have to learn how to prompt write prompts and code. But, there was one CEO who said, no, you want, within 12 months, you won't have to do any of that.

00:04:31:08 - 00:04:56:18
Unknown
The AI will do it for you. And I'm guessing that's, where automated agile is heading is that you don't need those technical skills. You just talk to your engine. And, it creates from there. Is that right? I've. I've got that right. I'm a business analyst originally. And business analysis. Yeah. I think comes from the place of anybody could do it.

00:04:56:20 - 00:05:15:01
Unknown
But you hire a business analyst because they've got the skills to do it in a way which is lean, and not wasteful. And I think that that's all, prompt engineering is for me. I think anybody could go through 30 cycles of asking some, you know, an AI to produce something. And it turns out it's not quite right.

00:05:15:15 - 00:05:34:05
Unknown
They could make sure someone has got that precision of language so that when they ask it a question, you get in a very specific answer at the back of it. And, you know, I like people who help it help itself. So I think, yes, you know, you certainly could get, the prompt engineering piece to be done, by AI.

00:05:34:05 - 00:05:50:05
Unknown
And at some point in the future, there will be a level of understanding of the AI. So it will help somewhat with that. But you're never going to get to the point where precision of language doesn't matter. So I think, you know, maybe prompt engineering will change as a methodology about that precision is always going to be necessary.

00:05:50:07 - 00:06:09:23
Unknown
Yeah, yeah. I mean, certainly with with the work that I do, if I write a prompt, and it could be a, you know, a large prompt if I, you know, type one up, I will then ask the, the alum that I'm talking with, to, reshape that prompt to make it tighter and more accessible to that, that engine.

00:06:10:10 - 00:06:27:07
Unknown
And yeah, sometimes you just look at it and you go, that's a third the size of what I typed up. And it's, you know, and it works brilliantly. Yeah. And there's a whole industry, isn't there, around prompt tuning where, you know, you you might be a great create a great answer from a prompt. But is it token efficient?

00:06:27:09 - 00:06:47:00
Unknown
You know, and I think once we get to that point. But, you know, really what we're doing is we're trying to take work away, from, the delivery life cycle to reproduce that productivity. And we, you know, we're seeing maybe 30%, increasing productivity off the back of this kind of methodology and thinking, so, you know, when you're doing that, you can you can pull that into creativity, really.

00:06:47:18 - 00:07:13:08
Unknown
But, you know, in terms of, prompt engineering, definitely. You can utilize AI to help you with that. And, that's what we think across the whole product delivery lifecycle. It's just how how to get the most out of it. What benefit can I get right now for that. But then leads me to think about context. And I'd be interested to know why context is so critical in, in, AI in particular in automated agile.

00:07:13:08 - 00:07:45:18
Unknown
Why is it why does it become smarter? Yeah. Well, it's about making sure you ask it that very specific question again. So I think in terms of, Artemis fragile, you know, one of the key fundamentals of it is you build a series of documentation which accurately and fully describes the products. Now, that would be contrary to, you know, you should use agile methodologies which say that you can't really think, understand what you want at the beginning of a projects because, who does, you know, you want to iterate, once you gather customer feedback.

00:07:45:18 - 00:08:18:04
Unknown
Well, what automated agile tries to do, is it creates that 100%, understanding, of where your product is right now based on the information you've got right now. So you can build that product to 100% level and get that immediate feedback. Rather than spending weeks on your document, you know, you want to make sure that it can be done inside the call that you have the customer straight away, you know, and that's where, automated agile kind of flips some of these traditional methodologies on its head where, you know, documentation might not be as important as the people in the understanding in the room.

00:08:18:06 - 00:08:44:20
Unknown
Well, when the reason for that is because of all the, time that you have to spend producing that documentation, it makes it, a wasteful action. Whereas when you're utilizing, things like AI processes, it'll produce that documentation immediately. I mean, you can check immediately. And then when you're asking the AI a question in the future, it's got all of this information that it utilizes alongside what you're saying to it, to then give you that better answer.

00:08:44:20 - 00:09:06:05
Unknown
So it's this is where some of the prompt engineering, methodology is. And we're talking like, you know, you talk about single shot problems and stuff like that. And each of those will give you an answer. And that answer might save you 30%, 40% of the work. So what automated agile tries to do is build structures of information so that instead of that, 30% were aiming for 60% or 70%.

00:09:06:05 - 00:09:28:17
Unknown
By giving that additional context of what we're trying to do to the AI when it produces that answer. So then let's let's bring it back into the front from, a conceptual theoretical into the real world. Could you, could you walk us through a typical day for, a team that's using, automated agile? Yeah. So I would say there's no team using automated agile right now in anger.

00:09:29:02 - 00:09:50:22
Unknown
It's at the experimental, phase. So we've done, three tests on it so far. So the first test we did, was with a team of ten consultants. What we did with that is we tried to produce, a prototype, and that was very simply, ask it to produce, requirements documents, technical information, based on the feedback that we gave it.

00:09:51:15 - 00:10:10:13
Unknown
And what that allowed us to do is produce a set of documents, which we can then add to what we call a ragged database, which is retrieval augments each generation. Now, in this scenario, we were utilizing Claude. Claude as a, is a ChatGPT type of a model. And and what we try to do, is utilize its project's functionality.

00:10:11:02 - 00:10:29:14
Unknown
And that allows us to essentially add information into a database so that when we ask it future questions, it's got the context and information to give us an answer. So what we did is we produced, a product design and those requirements and technical documentation, we added it all into the RAC database so that we're asking it future questions.

00:10:29:16 - 00:10:52:20
Unknown
It knows what to do. And then we ask it to produce a certain prototype, and then try to fix that. And we did that in about 75 minutes. That was stage one. Step two was like kind of build upon that because what it did is produce a lot of code, which, you know, isn't necessarily very readable. So the next stage was about, well, if we're asking it to produce these documents, I'm never asking it to produce code to what standards should it be doing now?

00:10:52:22 - 00:11:10:24
Unknown
And one of the key things I see people talk about is we want it produces code that nobody can read. For example, is something I've heard frequently. Well, I'll only do that if you don't sell it not to. So it's about producing, a coding standard that it's happy to work to, and in an automated agile, what we want to do is make it highly modular.

00:11:11:13 - 00:11:28:06
Unknown
And, each individual code file a single responsibility. And we do that because what we want to do is ask the AI the simplest possible question we can do, and then give it the best possible context we can do to answer that question. Well, so the second version of that, we did it with the same set of ten consultants.

00:11:28:19 - 00:11:49:21
Unknown
And that produced, it took us a little bit longer, not 15 minutes. It was an hour and a half to produce a working prototype, but instead of it being a single Python file that did everything, it was a series of modular files. All were built to a set standard. And then what we could utilize that with then is there are code tools such as, you know, GitHub copilot common one book cursor is one that we're utilizing.

00:11:50:22 - 00:12:10:08
Unknown
And what that cursor allows you to do is to take those files, put them inside, of cursor, and then put these files in a chat with cursor and say just build this and it'll just build it straight off the bat for you. Now, it was really key. And, in our second test that we saw is a bug came up.

00:12:10:22 - 00:12:28:00
Unknown
And when you get that book, you try and understand what an automated Oklahoma does that book come from? Is it a process problem? Is it a requirements problem? And we found that it was a problem in the process because it was an error that wasn't written in our code documentation. So what we then do is update the code documentation.

00:12:28:00 - 00:12:51:01
Unknown
And next time we run that process, that code will exist. So it becomes self-improving, gives you its own feedback and allows you to respond to that feedback. And that's what you lift it from 40 to 50% of time saved and productivity increased. Open up and up by continuously improving based on that feedback. That's I mean, the numbers that you've, you've given me so far, but that's impressive.

00:12:51:03 - 00:13:09:24
Unknown
I mean, you know, different looks. Three, four months ago there was a big Google or about, ChatGPT. One and, you know, how it could reason and how it could learn from itself. And everyone was, you know, raving about that. But this quarter, in a way, takes it to the next level. Is that right?

00:13:10:01 - 00:13:44:20
Unknown
Yeah, it's trying to it's trying to build a process which is self-improving. And you understand the weaknesses in that process. Based on the feedback that you're getting from your customer and from bugs, effectively bugs the feedback. And because that process is machine built and essentially built on top of statistical models, it allows you to do things like probability prompt in which is you ask a question instead of asking the one question, and get an answer, you'll ask exactly the same question 20 times, and then you coalesce those answers down.

00:13:44:22 - 00:14:06:04
Unknown
And that gives you that kind of steady way of working, which you can then iterate, and improve upon. So if you find that your, code is being produced in a way which is unreadable, then give it a feedback, improve your documentation so that it doesn't get produced in that way. Test out successful that is and keep iterating that until it's producing something that you comfortable with.

00:14:06:06 - 00:14:20:16
Unknown
You know, it's about taking responsibility for making sure that the AI produces the right data by giving it the tools to do so. And if you think about it, that's no different than humans. You know, we wouldn't get a group of humans and put them together into a process and just expect everything to work without some training and development.

00:14:20:20 - 00:14:28:12
Unknown
How often I've been in enough companies to do this, but you're not supposed to.

00:14:28:14 - 00:14:54:23
Unknown
The elephant in the room at the moment with AI is, if I'm being asked to do all the work, if AI is picking up, the slack and doing a whole lot of tasks, what happens to the people who, doing it before I have a right sized as, companies have always said for decades, or is this some other way of tapping into the brains and experience of those people?

00:14:55:00 - 00:15:13:02
Unknown
Yeah, I think this is an important question, and I think it's one I get asked a lot by people. Look at CERN, you know, is I going to come to my job? And I think it will be companies where they will utilize some of these, processes in place and they will lose 30% of people, and I think that will be, the wrong thing to do.

00:15:13:04 - 00:15:38:06
Unknown
Because really, if you look at what's happened with tech, you know, when Stack Overflow came out, when the internet came out, you know, these kind of jobs explode because what this is doing is, is increasing productivity. And when you increase productivity, you reducing the cost of implementing change effectively. And that change often has value. And every organization I have worked in has always had a backlog of value that she's three times as long as the people that have got to deliver it.

00:15:38:08 - 00:15:53:22
Unknown
So I think you will find that companies will cut people and their competitors that don't, and use that extra productivity to innovate, and produce better experiences for the customer. I mean, how many, how many websites have you been on where it's been an experience, if that all this is really good compared with ones that you've hated, you know.

00:15:53:22 - 00:16:12:14
Unknown
Well, you've now got more opportunity to be able to challenge, those, you know, original ways of working and, and fix those problems. So I think this is going to cause an explosion, in, technology jobs, they might not be exactly the same jobs because what automated agile does, he supports everybody in each one of their roles.

00:16:12:14 - 00:16:33:00
Unknown
All of their roles are still important. They still, fulfill a specific need, and they'll just be able to produce them. I'll do them, more efficiently. So I think you might see some changes in jobs, but ultimately you're going to get, you know, much, bigger tech industry, I think, at the back of this. Totally agree with it.

00:16:33:00 - 00:16:51:09
Unknown
I mean, one of the things that we do at Amigos is that we're very hot on on people focused AI so that, you know, if we go into a company, talk to the CEO, and you know, the, the elephant in the room, you know, are people going to go? And we always say to them, no, don't get rid of your people.

00:16:51:09 - 00:17:13:07
Unknown
Keep them. Let them be more creative. Give them, you know, free rein to come up with a whole lot of stuff and listen to the youngest in the group, you know, set up teams, listen to the youngest rather than that the highest paid person in the room. And they and they will create an environment that will make your business sing because it's taking you in directions you wouldn't normally think of.

00:17:13:07 - 00:17:35:23
Unknown
So, yeah, like you, I think, yeah. Keeping you staff and letting them be creative, is far more, intelligent thing to do. I yeah, I think so, because like anybody who's a tech professional who's worked in an organization could sit there and write to you about, like, themselves, things that they can see and improve inside the business, you know, help better ways of helping people work.

00:17:36:00 - 00:18:11:10
Unknown
And they might get to a point where, you know, the business is so efficient that, they don't need as many people, anymore. But we're years away from that, I think, in the vast majority of businesses. So, yeah, I don't nothing to worry about. I would say in the short to medium term go back to the the late 90s and the early 2000s, before social media came around in and in 2004, I was building websites and using, you know, Macromedia Dreamweaver, and, you know, the websites like I will be talking websites a couple of minutes against the websites were pretty crude and pretty horrible.

00:18:11:12 - 00:18:38:10
Unknown
And then CSS was introduced and suddenly websites went into a usability sort of overdrive and they became fabulously, they looked gorgeous. They really were ergonomically simple for the user. And the whole industry exploded in terms of people suddenly getting in and and making beautiful websites. And, and I think I was going to do exactly the same thing. It's not going to become this, this monster that eats the population.

00:18:38:10 - 00:18:55:24
Unknown
But it'll be something that just takes business and humanity to a new level. If you work in the tech industry, you're probably right. You know, there are other industries that will be less, less, nicely handled. You know, if you work at the call center in New Street, you're in trouble. Yeah. But, like, there are some industries that are going to get decimated by it.

00:18:55:24 - 00:19:18:03
Unknown
But I think, you know, certainly in the tech industry, the, you know, people who work in the tech industry are going to change the world, I think. And and if we can do that twice as fast or more than then, all the better. Let's talk about automated agile and, you know, businesses and getting started with, with automated agile.

00:19:18:07 - 00:19:42:17
Unknown
What's the first step for an organization who want to move into that space? What should they do? I think the same thing as you would do with any technology or any methodology in that set up, something like a working group or a community of interest that, assesses each of these new methodologies or ways of working and process changes and gives you an objective view of whether it's success or not.

00:19:42:19 - 00:20:02:19
Unknown
So that would be the first thing I would do. An automated agile, you know, it's a series of steps you can take, in that process. But I wouldn't do any of those steps without verifying whether it works for you. It was an organization. So step one build a structure that allows you to verify. And then step two, is around making sure your data is right as an organization.

00:20:02:19 - 00:20:24:21
Unknown
So, any AI led process thrives on well-constructed, easy to read data. And so if you're making sure that you're utilizing tools like Jira and as your DevOps, for example, if you know that common tools in organizations are having that structured in the right way so you can get the right data out of it, it's important, especially as those tools, are coming into this new age.

00:20:24:21 - 00:20:44:12
Unknown
Now, Gia has got agents inside it that allow you to do some of the automated agile process utilizing Azure DevOps or something similar called modern requirements, which is a plugin that does something along those lines. So yeah, get your data right. Build a process that allows you to judge whether what you're doing is the right thing or not.

00:20:44:24 - 00:21:01:16
Unknown
And then either find yourself a partner who knows, about these things strongly or, empower your people to learn about it. Because it moves really, really quickly. It's almost a full time job keeping up with what's going on. You know, there's a lot of information out there on the internet, but. So how do you get it?

00:21:01:16 - 00:21:27:02
Unknown
Make it work for your business? Yeah, absolutely. I mean, and and, come back to the point is, you know, if you're putting teams together to, into, so either design or implement, some of the business ideas that own, you know, directions we should be going is listen to the youngest members of the team, you know, just don't discount them because they're not, you know, the highest paid person in the room because they their ideas will come out of left field.

00:21:27:03 - 00:21:51:18
Unknown
And you will suddenly, if you know, the old people will have to go. Hang on. That's not a good idea. Yeah. So. Okay. What am I, I still can't we I think I think, you know, it's good. It's good to listen to all levels, you know, to these kind of things. And, and even, one of the, one of the great things about AI is some we call system zero thinking, which is essentially you can utilize it as a hard drive that sits outside your brain.

00:21:51:20 - 00:22:15:10
Unknown
So don't think that you because you've got no knowledge and things like machine learning or things like that, that you necessarily are at a disadvantage because what you can utilize is AI to get you to that 80% knowledge in those kind of things. And then you can understand whether it's a reasonable approach to do as a business, and then you bring in the specialists to do it, you know, so I think you can utilize AI is a force multiplier even in the thinking of how to apply AI.

00:22:15:12 - 00:22:37:04
Unknown
Most definitely. All right. One last question for you, Paul. Where do you think, automated agile will be in, say, three years time? Yeah. Well, I mean, automate agile. I mean, it has no intrinsic value. It's just a brand name for for how do you use AI in the best possible way? To, to have a PDL C, which is productive.

00:22:37:16 - 00:23:01:13
Unknown
So I see, that thinking of, you know, that I am pdc to be something that strong investment in right now. I think you look at consistently more and more productive. And I think what I see automated agile in is trying to keep up with that and, and putting it into that kind of, way of working that people can understand and apply in their own business.

00:23:02:10 - 00:23:25:00
Unknown
So what what I'm hoping automated agile becomes is something that people collaborate on, you know, in different countries, the different ways, ways of working so that we can all, sift through quite a lot of the, you know, the the noise that's out there really and really get to the value, in a, in a way which, people can collaborate on together and improve.

00:23:25:02 - 00:23:46:02
Unknown
I, I've got a, well, a friend that I've known since my early 20s is a mechanic. And, you know, over the years, he said, everyone thinks a mechanic, you know, blue overalls and, you know, wrenches and screwdrivers and spanners. And over the years, he's. But he's been blown away by how technology has come into his industry.

00:23:46:04 - 00:24:20:01
Unknown
And now he doesn't get his hands dirty because everything is so, you know, tightly wound together. He just plugs in a little box underneath the, steering wheel, and that little black box just tells me everything he needs to know about everything in the car. Not just the engine, but the electrics, everything. Every component. And so it just saves him hours and hours of, rather than, you know, undoing stuff, having a look at it, putting it back, you know, just instant and and that's happened quietly with no one realizing it's going on except for the mechanics.

00:24:20:07 - 00:24:43:16
Unknown
And I think unfortunately, you're right. You know, I at the moment has this big, bubble. Oh. You know, there's the name around it, whole industry of people hyping it up. But, you know, in two years time, when it's all calmed down, people will just see, actually, this is pretty bloody productive. And, you know, I don't know what we were scared about.

00:24:43:18 - 00:25:05:03
Unknown
Yeah. And we want to change very much what it is we do. It'll just be quicker and more effective. And, and I think, you know, one of the, slides, I have a a and a deck. I show people when I'm talking about AIS, maths teachers with big placards, saying the no to calculators, you know, and I think I'm old enough to remember that.

00:25:05:05 - 00:25:24:22
Unknown
Yeah, yeah. And, and I think, you know, and that's where that's where we're at. It's the same, like, Luddite Luddites conversation again is, you know, is and yes, it'll be disruptive. Yes, people will lose the jobs. But there'll be new jobs that pop up, and those jobs work building in times tables, you know, and we'll be we'll be taking away a lot of effort.

00:25:24:22 - 00:26:02:08
Unknown
And especially when robotics comes along. But there'll always be a need for human creativity. I remember, just going back three, maybe four years ago. Marks and Spencers, owning Marks Spencer in England were putting their staff through courses. I mean, s is expense on, coding and things like that and I can't remember I think it was before AI, but they were putting the staff through various coding and you could get a qualification, a recognized qualification, in, you know, computer programing or whatever.

00:26:03:00 - 00:26:22:24
Unknown
So that it because, I mean, I saw that they were going to, you know, people were going to, you know, move out and they were going to shrink the workforce and everything. And so they were preparing their people as much as they possibly could as a good employer to, you know, make sure they had something to go to after, you know, leaving a mess, which I thought is brilliant.

00:26:22:24 - 00:26:41:04
Unknown
And if companies can do that here with the AI rush, then, more power to them. Yeah. And I think that kind of retraining people who are already advocates for your business and already know your business very well into the kind of role that adds value to that business in the new age. It's a powerful thing to do.

00:26:41:06 - 00:27:03:11
Unknown
And, you know, more and more organizations are working in that way. But I think it's something that should be, you know, government led, in terms of supporting organizations in doing that because, you know, if we utilize AI well, it'll change things quickly. And for the better. But in doing so, there'll be some friction around people's roles.

00:27:03:21 - 00:27:20:16
Unknown
And we need to make sure we support them into the new roles I creates. And the private sector will cover that, and and I think he should, but, I think strong government support is, is only going to benefit everybody. You know, some say this is the fourth industrial revolution, you know, and I and I, you know, I'm a Manchester by.

00:27:20:16 - 00:27:46:24
Unknown
So I'm, I'm from the epicenter of the last industrial revolution. And I think every time that this happens is disruptive. But the governments that push, this kind of thing, happen in inside of the country, and support it, will see, strong benefits because much like a company, which is embracing this, will be able to be more competitive against a company that isn't.

00:27:46:24 - 00:28:04:18
Unknown
It's the same at a country level, too, and especially in, countries where the strong, like, knowledge industries like the UK, so I think I see it as a really massive opportunity. We need to reach out and grab a performance as quickly as we can. And that's, that's that's what, you know, we're trying to do.

00:28:04:20 - 00:28:34:22
Unknown
Pull. I've got one more question for you. What would be your parting words if you if a companies listen to you now, senior leaders or a CEO or an entrepreneur. What would be your parting words to encourage them to get on board? Yeah. Start now. Don't wait. Because all of this is very much layering. So, you know, you can put some effort into it and you can start to get some benefit, right now.

00:28:34:22 - 00:28:58:06
Unknown
But once you get that structure right, you'll be able to implement new tools. It's about being fit for the future. So I think the further the longer you wait, the longer procrastinate around. Well, I'm not hundred percent sure what to do. You know, the further you're going to get behind on that, because, I think if you can get the PLC right, and that process and start from that, and that can start improving your business.

00:28:58:06 - 00:29:17:01
Unknown
Well, at 30% faster than you're doing it right now, you know, think how much that layers in a year by year basis or a month by month basis. And how far behind you're going to get your competitors, if they're doing it new or not. So, I do think it's something, you know, start straight away because there's, you know, there's money at stake at the end of the day.

00:29:17:01 - 00:29:37:10
Unknown
So and, you know, this is an opportunity to improve your productivity. And you should see it as an opportunity rather than a fact, you know, and if there's an opportunity there that other people and other organizations have found great success in, then just test it in your organization. Start to build the skills. Now, because these skills are very expensive in the market.

00:29:37:10 - 00:29:59:23
Unknown
So if you've got the opportunity at the moment to to train people, to give people the opportunity to self train, and still stay, you know, roughly ahead of where the market is, whereas you wait six months to a year, you know, you're going to get to the point where as an organization, you have to start hiring these people in, and then who wants to be the person who comes in and teach an organization this stuff?

00:30:00:00 - 00:30:17:10
Unknown
If you're doing this, you want to work for an organization which teaches you. So, you know, I think there's an important gap that's going to start happening in between the organizations that are working in these kind of I first ways and the organizations that and and that gap is going to start to widen over the next 18 months.

00:30:17:12 - 00:30:45:09
Unknown
I think the two things that come to mind with that one is it's an opportunity cost. If you don't get involved, it's going to cost you business. But secondly, if you delay getting, your people trained in how to write prompts, for example, I mean, Forbes, about a month ago, Forbes, a magazine ran a, a few articles on prompt engineers and what they're charging at the moment.

00:30:45:09 - 00:31:09:23
Unknown
And if you are really good, well, you know, and these guys, I mean, these guys and gals within, you know, 2 or 3 years that probably be out of by themselves in Ireland because they're making so much money. Now, you take that forward 12 months and some companies are going to be able to get on board in 12 months time and afford, as you quite rightly say, and afford to, to get someone involved to train their people up.

00:31:09:23 - 00:31:33:13
Unknown
So you got to get involved now because otherwise it's a cost. And the same goes for individuals like think you're you skills the marketplace, you know, where do you where do you want to position your skills in the future. You don't you don't want to be the last consultancy that does Prince to. So I think, you know, you want to you want to make sure that your, skills are at the forefront of the market so that you can get the most value from them.

00:31:33:15 - 00:31:58:00
Unknown
There's a colleague, in, the communications and marketing industry here, here in Adelaide, my town. He, has for, you know, many years written, you know, small, medium sized businesses, marketing plans, and, you know, he would go research the company, get all the literature, find it, you know, read through it, come up with their marketing, voice and all this sort of stuff.

00:31:58:02 - 00:32:22:06
Unknown
Now he just goes to, perplexity or ChatGPT and says, analyze this website, which is their website. And he has some PDF documents. They've got, tell me what their tone of voice is and write a 12 month marketing plan, you know, written words and everything. And, you know, three minutes later, he's got 12 months worth of marketing, broken down by social platform.

00:32:22:22 - 00:32:42:22
Unknown
And all in the voice of that company is just brilliant. And that's essentially what automated agile is in the, in the, in the, technology process, you know, you go and you give it some information, some domain information will produce a requirements document and that won't be the requirements document use. It'll be 80% of the way there.

00:32:42:22 - 00:33:02:13
Unknown
Like a job I produced it for you. But you'll have to tweak it and get it right. But it just saves you so much time getting to that point. And this is some of the thinking that we've got around automated agile that is much easier to, have something that you can criticize it, like if you want an answer, on the internet, you could give the wrong answer and then loads of people come.

00:33:02:13 - 00:33:18:06
Unknown
Correct? Yeah. Well, it's much easier to look at requirements document, which is wrong. And then fix it than it is to create the whole thing from scratch because it's, it's just easier to think that way than it is to create from brand new. And so that's what we're trying to do with, with automated agile is, is, is give you an answer.

00:33:18:06 - 00:33:35:22
Unknown
It might not be the best answer, but at least it's summat for you to start from. And that'll save you time. Paul, thank you so much for, this, this chat today. Really appreciate it. If people in, Britain wanted to get how old if you. How did they do that? Yeah, they can contact me on my email, which would be, [email protected]