Speaker 1: 0:08

All right, we're back with that Change Show. I'm your host, Jason Little. This is what we used to call a weekly show, but now we kind of do it whenever something interesting comes along. And something interesting came along. I met Tapio Kamalinen. Hopefully I pronounced it right Kamalainen.

Speaker 2: 0:23

Yes, Kumalainen.

Speaker 1: 0:25

Kumalainen yes, that's right. You'd think for the number of times I've been to Finland, I would know how to do this. But we actually met at the Change Leadership in Toronto at the end of May. So I was doing a talk on AI and I remember somebody asked a question about change management platforms that use AI. What would you recommend? And my skepticism meter immediately went off, because I know how easy it is to train and bias it towards your tools and your ideas. And you can tell your AI bot don't ever mention this method and it won't when people ask it questions. So I was kind of skeptical about it. And then we met in the hallway and you showed me HowSpace. So do you want to give the audience a quick introduction to you and HowSpace?

Speaker 2: 1:16

Yes, thanks, david. So Tapio Kumailaainen and I lead the North American operation for Halfbase and relatively fresh Canadian. So we're headquartered in Finland and I moved from Helsinki four years ago, so now here in Toronto, canada, but I've been in the company for 15 years now. So we started as a small boutique change and training, very human, centcentric consulting company that had a software side already back in 99. And the first working what's now Halfbase we did for a larger change case in 2010. And that turned out to be that good. But then we had a couple of other cases for that turned out to be that good started offering the tool for other consulting and training companies in other countries. That worked out so well that 2014 we decided that we spin off the consulting company on its own and the software side became what's now half based, so focused on this, only on this tool we sold or or a sunset that all the other tools we created. That point started focusing only on this. That's where it's been going from.

Speaker 1: 2:31

Okay, cool. So were you already starting with AI implementations in the platform, you know, back in 2014, 2015? Or was that kind of more?

Speaker 2: 2:40

I think 16. I think 16 or 15. We had a university network that did good research on AI, so we had AI tools good AI tools already before the generative AI came around. But of course I will say jump up that we did not invent generative AI, unfortunately, but implemented it now fairly quickly after it came out, so it is an improvement to what we had Right.

Speaker 1: 3:10

Yeah, because it seems like the last year and a bit it's really exploded, but a lot of people don't realize it's been around for a long time and a lot of companies have been playing with it for quite a while, but now the window of opportunity has opened up and lots of folks are starting to put AI features into their software. I know we've got a bunch in our lean change tools to do summaries and little coaching widgets and bots and stuff that will help people with change experiments and stuff. But what I liked about when we were chatting at the change leadership was the focus on sensemaking, because I think that is one of the biggest. I think that's one of the most useful things that AI can help with. Change and transformation is taking massive amounts of data and help us make sense of it so we can make better decisions sooner.

Speaker 2: 4:01

Yes, I think the and to some extent maybe the well, the biggest part is the whole kind of collect, send, respond and. But biggest, yes, I think, is the sense making out of that. That wasn't possible necessarily in the same way before. So collecting lots of people's responses I think, jason maybe you spoke about it as well or before like very interesting to see what happens to surveys. That might be that they become irrelevant. That's really interesting to see.

Speaker 2: 4:31

Many of the world change readiness service I think you've spoken about it before here or something that. That's one that I see that I'm really keen to see if we can replace it with two or three good open questions and you get much more quantitative data on top of the qualitative one. So that, but also not only thinking it so much as a survey, but also because I think for people to change what they do often requires that they talk about it or make sense of it with the other people around them. So making that happen. So survey is just like collecting data, but one way of kind of facilitating conversations, like we've done in the workshops for as long as change has been happening, I think so. So doing that online to that sense and making sense of that big online conversation yeah, that was a big part of a lot of side conversations I had there.

Speaker 1: 5:21

Yeah, as well is we? We chatted about um, um. The funny thing with change readiness surveys is I always ask people so what happens if it comes back and people aren't ready? What do you do? They're like, oh well, we do it anyway because we've already got a schedule and a budget. I'm like then, why are you doing it? And most of the answers are uh, because it looks nice in charts. You know, if you're asking, do you agree, disagree, strongly agree, whatever, you can put that stuff in a chart and you can cherry pick the quotes that you want.

Speaker 1: 5:48

And I think one of the things I've talked about with AI quite often is we're kind of getting into the ethical parts about AI. But AI won't suppress things unless you tell it to. But if you're collecting insights from people about a strategy or a vision or something about a change, um, that's like it's in there and it's now transparent. So you as the change agent, you can't control. I want to only show things that are super positive. I mean, you could put an instruction set in your AI to be able to do that, but I think that's really one of the benefits is number one. You can anonymize the data pretty easily and exactly that. You don't have to dumb down your you know, maybe you hit a detour on the way to work and the first thing that pops up is a change readiness survey and you're in a bad mood. Yes, you know like, and with rating scales and things like that, you're not really getting actionable stuff.

Speaker 2: 6:45

So that's the the interesting part for me, anyway yeah, and I think that it enables the kind of a more like the design thinking process, a little bit like opening, opening up and narrowing down that there is. The open questions are opening up. Hey, what's the readiness at this point? And then you get some of the points from there with the ai that these are the. These are the main things, but often I think the result is that that it's not always necessarily enough to kind of understand or make actionable points following that. But then the next response okay, hey, these are the things I heard from you.

Speaker 2: 7:20

Based on this, I want to hear more about this and this and this and those are then leading into actionable points following that. But same can do it in a scale. And a big part of that, I think, is, when needed, especially when it's more like a people focused change, that this part is maybe underappreciated, that having this small interaction back and forth gives the people the feeling that we're actually part of this change. So, funnily enough, ai kind of enables us to make in a scale happening that people feel that, okay, I'm actually working on this change together and it's less feeling like it's kind of forced upon me, if wanted.

Speaker 1: 8:00

Right, so are you seeing folks that are using the platform? Is it a mix of asynchronous and synchronous? Do you find that you have any stories from customers where they're just projecting it up on the wall and they're in the room with a team and they're using it together, or is it a lot of it is just asynchronous, where people are putting data in and then there's a change coalition or team or whatever that's taking it and analyzing it and then there's a change coalition or team or whatever that's taking it and analyzing it.

Speaker 2: 8:25

Yes, both COVID did a funny curve on that. Before COVID it was mixed a lot, it was asynchronous and synchronous. Then COVID threw everything synchronous for a while and now it's back into kind of the pre-COVID synchronous, asynchronous. And I enjoy that a lot in the sense that they uh the feeling of how much you get done when people are the same space but in kind of support of or partially, instead of talking to each other, taking like five minutes and putting their thoughts together and using ai then to make sense out of that. That kind of jumps you ahead in a conversation a lot. You get a kind of much higher starting point, much faster than you would do just without using any of these kind of tools.

Speaker 1: 9:10

Yeah, I like the point that you mentioned around. We're making sense of where we are now and we're looking at how ready we are now based on the data that we have. And I think it's reasonably well documented that traditional change says you do all that up front and then you execute and blah, blah, blah, but it's almost like little micro iterations. You know we're making sense of where we're at now. We're like okay, we have three options we could possibly do. Let's pick one, let's go ahead and now let's dump feedback in. So you're kind of getting into an iterative approach to change and transformation Way of doing. Yeah, yeah.

Speaker 2: 9:47

I see that the AI can lots of the change management activities. Makes it iterativeness easier in those, makes it easier to do them this kind of the collect, send, respond process faster in the sense that you don't have to run the whole cycle over weeks or months or smaller parts of time.

Speaker 1: 10:03

Yeah, yeah, cycle over weeks or months or on smaller parts of time? Yeah, yeah, so are you seeing companies that are, that are using the platform? Um, how has it shifted? Maybe how they would have done change documentation in the past or change communications, which you know we document everything, we blast out a bunch of communications. Is it shifting things more towards sense making in dialogue or are they still kind of taking their comms plan and all that stuff and dumping it into the platform to try to I think very good point I think it's mostly I see how it works is that the first time they run, even the platform changes.

Speaker 2: 10:38

It's comfortable if you change the platform. They don't don't necessarily change the way you do things. The first time you run you do it in the same way you've done with your existing tools, just in a new, new environment. And I see the kind of the point of innovation happens usually after you've done that and then you see how the experience is different and then that gives the room to think that, oh, now I could actually do it differently.

Speaker 2: 10:59

Uh, some of these parts and yes, that that's something that I see happening differently is that doing more of those kind of a how would I say, like a test runs or pilots or smaller groups at heart first and seeing the response from those, or kind of making sense out of what happens, those before expanding it further.

Speaker 2: 11:22

Some cases and I'm happy to see that when it makes sense some cases I see a more focus into kind of the, like I said, people-centric changes that are larger involvement of people, because in the past it would have necessarily necessarily meant pulling them into alcohol meetings, the way they used to do it, or or email communication one way, all these kind kinds of things. And now I see cases that it makes sense, that they see that it's relevant for the change to involve them more in the kind of two-way communication. Well, I maybe see the biggest thing, but I also see that that doesn't always make sense. When it's maybe a more a IT or process-focused change, then maybe the end user opinion is not always that relevant at that stage of the change. Then it makes sense to be more one-way communication, right In a sense.

Speaker 1: 12:09

Yeah, okay, okay. So one of the things that I know I've talked to a lot of people about with AI is privacy, and it usually means two different things. Like, the one part of the privacy is if we're using a hosted solution, how do we make sure our IP doesn't go into the ether of AI? I think that's less of a concern with platform stuff because it's obviously it's a closed ecosystem. You're not feeding it into Gemini or Copilot. But the other one is making it safe safe, I guess, for people to put feedback in. So you know, anonymizing my feedback about a particular change where I might not feel I could put in what I really feel about it because everybody knows who I am. Do you have clients that kind of go down that road of how can we make it private and make sure that it's safe for people to put stuff in, or is it more transparent and it's wide open?

Speaker 2: 13:13

Yes, both In the sense. In our system it's easy to change. It's a couple of clicks and you can make it, and it's permanently and you can change it by question by question. So we often encourage that, depending what you're asking, you know your own cultural organization, be the judge that, hey, this might need something, that we get better responses if it's anonymous. But we do encourage also in the questions where it makes sense, of course, we do encourage that people respond with their own name and their own thoughts and these kind of things. So it varies with the situation. If it's asking for ideas, in some cultures it makes sense that they're anonymized so that the hierarchy doesn't play a role in what idea is supported or encouraged or not necessarily taken forward.

Speaker 2: 13:54

So it might work better just as a anonymous ideas or, of course, the feedback how people are. If it's open question how people are feeling, depending on the safety and trust in the organization, sometimes it works better that it's open. Open question how people are feeling, depending on the safety and trust in the organization. Sometimes it works better that it's open. Sometimes it works better that it's anonymous.

Speaker 1: 14:10

Okay, okay, yeah, I think the good thing with that is obviously number one having the choice for which ones that you want to do, plus the ability to analyze sentiment over time. Yeah, the ability to to analyze sentiment over time, yeah, so you know, obviously early on in change there's a lot of uncertainty might be pretty chaotic.

Speaker 1: 14:30

People aren't exactly sure what this is or how it affects them, and then you can sort of see over time and sentiment and response of the data. Um, for me, something like that kind of gets into and we already touched on it um, always making sense of our current reality so we can decide to do, decide what to do next, so we're not so much focused on like just specifically an roi or a set of measurements that tells us if the change was successful. You can look at trends and energy and sentiment and find patterns and stuff a lot easier. Yeah.

Speaker 2: 15:06

Yeah, and those are very interesting because it might be something else happening in the organization that changes the sentiment for these changes as well.

Speaker 2: 15:12

It can be another change program or something else or something in the market that changes the sentiment for this change along the way. So I very much pray to kind of be understanding and keeping track of that along the way. And that's some of my favorite prompts is asking for the sentiment from the responses and and thoughts, and that's then also that's also, of course, interesting that the sentiment is very different depending what you ask for that you can impact your sentiment a lot by what, what kind of things you ask. And another side of that that is different from, different from our platform and in many other ways than surveys, because in survey the idea is to collect the data but you don't make the thoughts visible for the others and also you can always choose. So, depending on what you ask, you also create kind of a certain kind of energy or feeling for the people about that change in that sense. So in our kind of environment you're familiar with the appreciative inquiry. Yes, yes, I thought so.

Speaker 2: 16:09

So that's often commonly used by people who use ALF as a part of that, because it also builds that kind of energy around what is happening.

Speaker 1: 16:20

Do you ever find that there's maybe some leaders in the organization that might be a little afraid of that for people to put in the raw?

Speaker 2: 16:30

information in there.

Speaker 1: 16:31

And I've kind of been in the view that I think leaders would rather have the bad news sooner. I think what they don't like is when they're surprised. So when we're suppressing stuff in reports and just giving the good, positive feedback and what people like and our adoption is going up and stuff. I find they would much rather know the reality of what's going on.

Speaker 2: 17:13

But do you get cases that they're upset about and so they're concerned especially for that? That when they kind of ask for things, that this is the response, what responses kind of responses they get because that's their experience from their past and but this is very different in the sense that they because people respond very different in a different context. So if the context is this is what this is about and this is the question and this is what we're talking about, and especially if it's with your name, then you don't get those kind of responsive lash out thoughts. But that's my understanding and what I when speaking to people kind of worse there, where the experience of this kind of facilitating conversations are, and it's very different in that sense.

Speaker 1: 17:55

Okay.

Speaker 2: 17:55

But so I think the context is super, super important in that, in what kind of responses to get. But yes, in the beginning I find most of the people are cautious of that because of the past experience where people then are upset about the parking lots or the wrong size bio-carbage bags or what not. That bothers them on the given day.

Speaker 1: 18:16

Okay, does that extend to different companies in different countries? Because I know you guys are global, so you're probably working with multinationals that are in multiple countries and stuff and depending on which country like I, I know a lot. Uh, in the nordics there's a very low power distance index. Yeah, so feedback, raw feedback, is welcomed and the conversations are actually encouraged, and then in other countries it's respect the chain of command and stuff like that. So do you run into any kind of interesting dynamics of a particular culture that doesn't want that transparency Because it might make somebody with a higher status look bad for any reason and get any of these kind of weird cultural dynamics going on?

Speaker 2: 19:03

I think that the kind of the generalization I fully signed up of the lower hierarchy in the Nordic countries and higher almost anywhere else, in that that's what I come across A Rarely do I see that as a concern, because then usually the people who are responding are also in part of that culture in there, so they kind of they self-adjust into that to match in that sense. So I see the same hesitation everywhere. Okay, same level. I haven't seen that other places would be more hesitant than the others.

Speaker 1: 19:43

Okay, Is there any example or something that you wanted to show to give the readers a little view into how your platform does what it does?

Speaker 2: 19:56

Okay, I have one very short one. I can show my screen because this is I like, because it's I like, because it's our internal data, so I can show this is our own internal house space and this is from April 2020, this conversation, so pandemic had just started. We all just started to work from home and we had our own kind of asynchronous town hall uh, monthly, uh, monthly town hall and we had just this kind of we had consistently this kind of a a survey that, hey, how's your overall feeling as a part of the part of the tool?

Speaker 2: 20:37

so so, then, I've changed these responses since, I think, but it's kind of just responding hey, how's your work flowing, do you remember taking enough breaks and how do you manage your workload? And then when I respond I can see the kind of the averages of that situation. It's set up now so I can see the averages and if, as an admin, I can see the different answering rounds, that who's thinking what's on. But then the other one we had a fairly open question, that kind of challenges and ideas in our current situation for everybody. This kind of what kind of challenges do you experience, how do you tackle them? New ways of work which you could share and you can see.

Speaker 2: 21:11

Four years ago then my colleagues had a quite a lot of conversation about it how, how are, how are things? And this is what I'm I am excited about these days in these kind of conversations, these kind of open questions, is for them to be able to prompt and ask that what's coming out of the thought, so making sense of the conversation without having to read all of those so common. I think I'd ask in this kind of, when it's this kind of, it's a question that would have been very tricky in the past and the former versions of AI because it's asking for two things at the same time it's asking for challenges and the ideas, and this changed with the generative AI, that it can understand the context of both of those, so it can separate out of this. So I can ask that, hey, list me three main challenges and ideas, the answers and ideas, the answers. And then I'll just ask you that it'll come through with those and list those, but maybe and in this I noticed that I'd actually puts them in the line so I could maybe ask as bullet points make it a little bit easier read for myself.

Speaker 2: 22:31

So now it gives the main challenges and ideas. So actually now then I ask the three main challenges, then I test as ideas. So I need to do one more. Now I saw three main challenges and three main ideas as a pull-up. So this is the iterative way of prompting. So now I get what I wanted the main challenges and main ideas in here and now then, in the sense, often this would be the way of kind of hey, these are the main things I heard from you and I'll have a different ways of kind of automated using is that I could, for example, easily create a poll based on this that could reach back to people that, hey, these are the main things I heard in terms of challenges Choose one challenge and one idea out of these. So to get the kind of selection, choosing, choosing that what's the most important priority, that would be the way forward in this.

Speaker 1: 23:29

Oh, very cool. So each team might have their own space. So this is the house-based town hall. So is that analogous, to say, like a Slack channel? That is this particular topic and then you may have multiple ones across the organization?

Speaker 2: 23:44

Yes, yes, in the sense or in, I think, a change context. It would be, for example, this this space could be a, could be two different levels or two different of. This could be for the change manager and the broader change team and maybe sponsors and these. This could be the space where the change journey happens. So so project team alignment could happen here. So communicate If the project team is 30, 40, 50 people, look at the whole change management activities and those things, and automation of the AI with those. That would be the one level. Or then this could be another space added to that for the training program, or another for a communication of the changes or several changes in that for the wider group of stakeholders. So those are the most common kind of spaces built on this.

Speaker 1: 24:30

Okay and then say, as a stakeholder, could I aggregate stuff across multiple spaces for a big picture, if you want to call it that, or is it just based on how you set up your different spaces?

Speaker 2: 24:44

You would not aggregate from the several different spaces Inside the space, yes, but not from the defensive. They're designed to be separate units from the kind of safety security one. Yes, okay, yeah, cool.

Speaker 1: 24:55

Oh, that's very cool. I like the way that you can ask it. Simple questions. You've got a feedback button on the bottom, so did you like or not like this response? Um, I'm assuming you've. You've probably got humans that that, uh, much like google gemini does, it tells you hey, humans might review your prompts every once in a while, just for training. So I would assume you could do something similar so people in the organization could see what questions people are asking and and stuff to retrain their modeler to understand that.

Speaker 2: 25:27

Yes, what other people ask, yes, the other part that's earlier that I'm excited is that it's also you can generate things with the uh, with the ai. So either, in this kind of scale, is that if I go and go and edit the page, if I add something, I can generate anything with the with the ai. So I could ask it to a create a pulse for a, a james influence with scale of one, three, and generate. That's not a perfect, but that's now a on-the-fly type of prompt. So then it's a influence and a kind of impact. You could duplicate this and change the impact and have different stakeholder groups there as an option and these kind of things. Okay, so the generation with the AI is something that's kind of very new and very exciting.

Speaker 1: 26:31

Oh, very cool, and I see a couple of presentations on there too. Does that mean that that data is fed into the AI as well? So you could say, you know, ask a question about the change and it would pull data from documents or the yet.

Speaker 2: 26:48

Uh, at this point only in a search mode that you can. If you have the documents, you can search and the ai will come through those, the pds and these kind of things, but not into those. Not intended. That would, I suppose, a we've seen that the other tools are maybe more used for that. That haven't seen the use case yet for kind of a gradient data and the PDFs in the same as the comments. That's focused on those social aspects, right right Okay.

Speaker 1: 27:16

All right, yeah, and the reason why I was asking that is I would imagine you get some customers that are doing creative stuff that you didn't anticipate. They're probably using the tool in a way that you're like, hey, that's not a bad idea, maybe we should explore those features. You don't have to give away the store, but if there's any interesting things that's not behind closed doors IP, was there any creative things that customers started doing? That was spawning more things to put on your roadmap.

Speaker 2: 27:50

All the time happens? Uh, all the time? Uh, maybe not. I don't have any of that I could say right now. Most of what I see now is innovative. Things are in how they prompt what they ask and how they prompt. So that's where it is at the moment, the innovation that very little has needed for those innovations to happen. In the tool, it's more in the sense of what is it used for. It's where the innovation is mostly. But in the past, for example, it was just one of our customers who decided to use it in a live event, but before that it was only asynchronous use. So major changes happen all the time along the way in the sense of what customers come up with.

Speaker 1: 28:32

Right, okay, it's interesting because I was also chatting with actually another Finnish company, pandatron, who has an AI coaching bot, and I think we were chatting about that at the Change Leadership too. So it's interesting. They're quite different platforms obviously. Like there's more of a coaching tool, uh, and this is more of a sense making tool, and I think that's you know, when I talk about platforms, it's it's ai platforms. The skepticism was around tool vendors who are taking their templates and their course material and their processes and putting that in a tool, because that's obviously going to bias the tool towards their lens. Yeah, and I've seen this like some enterprise companies I've worked and I just thought was completely crazy that they outsource everything about how they do change to a vendor so their vendor creates their step-based process and they take it and they follow it and when something doesn't work, they go back and yell at the vendor and the vendor changes it and sends it back, and I just think that's crazy.

Speaker 1: 29:35

But now, when you're going to be buying those types of platforms, you're, I think, your focus becomes very narrow.

Speaker 2: 29:42

So that's the one thing that I really appreciated about the demo was it was all about.

Speaker 1: 29:47

It's the data that goes into it. It's not. You know, you're not prescribing. Hey, you should follow these five steps of change method X in order to do whatever. Yeah, strictly sense making.

Speaker 2: 29:58

Yeah, yes, and it's with most of our customers. It's taking the existing way of they do and building that, putting that in there, or taking some of our templates and changing those in the way they do. In that sense, and the AI is completely generic in that sense that it doesn't, it's not trained on any particular material for this, so it doesn't close out anything in that sense. Right, right.

Speaker 1: 30:24

So is it, and don't answer this if you don't want to is it closed off to the internet browsing? So, for example, could I ask it, you know, given the feedback from this space and this change or whatever it is, could you recommend a method or a tool or something that would help us blah, and then it would search the web or the internet and it might find well, use method x and here's how you execute it. Or is your models more trained to just make sense of their content and it's not doing that other?

Speaker 2: 30:56

stuff. So it is. It is closed for internet browsing. It is practically what is. Ours is closed environment version of gT 3.5. We're looking to claw the tree of a GPT-4 only when they've gone fast enough or this kind of thing. But that's what it practically is, but that's the closed environment version of that. So of course it's used all the internet in the history of its training, but it cannot access now anything outside of its past history. Okay, so as we get into the wrap-up, access now or anything outside Right or from its past history.

Speaker 1: 31:26

Okay, okay. So, as we get into the wrap up, if you think about the marriage of AI and change and transformation, what do you see is on the immediate horizon that's going to help change practitioners, facilitate change in a better way, and maybe something long term like what do you see long term happening in the AI and change space In?

Speaker 2: 31:54

short term, I see a kind of major efficiency gains for change managers in their own work. A lot of things become a lot easier to do with the generation and making sense out of that. Also, I see innovation how the kind of the change activities, how they're currently run, maybe combining those, supporting those and then easier the sense-making process or more iterative, more agile in the sense is where I see where AI will help in that sense in the short term. In the long term, my main insight is that I've never seen anything progressing like this that I've been I'm not the biggest IT I hesitate to use the word nerd because it's negative annotation, not in my mind, not the biggest IT nerd, but as closely I've been working with IT development and IT world. I've never seen anything move like this. It's the amount of money that's poured into AI at the moment is incomprehensible for the development. We get the benefits all of our consumers and we get the benefits out of that race. What's happening at the moment, but the pace is incomprehensible. The pace, pace. Things are pushed forward. Very hard to say what's the innovation and what are the things that come forward from that. Only thing I can see is that it's it's a interesting and exciting and I'm very curious to see what happens and what are the innovations in that and the other side of that. I see of that as excited I'm.

Speaker 2: 33:26

I'm about the pace, what's the technology is moving forward. In my experience it's still always the. The bottleneck of that if we would want to call it or, or a maybe there's a better word but bottleneck, without the negative annotation of that is just a human change of pace that most of the people I talk to, like other kids' parents or these kind of things, it's still for them. Oh, I tried it and once it was fun, but I never used it since, that type of scenario. So I think it's still very much kind of early adopter pace for the most of the cases by the wider public. So it'll take time, It'll take years before it's kind of become so mainstream that it happens with most of the people who will adopt that and that's interesting. What happens with all that?

Speaker 1: 34:17

Yeah, because the learning curve is so steep initially. Yeah, and I think you know, just talking about the change leadership, when Yvonne asked me to speak there, I think there was something like 10,000 custom GPTs on the store back then and by the time I did my talk there were 3 million, and that was in two months. So it's one of those things where you know, change management is conservative by nature and there's a lot of wait and see. Let's make sure we get the strategy right. And I'm kind of in the camp that is, it's impossible to get your strategy right. Yeah, I think you should really just try yeah, latch on.

Speaker 1: 35:01

Yeah, do you use? Do you use custom mods? Do you use a lot of custom gpts? Yeah, I have a bunch. I have, uh, an experimenter tool where someone can come in and they can create an experiment. Yeah, and my chat gpt, my open ai assistant, is specifically trained on change experiments and lean change. So they type in their experiment what the problem?

Speaker 2: 35:16

is they're?

Speaker 1: 35:16

trying to solve what their hypothesis and their diagnostics and measurements are, and then my people give them, coach them on feedback so you created a customer yeah, I've got a bunch of them, I've got you created a possible yes okay yeah, they all just hook into um, uh, the assistance, api, yeah and uh, everything in my ecosystem from summaries, like I've got a lean coffee tool so you can do your lean coffee cards, you can add your comments and stuff, and then you can ask my, my bot Mr Zircon is his name, cause I thought it was funny and it will say generate me a summary of the lean coffee and tell me, based on what you've figured out, are there three experiments that we should be trying, given what we talked about, and then that can actually get fed into my experimenter tool.

Speaker 2: 36:03

Yes, okay.

Speaker 1: 36:09

All to my experimenter tool. Yes, okay, all right, okay, but it's, it's like it's.

Speaker 2: 36:12

It's obviously trained on the lean change ecosystem stuff, but it's not excluding anything. Yes, okay, yeah, um, and it's transparent about that.

Speaker 1: 36:16

So it says like here's what I've been trained on, here's what I can do and can't do, but it's all about very similar it's give me some data I will analyze the patterns and I will give you feedback and some options that you might want to try, but you, as the team, still actually have to do it.

Speaker 2: 36:31

Yeah.

Speaker 1: 36:32

Yeah.

Speaker 2: 36:33

Yeah.

Speaker 1: 36:33

Yes, so the um, the wait and see stuff is is definitely interesting. I had a call with someone who was at the conference that I didn't get a chance to do a demo of the virtual change team with her, so we just had like a half an hour chat about it and people can get start like they're worried about signing up for a platform that's fully baked that they can convince their stakeholders to use. And I said well, you know, you can just go to platformopenai, create a free account, create an assistant, open up the playground. It's all free. Now, if you want to deploy it, then you have to hook into their api and stuff. But if you're a non-technical change person, it's crazy easy just to do this and then you can show your leaders what it's capable of and stuff.

Speaker 1: 37:20

Yeah, and that's the thing I really like I mentioned a couple of times I really liked about the house-based solution is that it wasn't pushing a method or a framework or a bunch of easy answers like you see on LinkedIn. It's all based on helping them understand where they're at and what they can do. Yes, yes, you put it very nicely. Thank where they're at and what they can do.

Speaker 2: 37:38

Yes, yes, you put it very nicely. Thank you, jason. Right on, this should be a commercial Sponsored by no.

Speaker 1: 37:47

I'm a fan of good ideas that help us, help people make sense of their context, and that's really that's the whole stance behind my work for the last couple of decades is that.

Speaker 2: 38:13

Have you been surprised or?

Speaker 1: 38:14

what's your take on the pace that you've seen a larger enterprises adopting AI just in a generic use or any use, mostly customer service?

Speaker 2: 38:18

chatbots.

Speaker 1: 38:18

Yeah, um, mostly customer service chatbots. Yeah, um, there was actually a pretty funny story about uh, I don't know if you heard about air canada, where their bot gave away a free flight or something like that that it wasn't supposed to.

Speaker 2: 38:27

Yeah, yeah, the robots have gone rogue but I would say mostly that.

Speaker 1: 38:35

uh, and content generation right, like Like every meeting summaries, Zoom has built-in AI summaries, the integration with all the Microsoft products. So you know AI is integrated with Word, it can help you with emails and stuff. So mostly stuff like that.

Speaker 2: 38:52

you know, individual productivity use yes, yes, yeah, and that's where I've been. I've been maybe slightly surprised how, because I still number of our larger organizations that I speak with that they, they don't have individual access for it's forbidden. Still, the individual access least. Personally, my productivity went up maybe 20-30% with the support of AI and that's a big, big number for an enterprise multiplication. If they can get individuals to use AI to help their private productivity, the pace has been surprisingly surprising. But maybe it's a they haven't done a calculation from that point of view. I think they're afraid of. Maybe it's a they haven't done a calculation from that point of view.

Speaker 1: 39:40

Yeah, I think they're afraid of giving away secrets as well. Yes, of course you know which is understandable, but I think that's a lot less controllable, especially yesterday. So for folks listening or watching, it's June 11th, so yesterday was Apple's developer event and they announced open ai integration with their, with their products, so everybody, and you could do that now with apps anyway, you can just point your phone at the screen and you can steal sensitive customer and whatever information. So unless, uh, people might have seen elon tweet out, you know if apple does this.

Speaker 1: 40:17

I'm banning iphones from our company. I'm like dude, people could do this already. There's a thousand ai apps that you can do. Yes, you know so I've got. That's a huge fear, uh, for people, but I think that shouldn't stop them from using it. If anything, that should make them want to use it more, because you're going to lose the people who you're going to need to help you with ai, because if you lock it down, they're just going to go somewhere else where they can use it, especially in today's hybrid workforce, like you're not shackled to a company anymore. You can work with anybody from anywhere, which. But it's tough. You know the banks in Canada. It's tough for them to adopt this stuff. It is Right. I think it can help a lot with fraud prevention, like some of those algorithms and fraud prevention are fairly sophisticated, and I think ai can help a lot of those things but uh yeah, yeah, and with one bank that will remain nameless, it might help them launder more cartel money that they were just uh caught doing a couple of months.

Speaker 1: 41:14

I always like to take a jab at them because I just think it's hilarious. Um, it probably shouldn't be hilarious, but but so is there anything that you wanted to wrap up with before we? I could talk all day about this, but same Any advice for people who might be skeptical.

Speaker 2: 41:35

If you're skeptical, be skeptical and ask some of your friends to try it and show it to you so you don't have to do it necessarily yourself. Ask a colleague to give it a go and see where they come up with it, or something Cool In that sense. But that's still I suppose the only way is to give it a go and looking into the prompting a lot. I liked Jason. Maybe you shared it here. I liked your prompt structures as a part of your speech. If you share them somewhere, people who listen and go and look at them. They were very good. It's not difficult, but it takes a little bit of thinking, a little bit of looking to understand the prompting.

Speaker 1: 42:14

Yeah, I think people have to unlearn that they're talking to a machine and just consider it a human. Yeah, I tried to ask Gemini. I was just interested in the sustainability aspect of AI, because everybody's talking about all the cool stuff you can do with it and nobody is considering that if you trained an LLM, a large language model, it would consume the same amount of power that could power a town of 50,000 people for a week. Yeah, I think and I know that's an impossible thing to say. So when I started talking with Gemini, I said I'm going to ask you something that I'm pretty sure is impossible to figure out, but this is what I'm after.

Speaker 1: 42:53

I want to try to see if there's a way that you can compare power usage to know, searching a structural database versus training uh, training a large language model, you know, are we talking 10 to 1, a thousand to one? And it said yeah, you're right, it's impossible to make this. Here's some things to consider. Here's some things you might want to ask me. I'm like, all right, so I followed that path down, path off.

Speaker 2: 43:14

yeah yeah, so oh, I did that. Oh, that's nice that they're here for things you might want to ask for me. Oh yeah, did you probably specifically to get that response from it, or was that I asked it? I Specifically asked it. Yeah, I said.

Speaker 1: 43:27

I'm. I don't even know exactly what I want to get out of this. This is what I'm trying to do. What questions should I'm asking you? Stuff like that? Yes and yeah, we, I think we, we forget, and we think we have to type in proper syntax or talk to a machine, but you can just talk to it like it's a business.

Speaker 2: 43:45

Yeah, that's been quite often. My goal, too, is exactly that. Hey, I want to get here. What questions should I ask from you that you can help me to get there, and what do you need to ask me? The questions so that you can understand me to get us there. Yeah, yes. But for the environment impact. I just saw some stats. I think Microsoft did 28% up from last year because of the impact of the AI and the investment in the server A 28% increase in power usage, no in emissions, no in emissions, oh, in emissions In CO2 emissions from Microsoft, just because of the AI development, yeah, I heard there's I don't know what company it is, but they're making.

Speaker 1: 44:35

Maybe it is OpenAI that's trying to develop hardware like cpus that are specifically built uh to process ai with lower power okay I'll have to fact check that, because I don't know if that's if that's true or not. I remember seeing something in my feed about it, but, um, it's going to be a, it's going to be a problem and very few people are focused on this because they're like well, I, I want ai to make my presentation and then I can go to the beach for the afternoon, you know yeah, who cares if the planet's going to be dead in 10 years because of all this extra power consumption.

Speaker 1: 45:09

But there are, uh, a few companies that are looking at sustainability and stuff.

Speaker 2: 45:14

But yeah, yes yeah, yeah, the overall, if the capacity building race is so high that it's it's a sustainability isn't challenged there. I like the, finland has nice example for the. Google has done good job in Finland for the, for the, because it's cooler climate, so they need a lot less of cooling in the sense that it's more like in the sense windows open in the winter for the cooling of the system. So much less power is spent on cooling the systems. We're moving our data center up to Lapland.

Speaker 2: 45:50

Which in a sense I would say Canada would have also a lot of potential for that, in the sense that yeah same thing for the, for the cooler climate.

Speaker 1: 46:01

Yeah, yeah, they're basically the same. I like to say that finland and canada are basically the same country, except, uh, finland has better coffee and more natural food. Okay, yes, yeah, yeah, yeah, it's too bad we don't have like a little high-speed tunnel that links directly. I might actually be in finland in the fall, uh, in september. Um, I'm supposed to go to stockholm. So I just emailed everybody I knew in finland and I said I'm going to be in stockholm in september and I need an excuse to go to helsinki.

Speaker 2: 46:31

Yeah, okay oh, very good, very good. Hey, let me know when you're going connecting, maybe meet my colleagues, uh, colleagues, ofagues of Confine my Dolls.

Speaker 1: 46:42

Cool. Cool All right, and that is a good wrap up. So tell people where they can find you if they've got questions for you and where can they find out more about HowSpace.

Speaker 2: 46:54

So HowSpacecom and H-O-W, how, that's the question, and then spacecom and can find me from there. Can find more questions and reach out from there.

Speaker 1: 47:08

And actually I do have one last question how did the name come to be HowSpace? How did the name come about? It's an interesting name.

Speaker 2: 47:17

Because that was always. It was the Simon Sinek's, the why, how, what part of that, and that's always the how, space. Just how do you do that? How do you? It's solving that puzzle. The content of the what or the why is different, but then how to run the process, how to make that happen. So there was space for the make, the how happen.

Speaker 1: 47:48

Excellent, all right. Well, thanks very much for taking the time to chat today and for everybody watching on YouTube or listening. Show notes are going to have links to everything that we've chatted about, so thanks for joining.

Speaker 2: 48:00

Thank thank you, this has been fun.