Change Management in the Age of Agents

[00:00] David: Hello, good morning, afternoon, wherever you're joining us from. Uh, it's great to be back and with new faces on the cast and one face that I actually can't see because he is just gone into dark mode. Um, it's, it's funny, I, it's always **Amber:** one [01:00] David: inside baseball, sort of. I've got all of you guys in front of me, but I've got the little strip of guests down on the, on the right edge there, and I could see the four little circles and I had three circles and a, like three faces in a silhouette. And I was like, Ooh, who's got a silhouette joining us? Uh, the faceless man known as Mr. Matt Slager, mate. **Matt:** Good. Yeah. I'm joining you all from the, uh, the coast of Adelaide. So hence the B background here and the unprofessional camera and lighting. But yeah, good to be here. **David:** You, you look great, mate. You always look great. Um, and what, what sort of a uplink are you using? Are you on starlink 4G, 5G NBN? **Matt:** Honestly, just the, uh, the, the Ben Breakfast, um, wifi here, so hopefully it's secure. I didn't do any testing. **David:** We're going out to YouTube, I don't think. Needs to be too secure for a YouTube live stream, but good. But thank, thank you for making it, Matt. So, uh, this is a live show, everyone, if Matt drops out, he tried. Um, **Matt:** yes, we **David:** appreciate that. **Matt:** Same with connection quality. [02:00] David: Um, but we've got, we've got two, two new faces on the show and you'll notice that they're not meant for a change. Um, coincidentally, nothing to do with International Women's Day that just so happened to be on the weekend. This is just because these are two friends that are excellent people, uh, and knowledgeable in this space. Uh, and so I want to bring them on to have a chat because, uh, mano and Richard couldn't join us this week. So in just, you know, top to bottom order. Joe, welcome to the show. **Jo:** Thank you so much everybody. The uh, AR operators. I'm happy to be here. **David:** You, you fall into the long time listener first time, uh, attender kind of category, right? You were often in the chat chatting with us. Now you're on the show chatting with us. **Jo:** Yes. I love, I love being in the chat with you. Um, all those questions that I throw at you, I say to me, what, or, yeah, get a really good answer. I really enjoy the show. Um, there's, yeah, lots of tips and tricks and just lots of information that actually helps me with my AI journey, so thank you. **David:** Brilliant, [03:00] David: **Matt:** Dave. Brilliant. I [03:00] David: don't know how I'm gonna feel today without looking at the chat and seeing Joe's questions there, so hopefully somebody else asks questions in her, in her place. **David:** Well, that, that, that's a challenge for the audience. Make sure you're asking us some questions and dropping there, interacting in the chat, telling us hi, where you're joining us from. We are all over the place. So I'm in Sydney. Um, Joe, where are you based? You are, **Jo:** I'm in Perth. **David:** You're in Perth. Maddie, you are down in Adelaide, I think you just said down in South, south Australia right now. Yep. Uh, and, and our other new, new new panelists tonight. Today. Uh, Amber. Amber, yes. Welcome to the show. Where you, where are you joining us from today? Amber? **Amber:** Ah, Brissy Sunshine. Yes. **David:** So I'm trying to **Amber:** see, **David:** we literally have every time zone in Australia represented today then. **Jo:** Yep. Yes we do. Yes. **David:** That's incredible. [04:00] David: Um, I, I said to the boys last week that, uh, on the thumbnail I have the East coast, the eastern St stand, daylight saving some the Sydney time, Sydney Melbourne time on the, on the thumbnail and the Perth time, because we've got man normally and I've gotta represent the West, [04:00] David: but I didn't put like the other two time zones in there. 'cause I've done other shows and podcasts before. And you'd give the thumbnail and you have like four or five time zones and it just looks ridiculous. It's like, that's crazy. Like. I don't know. How do I mean, Amber, you're up and Brizi, how do you feel about being excluded from your time slot being on the thumbnail of a show? Right, **Amber:** that's fine. I've put it all through my socials so everybody knew. 11 o'clock Brisbane time, so, **David:** yeah. And, and, and Joe, you're, you're in wa The wa Hass always got a chip on its shoulder about the, uh, the East coast ride. **Jo:** We have a bit, yes. Yes. We do feel a bit like our own dominion here. **David:** And I, I, I, I feel I can still say that even though I haven't lived there for over 20 years, because I was born there and lived there for over 20 years. So I, I understand. **Jo:** Oh, I was actually born at Eastern States and I've been here for 30 years. I'm not sure if I'm still, uh, I've gained my residency status yet. It takes a while. **David:** Yeah. I think, I think I cross over this year that I've lived more of my life on the East coast than the West coast. [05:00] Jo: Yeah. **David:** So yeah, it's, it's good. We don't, we don't have the same mobility in Australia that you see in other places like America. Right? They, they, they do have a lot more interstate mobility. Do you guys notice that? Or is it just me? **Jo:** I have noticed that. Yeah. I totally agree. Like, I just find people pick up and move when I'm listening to different people. They're like, they've moved states across states, across states, and they just, it's just normal. It's like, sort of takes a lot for an Aussie to move from one state to another. **David:** Yeah. Do you think, do you think, uh, AI's gonna change that with the job mobility, remoteness bandwidth? Or is it just, oh, that's Aussie culture and that's the way it's gonna be? If we invest in the cities, people sort of may gravitate to one place and they kind of generally stay there. **Amber:** I, I, I think the younger generation, like the, the new, you know, the younger generation that have really, I guess, uh, very f familiar and it's normal to just be working from home or, um, be on socials and be influencers and things like that. [06:00] Jo: I just think. It'll be, we'll see more people [06:00] Jo: just, you know, moving state to state will be very common, **Matt:** especially like you said, Dave. 'cause you can do everything remotely. **Amber:** Mm-hmm. **Matt:** And, you know, be able to spin up everything just as long as you've got a good internet connection. Yeah. Um, you basically can live anywhere. **Amber:** Mm-hmm. **David:** Yeah. I, I, I'd, I'd like to see more states in Australia. I think, I think we've got, we don't have enough. I think they're too big. Um, they, they're, I mean, like, as if, as if think like as if from like a New South Wales point of view, how are the guys sitting in the middle of the Sydney CBD making good decisions for Dubbo and Wagga and those, those country towns over in wa it's so big. **Matt:** Mm-hmm. [07:00] Jo: **David:** Um, how are they, how is Perth making good decisions for Kalgoorlie? That's also good for Perth? Like, and so then you think about business opportunities and what we are trying to do with, with starting businesses and businesses, investing in ai. I think one of the things in the place like the US is they have different tax laws and they have different business policies, so they try to attract talent from other states based on [07:00] Jo: their business laws and things. Here we're all very homogenous. It's all kind of the same. So it's like, oh, okay, I just like the weather better there. All the rules, all the laws, all the things from running a business are pretty much the same. Maybe Victoria's a bit harder given, given the, their state government and, and the legislation down there, but yeah, it's all kind of same. Same. **Jo:** Yeah. I think we should, um, ask for Mano to be, um, mayor because he would, um, mano for mayor in Cal, he'll be bringing up the wa get things sorted. **Matt:** When, when can he start that journey? **Jo:** Oh, well he's already in the throes of it there, so we should do, you know, put him forward **Matt:** mano for mayor. **Jo:** Yeah, I think you're right though, especially like just the distance and the spaciousness here. [08:00] Jo: I mean, I'm, one of my passions is to really get. Help towards the, for the remote regional communities, specifically focusing on WA at the moment, because that's my sort of heartland. But, um, yeah, I think we just need to, we, I'd like to see a bit more of us being [08:00] Jo: able to distribute our information and our support **Amber:** Yeah. **Jo:** Out to the, uh, regional areas. Um, I know there's a lot of good, uh, small businesses that have all started up. There's a lot of, um, ladies that are on farms that are running farms and also starting side businesses. Well, and thankful to starlink and being able to use AI so that they can have things running in the background while they're managing farm children, you know, the general day to day. So, mm-hmm. Yeah, it's really, it is really important to get to regional areas. **Amber:** Mm-hmm. Yeah, I agree with that. I share that passion. It's something that Joe and I talk about a lot. It's just how we can offer our knowledge and our, I guess, um, if there's tools that we can create and build that would be very helpful for. That would be, I guess, um, easy to use, easy to implement. **Matt:** Mm-hmm. **Amber:** Just to make everyone's, you know, life easier. Um, [09:00] Jo: **Matt:** what sort of areas, Amber, like, you know, I know you've done stuff with [09:00] Jo: NDIS in the past, like is that still like your primary focus? **Amber:** Mm, a hundred percent. I mean, that's basically where majority of where I spend my time is, like with NDS providers and implementing, uh, automation into the systems. Uh, I've got a bit of a journey of what that looks like just by tapping into all the different types of AI tools that are available and that have been, I guess, uh, advancing. But, um, yeah, ND Ices is still definitely, I guess, my bread and butter. I can't get away from it, even if I try. I'm very passionate about it. Uh, and yeah, it's, **David:** but anything that, that I, AI and tools can do for providers to help streamline costs is gonna be huge because we, we know like that's as much as this is, and I've declared me time, this is not a political podcast. We're not here to pick sides or talk policy, but, you know, we we're running businesses and building businesses and, and government policy does affect businesses. [10:00] Jo: And, [10:00] Jo: and so things like the NDIS we know is growing unsustainably. **Amber:** Mm. **David:** Um, and so we need to be looking at how people like UAM can be investing in that ecosystem with modernization and, and AI tools, et cetera, to try and streamline and cut costs so that way the people that really need the help will be able to get the help. **Amber:** It's really exciting. I was actually speaking to a, um, director yesterday about this just, and Joe, you mentioned about education, having, you know, having a conversation. 'cause it does take a time to bring awareness of what's possible and. I think that the more we talk about what's possible, then people can see solutions and I guess we can close that gap of, uh, I don't know what to do. [11:00] Jo: Feeling quite overwhelmed having all this information on socials where it's just nearly seems, you know, wild of what they're actually being [11:00] Jo: exposed to. So I think with the NDIS space, because everyone is so busy in the day to day, they don't have, they don't have a second to even really, um, to do, they don't even have the capacity to do their own job. You know, then everyone's working essentially like three different roles. So I'm really excited to, to help more NDIS providers to automate their systems. Uh, and I think what's been very helpful is knowing the operations of, uh. Of a sector, you know, of what's involved to run an NDI service provider and go in and, uh, I guess integrate the automations. Like we're just gonna see ease come into different providers, uh, well different businesses, right? **Matt:** Yeah, [12:00] Jo: **Jo:** yeah, yeah. I, oh, sorry. I totally agree. And I think that, um, speaking to the regional areas like the NDIS, uh, businesses that are in [12:00] Jo: regional are really actually needing a lot of support at the moment because sometimes if they've been locally based, they have a certain. Size, you know, population that they are actually have customers within. Mm. They actually need to be able to use, get a toolkit together, which, um, I call it for me on in our business, it's the mud map because it's never one straight road of it in incorporating one tool here, one tool there. Because at some point you may need to upgrade to another tool here and another tool there. So we call it the mud map. Um, and it just helps people get actual guidance. It's okay to go on the dirt track, it's okay to go, you know, left at the stump. It's okay to come to a decision point. There's a tree down, okay, what do you need to do? What's the issue, what tool do we need? You know, things like that. [13:00] Jo: But with Amber's, um, having such experience, extensive experience within that NDIS space, uh, I think it's just kinda invaluable. And yes, Amber and I do have conversations in [13:00] Jo: regards to the regional areas and what it is. Um, that we can actually put out there, um, and go and speak and go and catch up and have like, I think I called them the, she shed, she shed talks, you know. **David:** Nice. **Matt:** That's cool. That, um, that reminds me a lot of a, a concept that I've been trying to use for myself recently. You talk about like when the tree's down and you know, which path do you take next? Doesn't matter so much that there's no road there. It's that you are learning how to navigate. Mm. It's the same with all of these tools and stuff. You know, every single new thing that comes out doesn't really matter if you have to pivot to something new, as long as you know how to navigate in general, like what the base principles are, the fundamentals, you know, the building blocks of each thing. **Amber:** Mm-hmm. **Matt:** Um, yeah, **Jo:** it's **Matt:** very true **Jo:** and exactly, it's like what, what size tool do you actually need to move that tree? [14:00] Jo: Do you actually need like something high powered? Do you need a, a chainsaw or do you just need a little pruner that's just, or do you just need to push it out of the way? And it's okay. It was an obstacle that appeared to be an obstacle, but you can still keep going. With the same toolkit. Um, and I [14:00] Jo: think that's where, um, our businesses in the AI space come into it, where we really are doing a lot more education and a lot more guidance, um, and understanding of what is actually possible, which is what you mentioned as well, Matt and um, Amber. Yeah. **Matt:** That principled approach is really, really important. Genuinely because like, this stuff's gonna keep changing. **Amber:** Yeah. **Matt:** But yeah, if you help people with that level of technical literacy and how to navigate, they'll be able to spot the opportunities on their own. **Amber:** I think having a really strong foundation where you've got things ready to plug in, you've got, I guess systems a, a, a system that can talk to each other. I think for that's the starting point. And then as things change, 'cause they will change, you can easily just pivot. Yeah. I, **David:** sorry, Emma. [15:00] Jo: **Amber:** No, I was just saying I think that's the beauty of like when you engage with an AI developer or you know, when you start [15:00] Jo: thinking about how can I implement ai? I think the conversation is what foundations do I need to have? That's the key. **David:** Yeah. A AI is very much the marketing term and I think we, we lead first with the, we are in ai, we run AI businesses, we help businesses adopt ai, but we all know that often it's actually workflows and automation. It's getting their systems to talk to each other. It's getting rid of the swivel chair mechanics of switching system system and copy and paste stuff and automating some of those small processes, which doesn't necessarily need any AI at all. It's just workflows and wiring and APIs mm-hmm. Will have a big boost to begin with to do some of that. Decision making and routing, you might have an AI layer in there that helps to do that. And that's more accessible than ever before now. **Amber:** Mm **Jo:** mm **David:** But we market with ai often the solution does not need a lot of ai. [16:00] Jo: **Jo:** Mm. Yeah, that's right. And just trying to get people from paper. In the smaller businesses that I've spoken with, there's still all of the invoice papering, you know, it, that needs to be collected, it needs to be entered, and the [16:00] Jo: data needs to be actually organized and cleaned up before they start considering, you know, we can do an audit of their business. But it doesn't, it doesn't start there. They really need to start, like Anna said, at the actual building a foundation and getting that data into the system so you can actually, um, integrate something that's gonna be supportive. **David:** Yeah, it's cri critical to understand not only the customer journey, like what happens when a customer reaches out and interacts with your business and what are all the things touchpoint your business has with that customer to deliver your product or service. But behind that, what is the, the data flow map, what do you do with the customer information? At which points are you generating new information for the customer? How does that connect system to system? Because if you don't have visibility of those things, then you've just got this mess of oh, that thing's over there and that thing's over there. You've got this fragmentation, which really is the opposite of efficiency and being able to sort of streamline. **Jo:** Yeah. [17:00] Jo: **Amber:** Sorry, I'm just, I'm loving this conversation 'cause this is exactly, I guess what we talk nearly on a daily basis as far as like [17:00] Jo: efficiency, smart systems integrate, you know, a smart operating system that supports change, uh, as you said, Dave, of like having, or, and Joe not having sos and having not having everything where it can be. Found online that can an agent or something like that can review it and push out, you know, and change your system as things change. Like that's what I'm really excited about and kind of where I've been nearly every single day for the last few weeks. But it's yeah, exciting **David:** and, and, and it needs to actually deliver value. Sometimes these efficiencies don't drive a lot of value. I heard someone talking about, um, people get really efficient at using inefficient systems. Mm-hmm. So people, and you think about people in typing pools and you, those sort of keyboard operator people that are using terminal software that was built from the eighties. [18:00] Jo: They know every single keyboard shortcut. They can type 120 words per minute. They can navigate around those systems. I [18:00] Jo: mean, Matt loves his terminal interfaces. He's all over that sort of stuff. Um, people can get really efficient at using, arguably inefficient, I'm not saying terminals or inefficient Matt, but they can like, get, get a really efficient at using inefficient and badly laid out software because they just lure and the tricks and the traps and they know how to go through it. I remember talking to a customer many years ago in a previous life, and the, uh, I was in a presales, technical presales capacity, and the account exec took me out there and the account executive had given me this full brief, this is their problem. They've got all this issue with their, with their customer service process and this and that and the other. And I was like, okay, cool. We're gonna go in there to pitch a customer service solution. And then when I spoke to the customer directly, like, yeah, I mean, that's true. There's, there's things there, but we're actually pretty good at managing all of that. But over in the sales section, that's where we are dropping things and missing that and we don't have this. [19:00] Jo: And I was like, okay, cool. There's one thing to diagnose inefficiencies and say, yeah, we could deliver a solution, but the uplift might not be very big if there's no real breakage there, versus another area where you can really diagnose improvements and, [19:00] Jo: and generate a return on the investment. **Amber:** Yeah. Yeah. **Matt:** It's pretty similar to the, the question we've just gotten from, um, from Spector who says, you know, do you see AI being useful in the farming industry? So it's a good example of, you know, where is the value? And I feel like Joe is probably well suited to, to answer that kind of question. **Jo:** Uh, passion regional farmers. I just think that at the moment they, there's a lot of farming actually that are quite ahead. Um, and they have started integrating AI into the actual bigger operations like ai, um, tools and systems, but they're also using it for robotics. That's a huge testing area, not just in Australia, but uh, worldwide. Um, and that's for planting seeds. It's for, they've actually got cameras that assess the plant or the weed. **David:** Mm-hmm. [20:00] Jo: **Jo:** And then they know exactly what they're doing, what they need to hit it with, what are they targeting with. So, and it's also those sort of AI robotic like. Driven [20:00] Jo: efficiencies, it feeds back to another system in the farm. Uh, they can look through it. It's also gonna reduce the amount of chemical exposure. Um, and they've also, I I was actually watching a documentary in regards to what's going on in Australia with the robotics and one of the machines that they've created. They're fantastic. They take the machines, the robots out, they put them on a farm for four hours, do what the farmer wants, take it back. And they've been going backwards and forwards and developing these. Um, and they, some of them are actually removing the weeds. They've recognized the weed between the plants and they take them out, therefore reduces, takes out, takes out the chemical process of having to spray, which, you know, that's just been a, a huge part of farming. Um, it's a concern for to consumers and to farmers because the way you know that, that can have some effect if that's not properly processed. [21:00] Jo: **David:** Drones with lasers. That's what we need. Drones that fly along, they see the weeds. Pew, pew. [21:00] Jo: But, but I, no, I mean, I joke, but they are doing that kind of thing. And whether they're using la like you said, they, they actually are using a lot of intelligent drone systems. They're using artificial, uh, like sort of, um, machine learning patterns to understand which plants are which, uh, vision recognition and they're doing all kinds of stuff. Internet of things, which was kind of a big thing, pre cursing ai. But now the two together is crazy. That's where we get terminators in the future. But, uh, that's huge on farms, right? Like for water management, chemical management, animal management, gates, fences, everything. Like, the more they connect, especially Australian farms are so huge, some of 'em, like the big, um, massive stations. Yes. And being able to have this remote visibility of things, control things on and off. It's pretty big, [22:00] Jo: **Jo:** and especially with water. So if they've actually got the water levels being measured and things like that, that can be automated, you can put that to a system that I know that they do have those sorts of things that are easy. But just interestingly speaking to what you were saying about the flamethrower and the, they actually do have a, uh, they're [22:00] Jo: experimenting on one farm with a robot that's actually is flaming the weed. Nice. So, uh, they've, they've gotta get the flame length, height and intensity. Right. They had a few incidences, but that is actually part of it as well. So **David:** I remember seeing a while back, they had these robot dogs and they were equipping them with flame throwers. I dunno whether, whether, whether that was for a military purpose or not, but they these like little robot dogs that could shoot massive long flames. I was like, that's that's an interesting use case. That's, that's scary. Um, but, uh, **Matt:** jokes aside though, Joe, what, what sort of value prospects do you see for farms at like different tier levels? You know, so if you are just a, a farmer with like 10 hectares worth of dairy cows or something, you know, is it worth looking at or do you have to be one of these like, monumental grain farms that has like kilometers worth of stuff? [23:00] Jo: **Jo:** No, [23:00] Jo: I don't think so. And I think that's where it comes into the aspect of the tools also. Um, like what, what exactly is being introduced there? Which brands, which type of, um, actual robotic equipment or ai, you know, generated support. But that comes down to as well, like legislation. So our governments have really gotta, um, be supportive of putting these processes in for all variety of size farms. Um. Yeah, it's definitely, it's very exciting for people because where they're feeling that they haven't, um, that there's things that are deficit that they can't do because of our monitoring. Because the monitoring of, I remember working on a farm in Mount Gambia and just to go out and check and find all the sheep on the back of a four wheeler, but take us over an hour and a half just to find them. [24:00] Jo: Then you've gotta turn around and bring them back then. So there's just a lot of monitoring systems that can be put in place. Um, they've actually shown a robot that goes into [24:00] Jo: a, a paddock and has got the farmer's voice cloned and it calls them and they, it took about three tries. And now it can go out and just call the, call the cows. **David:** Wow. Ag Ag Tech's getting crazy. Yeah. I mean, it's **Jo:** amazing. **David:** J Jeremy Clarkson's not, not exactly known for being a high tech wizard, but even he was using sort of the virtual gating stuff with the, the goats on his farm, right? Where they puts these little collars on the goats, uh, and sort of sets up a virtual geofence. Uh, and then he can move the goats around and they learn where the fence is based on getting sort of buzzed by these collars. And so that way you don't need as much infrastructure investment in fencing, and you can deploy livestock in places with virtual fences that are all controlled remotely and alarms. It's, it's pretty wild. And given the stress on farmers and how much, how hard it is to be a farmer in any country at the moment, the more tech and things they can apply to, to simplify their job and, and reduce costs and stress and burden, **Amber:** yeah. The [25:00] David: better. **Amber:** Yeah, it's, it's pretty exciting. As far as the. Things that we can automate with machinery. Like my girlfriend who has 32 acres, she's just recently got a machine that does all the lawns and she just watches it on her phone and it, it crops it all. So, you know, that's just, I mean, even that alone where she's not out there and it will take her nearly two days, **David:** you **Amber:** know, to, to do the, it's just amazing what is coming out and what's being Yeah, **Matt:** that's, that's actually a really, really good point. [26:00] David: 'cause like my question before, like what, what is the value prospect, you know, that time saving where they're doing stuff that they shouldn't be doing. 'cause they can find a way to get it done in some really clever way. Like that, you know, the idea of, um, just doing the lawns, you know, for a, for a farmer like you would expect to be, be looking at ways to make their accounting better, you know, tracking their, their livestock like you said before, but. Just relieving them of their lawn pressure is, you know, something that's huge **Amber:** that people don't realize. I mean, I'm not a farmer, but I have friends that live on a little land. But even that, her time of not being on a chopper now and she can actually spend time with her kids, which is what she's doing in replacement of being on the chopper. Like, it's just huge. Such a time saver. **David:** Yeah. When, when I first bought a robo rock or vacuum cleaner for the house, I called it the, the a time machine. 'cause the idea was it was just buying time back rather than spending an hour vacuuming the house. You run this little thing and it runs around and sweeps and cleans and you've saved time and, and this is that on a bigger scale, these sort of auto mowing machines and things go out there and there's domestic versions, but the industrial versions are crazy. What's what's coming? And I mean, if we can have self-driving cars, why can't you have a self-driving mower? Obviously safety issues for the blade and stuff and object detection. But that can do all that kind of stuff now. Right. [27:00] David: **Amber:** Yeah, it [27:00] David: has the same, uh, I guess mechanisms as far as, you know, those little sensors. Mm-hmm. So technology is kind of the same. **Jo:** Yeah, I'm absolutely, um, I'm loving our little robot, our vacuum robot. That saves me. And I didn't actually even realize, I thought, how much time do I actually spend vacuuming and we're just pop on in the morning being renamed by the kids numerous times. Sometimes it stink and sometimes it's Harry ands, the Dore Dolores. They go on the and change it. And then I've also been at home by myself and I'm doing the washing and all of a sudden the vacuum cleaner starts. 'cause one of the kids is in a car three streets away and they thought they'd just turn it on and get it started. I'm like, it really is time saving. I think that's what, um, we all, as business owners in this space are trying. [28:00] David: To achieve for any, any one of our clients. Um, and obviously there's varying degrees of technicality. Um, you know, like we [28:00] David: all, uh, four of us are doing, um, quite different areas of work and it's just, it's all relevant. It's there's time consuming, uh, tasks that, like you're saying, the value mat is every single person has their own value mark on it. So, you know, things that are really important for me to ti save time and get time back is totally different to anyone of, um, anybody else or any other business owner. **Matt:** A hundred percent. Like where is everyone individually? Where can we find their, you know, home robot, you know, what in their business, where can we deploy something that gives them equivalent time back that they just weren't aware that they were wasting on other stuff, like they've said at the beginning, like inefficient systems that they just persisting with and getting really good at dealing with. You know, that kind of thing is where some of the biggest unlocks are, and it, it's not always immediately obvious [29:00] David: **David:** everything from answering phones. So the AI receptionist concept is, [29:00] David: it can be a huge one for a small business owner and missing calls and missing jobs and, and revenue opportunities because you couldn't answer the call or, or whatnot. So that's sort of that end through to the automated paper processing and, and digitization of the processes and things through to then actual service delivery, uh, AI agents for enhancing work productivity, for, for research and content generation and other bits and pieces. There's, there's no one size fits all. Uh, and that's one of the things we like to talk about on the show. Like, who knew we were gonna talk about ag tech and farm robots on, on this give of week. But these are all areas where these things have parallels and can pull out and you can adapt ideas from that industry, industry into another industry. **Amber:** Mm-hmm. [30:00] David: **Jo:** And just to clarify, I am not a farmer. I'm from a farming background. Mm-hmm. I, I lived on a farm with my grandparents for a period of time growing up, and, um, all of my uncles were farming. So that's where my farming passion, like passion for the farmers come in because they, they work so it, it is just, they work really, [30:00] David: really hard. They deserve a little bit of time off. And that takes education because we have this, uh, socialization. I see that with my, my parents, that there's, there's things that in them that says if they're not working really hard, they haven't had a productive day. So I've been shaping and re, you know, teaching them about what you know. Go and do your sewing, mom, go and do your gardening. Dad, you, you know, that it doesn't have to be a big, so I think that's also even for, um, my age group, uh, we are, we have been really, um, trained to be like, if you're not working really, really hard and achieving all these things at once, when you can actually get your AI to just get onto it and, you know, have your agent help you out with your mail and, uh, your business calls or, or whatever. And like you said, the reception or there's a whole variety of things there. [31:00] David: **David:** Yeah. So, so long, as long as you're actually using that time to not just do more work all the time. I mean, [31:00] David: we've discussed this on the show previously where, and there there's pros and cons to that. Sometimes it is good to free up that time doing sort of busy work, so then you can do more productive business growth work. Sometimes you should be, like you said, using that time to step back and just do something else that's not just work, like spending time with the kids or, or investing in hobbies and other creative pursuits because, and, and Richard mentioned last week. Human creativity is gonna be critical to actually really leveraging the most out of these AI tools. **Jo:** Mm. And I think Amber and I have had numerous conversations, um, over the last six months about looking at, um, personal growth and what, what you're doing in, when you've got these things are sorted out in the business, you actually can actually take that time out to go for that walk to allow that creativity to come up to actually then, you know, purse ideas that are your ideas that you want to then get AI get your agent to help you out with. Yeah, [32:00] David: **Amber:** we combining the two, you know, I think like [32:00] David: when you understand the capabilities of AI or if you've got AI to assist you, you can essentially still be productive and go for your walks. And AI will essentially continue the task or increase your productivity for the day. So like, for example. You know, I've got an agent that would send me updates that's happening either in the, in my business or in the AI space. And, you know, it can link up to now a voice ai or even note lm lm, you know? Mm-hmm. Where it's got the audio. It's like just having that automated and going for your walk and you're listening to a podcast and instead of you coming back to your computer and doing your research, you're gonna essentially go out and **Jo:** mm-hmm. **Amber:** You know, so integrating the two, this is coming from someone that is a little bit obsessed with ai, you know, so, [33:00] David: **David:** I'm, I, I, I wanna jump in into what you're doing with some of that stuff in just a moment, Amber, because I think it's really interesting and I want to pivot into [33:00] David: that, to, to, to, for the back half of the chat. Um. Yeah. And I had a segue that I was gonna say, and now it's completely lost my head around what you were saying before we jumped into the other thing. What **Amber:** capacity? Identity. **David:** Oh, the, the, the trap. The trap of, oh, before you go for your walk, just do this thing. Um, have any of you fallen into the just one more prompt trap where you are like, I'm gonna go for my walk and I'm just, I just get the AI doing this thing. Oh, oh, finish too quickly. I need to come back. It's just one and then like 15 minutes later you're still sitting here going, I, I need but one more prompt. **Amber:** Yeah. **Matt:** I'm terrible. Like, if I'm by myself. Yeah. If I have nothing else planned. So like, this, this trip has been, um, I don't know, crazy detox in that sense. You know, I feel on, on one side of the coin, I feel like I'm falling really far behind fit. Then on the other side, Dave, it's like all of that, just one more prompt thing is, um, I don't know. I've gotten like a little bit of a, a release from it. I. [34:00] David: **Jo:** It's so important to take that [34:00] David: break. I think that one of the things I came from, um, my previous chapter work-wise was extremely active. Very active. Um, and so I had 29 years of active bus, uh, working in an active environment. And I've come to being on the computer all the time and it's sedentary. And I actually have my kids saying, no, no, we are going to the beach. You're coming. No, no, you have to get off. You've got to cut. We are going for a dog walk. We got like, so I actually have to incorporate and put reminders in my AI system to say, righto, time out, time out. Come on. You know, so that, yeah, it's, um, it's a challenge because yeah, you get excited about things and you wanna keep going. **Amber:** That is it, that's the, the trap is, it's exciting. You see things. [35:00] David: Come to life, you see things working and you're just like, that is, you know? Yeah. And so I think that that dopamine, you know that here you're just like, okay, one more prompt. I'm gonna automate that now I'm gonna automate that. Now, experiment with this. But [35:00] David: it really is a matter of committing to that, you know, whether it's going for a walk or going outside, like, I have time on my calendar that it's family time. So committing to that, because yeah, you can easily be stuck behind your computer, but that's not life, that's not living. You still gotta have that balance. I, **David:** I have my 11-year-old daughter saying, dad, can you cook dinner now? I'm like, yeah, okay. Just, just, just, just a minute. And like three, five minutes later, I'm still in here. And she walks back and I'm like, okay, okay. And just like, just walk away from the keyboard. It doesn't matter. The agents will be there in, in an hour or so. When you come back, you can leave it for an hour. It's fine, Dave, but **Jo:** I could just see your laptop with its arms out, hugging around your neck while you're trying to cook over the stove, calling **David:** you **Jo:** back. Um, [36:00] David: **David:** but I mean, I, I, I took advantage of the weather this morning. I went out and, and gave my lawns a quick mow the old fashioned way today, uh, I don't have rubber mowers, but uh, it only takes me out for me 'cause I'm on regular lock. Uh, all right, we we're gonna try something new here. I'm gonna do something that we haven't done [36:00] David: before. Not that pop **Amber:** quiz. No, I'm kidding. **David:** No, no. I, I, we, I'm, I wanna start putting more segments in the show and sort of like really delineating different conversation topics. Uh, and so I didn't quite get the bumper fully ready in time to try and integrate cleanly, but we're gonna do this. That was, that was worth the suspense. Right? **Amber:** I love it. Love it. **David:** So, uh, I wanna, I wanna deep dive on a topic a little bit and get a bit into a, a deeper tech chat, uh, around AI operating systems and, uh, that sort of. Overall prevailing concept that sort of going around, you're seeing a lot of talk about it on different YouTube channels, on LinkedIn, every man and their dog, since OpenCL has kind of evolved, uh, and people have sort of seen what they can do with sort of the 24 7 personal agent. [37:00] David: Now people are [37:00] David: trying to evolve that into a broader thing. I'd like to say I was there first. I've been talking about this for months with what I've been doing in anti-gravity, uh, and my work called, uh, my anti claw system, which didn't, which, uh, yeah, I, I've been on this path for a long time and evolving it myself, but it's good to see everyone else catching up. I mean, good ideas. No one has a monopoly on, on, on good ideas, I suppose. Um, but, uh, panel, what are you guys doing as far as AI operating systems? What are you seeing? What are you hearing? Uh, Amber, you've got like 35 agents working for you or something. Is that what you would call your AI operating system? **Amber:** It's up to 40. **David:** 40, **Amber:** yeah. Yeah. **David:** That's wild. **Amber:** They, and look, I've only done that many because. To keep them under control. And so they operate appropriately. They don't go rogue. Um, but **David:** like they're really compartmentalized and very specialist. [38:00] David: **Amber:** Yeah, basically I set up the whole agency, like an automated agency, like I would as I would help a [38:00] David: business startup, so seven different departments and then under each department I would have, you know, different skills. And so the reason why I guess this agency or this, this army, AI army is growing is purely because of the fact that when I have a new project, I'm like, okay, I need a copywriter that just does for YouTube short and long, you know? And so that would be automated to, um, to different departments to go through to production. Uh, but yeah, it's fun. This is where it's like, get off your computer. I still go for my walk every morning, you know, because it is really cool to see things. Happening. And, yeah. **David:** So I mean, this, this sounds similar to what I was doing with my different superhero names and, and I had different people for different types of job roles. You've just continued that path for like 40 odd people. Do they have personal names? Are they just like defined by their skill name? **Amber:** Yeah. [39:00] David: **David:** Do they have managers? Uh, and what system, like what are you building this in? Like if [39:00] David: someone goes, I, I want an army of 40 people working for me, how do I be like Amber? **Amber:** Yeah, they don't have names. 'cause I thought, I get my kids' names confused. Like, if I was starting to name all these different agents, I'm like, who is Harry and who is Bob? So I'm like, no, no, no, let's just keep it simple that that's the copywriter guy, or that's the YouTuber or that's the, you know, whatever it is, the sales first and all the customer service. Um **David:** mm-hmm. **Amber:** So I just kept the, the names very simple as far as the skill, the what they're responsible for. An agent. Mm-hmm. Um, they all do have a manager, so nothing can actually go through to production or be, um, I guess actioned without the manager okaying it. The manager would occasionally, most likely, depending on the complexity of the task, would ask for my approval and I would say yay or nay. And then the manager would take it, you know, take the information and take it back to the, the people. [40:00] David: **Matt:** How do you, how are [40:00] David: you handling your, your jobs and task list there? Like how do you manage state and, you know, everything like that at this stage. **Amber:** Yeah. And this is why I was looking to zen flow. 'cause I'm like, okay, I need to, um, look into a way where I can keep things under wrap. Because at the moment things are coming to me through, like I said, the manager. Mm-hmm. Um. It depends on the project. I would still say like give the manual task unless it's already automated. Like for example, the research agent that happens every morning or if a client was to book into my calendar, the agent would do a deep research on the person who's booking in like their do web scrape. Um, so I guess as far as how I manage it, it depends on the, the project that's at hand or what's active at the moment, where other things, what have you built it in is **David:** just called code agents. Is this like skills and [41:00] David: **Amber:** well code? Yeah, [41:00] David: everything's in called code. So that's mm-hmm. **David:** And yeah, and you just talk to 'em because, so they're their sort of skills and you've just got skill definition files for like 40 different skills and they talk in that manner. **Amber:** Exactly. Yeah. **Jo:** And Amber, a question. So for just for example, for a small business owner, uh, maybe they've got this, four of them in total. Uh, you looking at doing that sort of number of build for a small business, like what sort of, because they're obviously using all of the departments that you've created, **Amber:** um, **Jo:** but if you are actually building this out for small business, uh, for people that are in the business, one owner and three, you know, employees currently maybe service based business, what number would you think that you might need for the number of agents to actually be managing? **Amber:** Yeah, this is what I say. **Jo:** If they were to go, like, you know, to go the full Monte with it. [42:00] David: **Amber:** Yeah. [42:00] David: I have this conversation nearly daily as far as like what kind of business would be suitable for this, how can it be implemented, how can it be managed? Uh, and something that I've kind of experimenting with at the moment. Which I'll share with you in a second. Um, but to answer your question, Joe, it could be suitable for a startup. You know, if someone doesn't even have a team that just essentially wants to have a business, has their sales taken care of, have their marketing taken care of, obviously there's parts that you still need to do, like it still needs, well, to a degree. [43:00] David: Like for example, you still need to be the face of the camera or you can get an avatar, but that still needs to happen. And still, once it goes to the, uh, file, then the AI agent would take that information, you know, and do what it needs to do. It's the same thing like at the moment, um, with my notetaker. So otta, notetakers, every single time I record in that, that automatically goes [43:00] David: to my, uh, one of my copywriter agents. It does, it actually goes to the research agent, finds out what's suitable and what's not. Then it categorizes into the. The pillars, the social media pillars, and then another agent would turn that, um, my voice notes, my, you know, into a short or long term, uh, form type a script. Um, so yeah, it'd be suitable for a startup who doesn't have any team. I always say that if you do have a team and you want to implement some of these, uh, agents, which are essentially automations, like let's just be real. **David:** Mm-hmm. **Amber:** Agents are just like another way of just controlling automation. Right. If you are going to implement these tools into a larger organization, there has to be a level of change management because **Jo:** Yeah. [44:00] David: **Amber:** And I, I really go emphasize on that because, you know, you've got a system that already works, you've gotta allow people to compliment the system that's coming in. [44:00] David: So, **Jo:** yes. Yeah, **Matt:** absolutely. When you say change management, like what? What does that even mean? **Amber:** What does that Oh, okay. **Jo:** Mm. Right now, **Amber:** you know, because I forget. I forget that like, devs don't understand, like it's just not common, right? Unless you've seen it and lived and breathed like a system change in an organization, holy duly like that. That is a project and a half. You have to essentially take everybody along the journey with you. There has to be a level of communication that everybody is aware of what's going on, because you can plug in an automation, but a, you wanna make sure that people are aware that this is how we are doing these things now. [45:00] David: You know? So this is why communication, that awareness, uh, it has to be well mapped out of how change looks when you are implementing ai. Purely because of the fact that, [45:00] David: uh, to keep your customers happy. So to keep your, for developers, to keep the, the client's, clients ha customers happy, there needs to be communication to them that essentially, if we're talking about voice ai, for example, I would actually put out some communication out to your client base and say like, we are **Jo:** making changes. **Amber:** We've got a new a, a voice AI employee. You know, Amy, like, let's introduce, make her feel welcome, blah, blah, blah. She's here to support you, but don't worry, humans are still involved. You know, so it's a, it's reducing that sense of, oh my gosh, like changes happening. I don't like this. 'cause nobody likes change. It's, it makes people feel uncomfortable. So change management map is essentially managing the expectations and the emotions that go with when you are implementing AI or any kind of change in that, that example. [46:00] David: **Jo:** Yeah. So the humans and then the tech. So you really gotta try and. Get them in on board and actually, so they can [46:00] David: see and they're aware that it's an actual complimentary system to their role rather than, you know, I mean obviously there are some positions that organizations will be, um, withdrawing from, uh, the human resources department. But um, yeah, the change management is massive. For myself and Amber, we've both been through big changes in, um, large organization and I think probably both of you gentlemen have as well, because, but it's usually the humans. That's disrupting the integration. Uh, it's not usually the developers that come in, they know exactly what they need to do, how they're gonna do it, and the, the hours of training for all of us humans that are working. [47:00] David: Like I think what, um, what you were saying, Dave, is that, you know, the, um, they've been doing it the hard way, the old way, the long way, the mm-hmm. And they've got really good at troubleshooting around all these issues that it's just like, it, it's sort of ingrained and that you [47:00] David: have to sort of take that layer at a, a layer, at a time is it's okay that we actually don't need to do it. It's gonna be like this. And I think that there's a very fine balance between the developers coming in the, the pre-conversation with executives and, you know, management. And then actually that change management process needs to be in place so the developers can, number one, get on with their job. So they can actually do the integration, um, and provide that nice and cleanly through the organization, but also the humans that are there. All of us, like anyone that's been an employee that's had a new system, it it is, you're nervous. It just makes people nervous and scared that they don't know how to use it, that they're gonna take practice and that sort of thing. **David:** It's translating the ROI from the organizational level to the personal level. [48:00] David: I think if I was trying to like summarize the value or the importance of change management, so at a company level, yeah, we're gonna put this in and we're gonna save the bottom line. We're gonna do this, we're gonna grow things. But for a personal and individual, what is the value of this change [48:00] David: to me? How does this help me in my job? How is this this gonna benefit me? And, and having run sort of a teams, I, I had this in not even with ai, I, I implemented a thing called quiet time once. So my team of devs was constantly getting interrupted by people from the sales team. They just come up 24 7 just, and just interrupt them in the middle of flow. And if they, this is. Then they, they weren't vibe coding. They were code coding. And so they, they were really getting into the flow. They were, they were there, they were focused, and then they'd get random salespeople just coming up and interrupting them and breaking them out to talk about stuff. And so I actually had to institute quiet time for my team, which sounds ridiculous for adults, that I had to put in a quiet time. [49:00] David: But I made, made a little sign and I put a little sign up and I corded our, off our, our area in the open floor office, um, with these signs. And so when the sales staff came up, they saw that and knew, oh, I don't just walk up to their desks. Um, they had to prearrange meetings. They had to do things because it wouldn't just interrupt one [49:00] David: person. The whole, the whole pod would get interrupted by these questions. And my team didn't love it at first. They pushed back a little bit going this ridiculous day, blah. But they actually, but, but, but the change management was, was highlighting the benefits of it for not just them for, but for everyone. And to the point where it became normal and it was okay, and the sales team would walk up and go, ah, they'd walk away. Or if one of my teams saw them want to talk to them, they would get up and they would leave and they'd go over to the kitchen area and they'd have a chat over there. They wouldn't chat at their desk. **Jo:** Mm. **David:** Um, and it was all around translating the benefit of this change of behavior, this change of, of workflow, why is this a good thing? Mm-hmm. And why does this benefit you? **Jo:** Yeah. That pro, um, productivity, ROI, so you actually see things being completed quicker, more efficiently with less mistakes. If people, the deep work **David:** and our teams that hate time logging and they, they hate logging where their time goes and where their hours go. And that's fine if the pro, if it's only being used for their manager to micromanage them. [50:00] David: But if that's being used, [50:00] David: because you, the manager needs to push back on another team and say, we're wasting time on this kind of work and we'll look at how much is being spent here when we shouldn't be doing that. We should be doing this. Mm, different conversation so that again, it's the change managers, like the ROI of doing the time logging the ROI of doing this, the ROI, of adopting the AI tool. What does this mean for me? It doesn't put my job at risk. It actually gives me more time. Makes me more effective. Lets me do other things. **Jo:** Yeah. Yeah. And operational vision. Like if a lot of people that work in for big organizations, if they actually understand why this is being implemented and what the road like the mud map looks like, they're usually a lot more, um, interested in being supportive in that and not objective, uh, and not, um, not objecting against, you know, any of the new introduction for, um, automations and things like that. [51:00] David: **Amber:** It's also like when I talk, think about automation and how it relates to business growth, like the [51:00] David: biggest, the, well, I think the greatest thing because this is something that I was trying to achieve without automate, without ai. Is essentially maximizing and, um, training up the team to know how to deliver quality care, quality customer service. So keeping your customers, your existing customers happy, but also then going out and, and building new relationships. But the thing is, is what happens is we get so bogged down in the day-to-day, the administration, the report writing, the shift notes, the logging times, or whatever it is, right? **Jo:** Mm-hmm. **Amber:** And so by having automation **Jo:** mm-hmm. [52:00] David: **Amber:** Put into place that takes out all those manual tasks that then allows humans, employees to spend more time on the quality, you know, whether it's relationships, whether it's the ser, uh, whatever it is, then business will [52:00] David: grow because you are, you created capacity for those people to then go out and, and, you know, do that bit of a. Um, to generate and, and have energy, um, with bring more leads in, essentially. **David:** So on, on your system, Amber, have you got any kind of visualization, any kind of dashboard, any kind of interface or just you're in Claude code on the terminal and you directly mentioned different agents. Uh, have you, like how are you automating them? Have they, are they tied in with Claude Cowork? Like what does it look like from a, a user? If someone wanted to say, I want that. **Amber:** Mm, **David:** is it just they're in the developer interface or is there something you've built on top to help you manage and see the 40 agents and who's working and who's on what and scheduling which jobs where? [53:00] David: **Amber:** Yeah, and this is what I'm having fun with at the moment and building a dashboard. So it's a bit more user friendly. 'cause at the moment I'm living in Claude code and, uh, in the terminal and, you know, wherever it is. Um, and I need something that will be, I guess, more visible [53:00] David: from a visual, visually will be, I guess, helpful for directors or even for me to talk to. Other developers to see what's happening, to see the activity. Um, so a dashboard, but the other thing that I'm wanting to explore is the thing that you suggested Dave. **David:** Mm. **Amber:** Was that little pixel movement I had **David:** a little pixel agents. **Amber:** Yeah. Yeah. So that way, you know, as the director, you know, you say to or yourself or the director, actual human says to the CFO or, uh, yeah. CFO of the agent. 'cause that's, there is a manager that talks to all of them. You give that order to say, Hey, this is the new project we are wanting to do. We're doing a campaign, you know, we, uh, we had this idea, blah, blah, blah. Can you send it through to your AI agency to see what task can be picked up? [54:00] David: And I would like to have, um, an. You know, something where you can see the [54:00] David: little agents moving around going into the Yeah, thank you. That's exactly right. Going into the marketing department, talking to all the agents. Um, I thought even having something where as a, another agent comes to life of like welcoming them in. Like, Hey, we've got like a new, um, **Jo:** got a new agent, **Amber:** new member. Having the agents having their own birthday have a bit of a celebration, like this is the day that that agent was born. Like keeping it quite human and, and realistic and so we can build a relationship even with our AI whole, it's like a whole agency, isn't it? **David:** Yeah. **Amber:** I'm trying to think of a better terminology of how I describe this. Like, **David:** like, like this **Amber:** is exactly right because Oh, that's **Jo:** fantastic, [55:00] David: **Amber:** even for myself. So I spent a lot of time with my quality team, AI quality team, just because it's. Getting everything working [55:00] David: right, um, and giving them real projects and real tasks. So like it is getting quite overwhelming for, I think even for them and their robots. So something else that I'm doing at the moment, and please, I would love your input with this, especially if this is something that could be, uh, duplicated, which it can be duplicated obviously, and, and spread into other people's businesses, is the quality team is now reviewing sos. [56:00] David: So with the tasks or the projects that I am giving them, I would create a SOP as well that would go to the quality team. And I have now an agent that just reviews those, those SOPs, the processes to make sure that they are aligning if I do need to make changes. In the agency, for example, there's been another employee, so that per, you know, that agent is doing something else, [56:00] David: all I would need to do is update the so. Mm-hmm. And it gets actually automatically pushed through. Um, I thought that would be an easier way to, uh, and again, I'm very early stages. I've been experimenting with this this week, but having something that would push out, um, and the reason why I was also, I came to that conclusion to do it that way, rather than me going in and saying to like, 'cause I'm updating the prompts, right? I'm updating the instructions with every single time I update the agency. I thought if I just updated something that I can just see on a document of a new process that's come into place, then that means that when this, when other business owners wanna use something like this. All they need to do through that auditing process that I would have with them, I would say, what does your process looks like? [57:00] David: And so not only do I have a full duplicated agency that's ready to plug in, again, I would plug in bit by bit because of the change management component, but [57:00] David: that SOP document, which can be copied, would just be modified to, um, their business. And that way it would just push out all the changes. So I'm yet to see how that all works because it's complex. It's insanely complex that we're talking about the, **David:** especially across that many agents. Yeah, **Amber:** I was gonna say the technical, like the guardrails, the things that they do, and I'm just like, hold on a second. How did that all happen? Like I'm spending a lot of time on ensuring that the quality, um. Uh, is there, the guardrails are there that they're not going rogue, is how I call it. Um, so [58:00] David: **Jo:** this product that, um, you would, you have here is really, uh, for enterprise level, the, you're trying to get it to a point where it's, um, going to be, um, happy with the complexities of SOS changing and updating, whether it's from legislation or whether it's from in-house, [58:00] David: um, you know, within the organization. **Amber:** That's right. Yeah. If I can get everything smoothly running and operating as it should, um, then it could, I would still always start from a smaller scale and start with the sales and marketing. 'cause I've spent a lot of time in that department, um, with obviously the quality, but. You know, early days, Joe, really early days and still experimenting. Um, I've only experimented with a handful of SOPs. **Jo:** I can say that I've spoken to Amber, uh, 10 days ago, and between in 10 days, this woman rocks. Uh, amazing. **David:** Um, SSOP SOPs for those wandering at home, uh, that is standard operating procedures. Just, yep. Just clarifying, just making sure I understand. Make sure we make sure all our, our viewers that, that, that may not be, uh, with, with such uh, terms now, I think what you're doing is amazing. [59:00] David: Amber, obviously I'm a big fan and try and, and this sort of code based IDE based [59:00] David: system is tricky for people to get their heads around and adopt, I think, a little bit. Mm-hmm. And so it's great to hear you're trying to build a dashboard and making that more interactive. I can't believe it was three weeks ago now. Apparently I published out, um, a gi repo with a basic version of what I've been building with the, uh, the anti claw system. Um. For people, which, you know, for people that are wondering, what is Amber talking about? The idea is, and correct me if I'm wrong, Amber, you've got these skills, now I've given mine fancy names. Um, but essentially like if I go and look at Lois, this is a definition for what my research specialist does and their philosophy and their directives and what they can do. And you, you're building out sort of these job descriptions for all your different specialists, right? **Amber:** Yeah, exactly. **David:** Um, and, and you are looking at how do you do things So in mind, I've got an onboarding one where you first set it up and it goes through and talks about what is your business entity, what do you do? [01:00:00] David: And you are looking at how do you do that for all your standard operating procedures, so that way it can ingest them, update the agents with them, [01:00:00] David: and then continue to be self-learning and self-healing as your business evolves. Right. If I understand **Amber:** exactly, **Matt:** you just said self-learning and self-heal. Yeah, I, and then Amber, you said exactly. I, um, I'm dubious that you actually are implementing any sort of self-learning or self-healing. Do you have any like recursive agent pedagogy or um, validation loops, any sort of self-evaluation like in that system at all? **Amber:** Is that to me, **Matt:** yes. **Amber:** Okay. I use, I'm using a third party, like a non-biased agent. So as I, as I test everything, I have a non-biased agent that's essentially cha not challenging, but, um, giving me, uh, an alternative way of doing things. That's to prevent, as we know, they like to be confidently wrong. Oh no, what is it confidently right when they're wrong. **Jo:** Mm-hmm. [01:01:00] David: **Amber:** Um, so [01:01:00] David: that's how I'm essentially managing that. As far as the exactly comment, that's how you've just displayed. Um, David, like that's exactly how I've mapped out and how I've built this. Um, and funny enough, Joe, you gave me the inspiration of how to start was essentially where I started was building an org chart. **Jo:** Yeah. **Amber:** A very detailed one. **David:** Yeah. **Jo:** Yeah, **David:** yeah. I mean, this is the challenge and I think Amber, you said you are, you are hitting up against guardrails and I think that's sort of like Matt Matt's point is these things become more complex. Where are the checks and balances? How do they learn? Where is their memory system happening? Now that open claw is a month old, everyone's going, oh, my open claw forgets things and it sucks. Now what happened to my amazing open claw? And its because it wasn't built for running over long timeframes and having, uh, sort of a longer recursive memory. It was really good in that Oh wow. In the first like five days, that's done all this amazing stuff, but it wasn't necessarily designed and built with, you know, working for, yeah. [01:02:00] David: Days, weeks and months. [01:02:00] David: Um, and so, and only any of these generative systems, how are these agents not becoming stale? How are they not suffering from entropy and just recursive, like, like going down the drain of the sort of, uh, circling themselves. So I think, you know, your quality team having a, that third party agents that are inputting in. **Matt:** Mm-hmm. **David:** But then how do they, are they rewriting their own skill files? Are they updating memory logs or decision logs? So in mind, uh, there was a whole memory folder where I actually have a [[state_of_play]] and a [[decisions_log]]. And one of the key things they do is they go through is update that sort of shared memory area to try and evolve the whole thing forward and just keep everything moving together. But **Amber:** that's exactly, that's what I've automated recently is getting that information. So getting it reviewed, and this is why I do spend a lot of time in the quality section. Um. Yeah. But even, um, outside of that, David, you made a really good point just before, is having it not only reviewed, but automatedly to push out. [01:03:00] David: I think, [01:03:00] David: uh, the reason why I guess it's getting quite complex now is I've purposely isolated skills. **David:** Mm-hmm. **Amber:** So I can, it's easier to kind of troubleshoot because when they were doing a few different things at the start when they'll, you know, had a bigger, larger responsibility, it was quite difficult. So I'm trying to like isolate, I guess, tasks and skills so I can update them. Also, the quality agent can kind of review it and make sure it's doing it properly. **David:** Do you have a HR skill that helps you hire skills and manage skills? **Amber:** I do have a hr, um, yeah, a department, if that's what you Yeah. I haven't spent much time there purely because the fact that it's not probably the very first thing that I would roll out. Um. **David:** That's interesting. 'cause for me, one of the first things I want is I want the agents to be building agents. I don't wanna be manually writing prompts. And so having a, a skill that can create skills and architect them well and get the task division and go, well this one's doing too much. This is complex. [01:04:00] David: We need to break this skill up and, and break into two [01:04:00] David: different roles. I don't wanna be doing that by hand. **Amber:** No, that's actually, sorry, I thought you meant hr. As far as HR for organization. **David:** I, I'm, I'm talking about, you know, calling the agents humans. Um, so a ar **Amber:** so my very, very first agent I built, well, there was three, and I know I've shared this with you guys before, but it was essentially, um, it was an agent that could build agents, but before that, it was an agent that was to test, uh, vulnerabilities, really understand around, um, guardrails and things to watch out for. Then I built the, the agent that could build agents. **Jo:** Mm-hmm. **Amber:** Um, and I, I don't know, by, I don't know if that order helped because there was no, I guess. Defense with the first agent being built because there was no agents essentially being built, you know, as far as what things should I put in there to protect our agents. [01:05:00] David: Um, and then there was one, um, that would review to make sure that everything, all the tech side is good. [01:05:00] David: And then, sorry, there was a fourth one that would actually push out the changes, um, push out the changes. If the agent, the review agent would say, technically this is not right, or these are, these are urgent, um, vulnerabilities that you need to action. And then I would have my third party non-bias agent that would review it. Again, very complex, a lot of time and a lot of like cognitive, you know. **David:** Yeah. You got inception going on there. That's, it's crazy. It's complex. Um, just, just, just to bring it back to one last, uh, question from the crowd. I know we're over time. We try to keep this about an hour. Thank you. Thank you ladies for joining us for this, uh, this episode and Matt on holidays. Um, but one good question around change management stuff. While, while I have, uh, you guys here for developers, would them getting involved in discussions with frontline staff around change management be worth looking at? [01:06:00] David: Or should it be something that's handled by the client's management team? So who owns that change management and do you want the, [01:06:00] David: the devs involved **Amber:** for developers? Wh when, whether the developers should be involved in that? Change management? Yeah. **David:** Yeah. Get, get, getting them talking to their actual end customers Or do you hide them behind the, the back down in the dungeon and just, uh. Ask messages to them. **Amber:** Yeah. So when it comes to change, it depends on also on the organization. Well, actually it doesn't because of this reason. When it comes to change management, you want to assign one person to be like a change leader. So just say whether that's your HR person, maybe you've got a relationship manager in the organization. [01:07:00] David: Maybe if it's a smaller organization, that's kind why I made the comment at the start. Maybe it's you, maybe it's a CFO. So it really comes down to, um, the, someone in the company that can be there to full time, to be approachable, to support, uh, support the change and support the team members as far as [01:07:00] David: what the developer's involvement is. The developer's involvement is really educating that change manager that's assigned in the organization to make sure they have all the information. To then, uh, roll out the changes because I guess developers aren't experienced in what, like in change management? **David:** So, so I suppose the question is, is should they be talking to their stakeholders? Should they hear directly from the stakeholders or should they just be getting Well, **Jo:** um, I'll just say these, it really depends on if you are the developer and you are the head of your, of your organ. If you're, say, a solo and you don't have another non dev person. So this is why I think the encouragement is with the partnerships is you developer, non-developer, because often the non-development person like myself or Amber or other, um, you know, ladies and men, whatever in, when joining together, I'm much more of an open spokesperson. [01:08:00] David: I [01:08:00] David: have an understanding of what my developer, we work closely together. So when, when they, I do the interviews and I get the information and then they do the interview if they want to. Or there's discussion of what is going to be coming in for the change management at the very beginning of this, um, quite often the developers are so busy and so under, uh, the pump, getting everything set up, organized, looking at what the infrastructure that they've gotta work with, you know, and maybe having to put in new infrastructure, like what are they actually working with in the, for, to get this development in the organization. [01:09:00] David: I think that it's you, it is often like for, for me, in our business, it's my role. So it's my role to have those conversations and be that contact point for the organization. So speaking to that, Sean, it just depends on, I guess, in what your business setup is. Like. If you are purely development, um, then, and you've got someone that's non development based, [01:09:00] David: that is your, um, shop front sort of person, um, then that's, I think that that's satisfactory that that person is doing that communication and leaving you outta the fray. Obviously you're gonna have internal meetings, but I just, what I'm finding in business is that the developer has enough on their plate and it's almost like what Dave's saying earlier. Put the shh, sign up, you know, put the quiet time, sign up for them so they can, whether they're, um, code, no code, like whatever, whichever part of the, you know, uh, project they're doing, they still need to be able to do that. Really good deep. Work time. Um, so I don't know if that answers your question, Sean. **David:** I think, I think it sounds, what you're saying is it's horses for courses, you know, go, go back on the ag tech thing. Uh, and there's no one size fits all. But if you've got an effective conduit that can communicate the needs and the value between the two parties, then the developer doesn't necessarily need to be frontline. [01:10:00] David: But if the developer's not getting an understanding of [01:10:00] David: what the customer and needs actually are and why it matters, and because, you know, we have technical people that go, but this needs to be like this for this particular reason. Sometimes not understanding the end customer benefit or the experience the customer needs. And so you need that translation. And sometimes putting the developer in the room will make that worse. Sometimes it'll make it better. That's right. Depends on the developer's. Depends on the front end person. **Jo:** Yeah. And that's what Amber was saying about having just one person within the organization that the actual automation team. Can speak to so that it doesn't get watered down with Chinese whispers and all that sort of thing. That they can actually have that one change person that can be the leader for the management of that process. **Amber:** It can get messy otherwise. **Jo:** Mm-hmm. **Amber:** And then nothing gets, nothing gets implemented. **Jo:** Mm-hmm. **Amber:** Um, I guess one last thing on all of this on change management, Dave, 'cause I know we've gotta go, is expectation and communication. [01:11:00] David: Like, I can't stress enough how important communication is in the role of a change manager and making [01:11:00] David: sure that everybody has, has been appropriately communicated to so they have an expectation of what their role is, what's happening. It's really taking them through that journey as well. **Jo:** Absolutely. **David:** Brilliant. Well, thank you ladies. It's been a great show. Um, Matt as well. Uh, great to have you from, uh, sunny South Australia. There. Those clouds look, delight **Jo:** it. Just look at the water, the color **David:** of the water **Jo:** there. I know. Oh my goodness. It's **David:** full of sharks though, so don't go on it. **Jo:** Oh yeah. Amazing. **David:** Uh, brilliant. Well thank you everyone. Thank you all. Thank you to everyone who joined us live. Thank you for watching this on the replay. We appreciate if you wanna learn more about us and our businesses and what we do, all the links. I'll add Amber and Joe into the links after the show so you can find more about them and what they're doing. [01:12:00] David: We will be back next week. If you liked Amber and Joe, drop it in the comments. Let us know if you wanna see 'em back, if you wanna see a bit more, uh, shuffling [01:12:00] David: around of the deck chairs. Not on the Titanic. No, no. This is a, a successful ship that we're sailing forward. So, uh, if you, if you'd like, uh, that let us know, drop a comment. Comments are great for the algorithm to let 'em know that we are here. And so, aside from sharing and telling your friends around this brilliant show, the AI operators, make sure to drop comments so strangers can learn about us too. And we'll see you all next week. Thanks everyone. **Amber:** Thank you. Thanks everyone. Thanks Dave. See you.