Innovating with AI While Honoring Traditional Skills with Douglas Daley

Mark Smith: 0:32
Welcome to the Power Platform Show. Thanks for joining me today. I hope today's guest inspires and educates you on the possibilities of the Microsoft Power Platform. Now let's get on with the show. In this episode we're going to focus on how a technology-centric technologist can evolve into a business-centric technologist and why that is so critical in the modern marketplace. Today's guest is from Sydney, new South Wales. In Australia. He works at IBM as an associate partner. He helps finance procurement, supply chain leaders, improve organizational performance and delight customers and employees by modernizing how their teams work with Microsoft technologies. You can find links to his bio and social media in the show notes for this episode. Welcome to the show, douglas Hi Mark.

Mark Smith: 1:30
Nice to be here Do you refer Doug or Douglas?

Douglas Daley: 1:34
Well, I think most people call me Doug.

Mark Smith: 1:36
Yeah.

Douglas Daley: 1:37
Yeah, but do you prefer Douglas? I actually don't mind either, to be honest, but most people call me Doug, as you know.

Mark Smith: 1:45
Yes, before we get started. And, by the way, I know Doug quite well. I've had the odd drink and meal with him over the last three years in the time that I worked at IBM. Are you still at IBM?

Douglas Daley: 2:01
By the time, when's this going to actually air? Months by the time this airs. No, I'm off to PwC. Off to PwC.

Mark Smith: 2:11
PricewaterhouseCoopers.

Douglas Daley: 2:12
I am, so I'll be there in about three weeks' time. Very much looking forward to that change.

Mark Smith: 2:17
Awesome, Awesome. Okay, Food, family and fun what do they mean to you?

Douglas Daley: 2:22
Oh, my whole family revolves around food. I come from a family that absolutely loves to eat and travel the world eating, and so I've got an 18-year-old stepson and a five-year-old daughter and I guess we spend the majority of our leisure time around Sydney finding new spots to eat. I think my five-year-old is the most excited little kid to walk out the door on a Friday night and find somewhere for Korean food in the city, for example, and our eldest is half contemplating a career in food. I think that's how passionate he is about it. But we're off to Japan for Christmas this year and that's going to be a three-week eating trip yet again.

Douglas Daley: 3:09
But yeah, that's food and family. Family's obviously a lot of fun. But you know, I had this sort of moment, maybe about 18 months ago, where I thought, okay, my daughter's getting to the point where she sleeps pretty well now and there's a little bit more sanity in the household and it's time to get back to doing something just for me, taking that kind of moment out. And so I picked up, like I generally do, niche little hobbies pre-industrial furniture making, so hand tool, woodworking, essentially Nice. So that's what I make some time for every week and travel for various courses. These days I'm off to do a chair making course in in Victoria um between IBM and and the new gig I'm very much looking forward to that.

Mark Smith: 3:59
Yeah. So one thing that struck me about Japan in my time there was miso soup is never as good as what it is in Japan.

Douglas Daley: 4:09
It's probably a lot of things that fit that bill yeah.

Mark Smith: 4:14
We're at Mount Fuji and, uh, uh, we, we ate at a local's place there and, honestly, it will be a memory that sticks with me forever just the quality, the, the flavors, the richness, the abundance of food and, um yeah, japan, amazing, amazing, um, so the woodworking side of things. So you're making things like is it just furniture or is it other things?

Douglas Daley: 4:45
it's, it's predominantly just furniture that's a big area, right, it's a lot you can do it's a big spectrum and right now I guess my core interest is in developing skills in two areas both chair making traditional Windsor chair making, so essentially mortise and tenon joinery, where everything joins into the seat and the seat is the rigid structure.

Douglas Daley: 5:09
And to make it super strong, you essentially try and build it from parts that are just split directly from a log with an axe. So it's um, it gives you that real appreciation for the medium you're working with and I think it's a huge amount of learning there. For me it's actually the the slow pace but absolute quality that gets produced, um, as well as the reduction in how many tools you're actually using. You know it's something that you could do without a single power tool. It's something that you could do without electricity, just by you know, natural daylight, and when you start to think about what that means and what it meant during history, it starts to bring up concepts like the seasonality of work, because you know in what we do you can get stuck into a project pace where it's a grind for two years on a big program of work.

Mark Smith: 6:07
Yep.

Douglas Daley: 6:08
And you never get that moment out to refresh and there's no doubt that that's impacting the quality of what gets delivered. So I think I was drawn into it just because I've had a long interest in woodworking and it was a hobby to pick up and I think there's a re-emergence of interest in craft at the moment A hundred percent there is. But I've found it much more interesting in terms of just bringing up ideas about historically how we worked and how they might be really meaningful to building better habits and moving past the burnout culture that's evolved in, you know, the white collar factories of today.

Mark Smith: 6:47
I love this. This is awesome. How can a technology centric technologist evolve into a business centric technologist was something that we have discussed and I really want to as an. I think it's so critically important in that my observation over the last probably seven years is that tech skills can be taught and if you have a bias towards them, you can pick them up. You can turn your hand to any type of tech skill if you're that way inclined. Not everybody wants to do tech, but if you are.

Mark Smith: 7:19
But what I am really seeing lacking is this business-centric view of tech skills. And I say that just coming off a call with a very large organization, maybe the largest in the world, a very large organization, maybe the largest in the world and trying to reposition them, or a group of people within that organization, to not lead with tech as in their product and lead with what are the challenges of the organization. And let's map our tech to that rather than just preach product, product, product and really understand fundamentally the need, and then the product will come. You just don't need to be lead with it and and almost, I feel to some degree ram it down someone's throat without understanding. And so these business skills, these business skills that allow you to engage with senior positions in the organization, not just the IT department, but senior.

Mark Smith: 8:18
It's funny, because we're in software, right? People think, ah, let's just go to the IT department. And the thing is, as a rule, IT doesn't control budgets, right? Strategic decisions for an organization are controlled by the executive layer, and I'll remind you of the last project I was involved with you on, which was a financial institution, and we got that deal because we engaged with the executive, who gave us the role to then engage in the senior architects in their organization and further down, right. So for me it's become a pattern of really importance of that business centricity. But what does it mean to you?

Douglas Daley: 8:54
I think, Mark, we can come at that from two different angles. It means two things to me. One, it means being able to apply a diverse perspective to a question that you get asked. That's probably quite an abstract answer, but the second thing it means to me is how you go about building knowledge throughout your career and what's the system and what's the model that you're trying to achieve. If we just start with an example, if you've got a technologist that's going in and looking to work with a team in a business, let's just pick up an example of that team Maybe that team's procurement example of that team, maybe that team's procurement. If you're relying on people in the business to translate their understanding of good ideas into a set of requirements that you can build in the technology, then we're getting to the point with AI, coding assistance and configurable tools where why wouldn't those business people just go and do that themselves? So if you can imagine individuals that can talk about good ideas and those good ideas span both how technology can be employed and what good looks like for that team in the business, that, for me, is a business centric technologist. And if I just break down my mental model for the way a business works and where you can change a business is just four layers. You've got strategy, and this is where typically the MBB firms like McKinsey are operating right and they're looking at what's the revenue model, what are the regulatory impacts, what's the market influences, what's the ambition, what's the opportunity, what's the the market influences, what's the ambition, what's the opportunity? How does that translate into changing that business for better? Um, better being obviously quite a subjective term.

Douglas Daley: 10:43
Beyond the strategy layer, then you've got how does the business operate right, and so that's the operating model. It's the processes and the roles in the business. Then you've got the capabilities. So a business needs a bunch of standard capabilities, like it needs an accounts payable team and an accounts receivable team and potentially the customer service team. And within that capabilities layer, then you've got the technology that enables that, and so some of that technology could be generic, like accounts receivable, accounts payable. Some of it's going to be technology capabilities that are quite specific to an industry, like if you're talking the steel industry or anything that deals with a commodity, then you might have a dynamic pricing capability. It's a tech enabler for the way those businesses work.

Douglas Daley: 11:32
So you've got strategy, you've got operating model, you've got capability. And then the last layer and this is the one that I think people often struggle to tie back in is delivery. How does a business deliver its products and its services? And if you combine those things together, really delivery is your traditional service, design type and user experience layer. The capability layer is your traditional enterprise and solution architecture.

Douglas Daley: 11:59
Your operating model is that traditional functional transformation with operating model transformation, and then your strategy layers are all your strategy methods, and so a business centric technologist has to develop an appreciation of all of those things in addition to just their core technology experience. And the way to go about doing that or to think of that is, I think, becoming a T-shaped consultant. You've got a breadth of knowledge and a depth of knowledge, and every year in your lifelong learning journey you've got to work out which little bubbles do I want to put in to my breadth of knowledge that year? Which little bubbles do I want to put in to my breadth of knowledge that year, and is there anything in there that I've already developed some appreciation of that I actually want to go deep in and make part of the vertical part of my tea right. That, for me, is a business-centric technologist and they should be able to carry a conversation or bridge the gap between people that have little appreciation of technology and people that have little appreciation of business.

Mark Smith: 13:03
Yeah, interesting. So how does one go about doing it Like specifically the business side? Right, Because we feel we're strong on the tech side, but the business side, how does one go about?

Douglas Daley: 13:13
So I spent a good portion of my career at KPMG. Um, I think that experience is part of the desire to go back to the big four for me as well, and it's about being able to be part of a diverse community of practice. I think so often our community of practice within the technology realm is quite constrained to people working within our technology domain like, for example, people that are power platform and CE practitioners or biz apps practitioners, alongside some other technology skills that support that, from data through to integration, through to cloud. The first step is to actually have access to diversity in your community of practice, and that could be embedding yourself in a business or embedding yourself in an organization or a project that gives you that access. I think that's step one, but the more critical step is how do you actually go about building that knowledge? And it's almost like creating your own knowledge base right that you can draw on when you're preparing for a workshop or going into a pitch, and I really love some work that a gentleman by the name of Tiago Forte did on this. And I think if you look at this model, it essentially breaks down into both a pipeline for acquiring knowledge and a structure for managing knowledge, and that pipeline is quite simple Code capture, organize it into a structured repository, distill it down into something that's actually reusable and then have a method of expressing it in content. And I think this is where, as technologists, we can probably push that idea a little further and essentially look at the concept of AI to pull from the knowledge base that we build over time.

Douglas Daley: 15:14
But then that knowledge-based structure that he creates is quite simple. It's projects, right? So I'm working on a project, and that could be something at home, or that could be a project that I'm delivering as a professional, and this doesn't have to just apply to your professional life, it can be more broadly used. Then you've got the idea of areas, and this is just where do you build a habit around knowledge acquisition? How, when you're listening to a podcast, do you actually capture a snippet and put it somewhere? How do you read a book on Kindle and take your highlights and put them somewhere? How do you grab things from the web as you're reading them and store them somewhere? And that's areas. This is this idea that you want to turn curating knowledge into a habit. Then you've got resources and then, finally, you keep it clean. So you've got archives, so para in essence, I think that's a really great structure to pick up.

Mark Smith: 16:58
Did he because the name rings a bell to me did he write the book the Second Brain? He did, yeah, yeah.

Douglas Daley: 17:03
Absolutely.

Mark Smith: 17:03
I've read, I think, two or three of his books and definitely a model. I've been working that power model into my life around knowledge acquisitions and putting it in a format that's retrievable right.

Douglas Daley: 17:15
Yeah, and I um the areas section and I break it down into strategy, operating model, capabilities and delivery, and then I take some standard models. Like there's an organization called apqc. They, I guess they're a process standards organization. They produce a table of contents for processes, a process taxonomy. That's cross industry. So I start with that and then every time I'm gathering some knowledge if it relates to order management and delivering goods or it relates to dealing with escalations in customer service, I just put it all under the right kind of process. That way, if I'm ever having to work with an organization and the scope of work covers those process areas, I've got a repository of I don't want to use the term better practice or leading practice, I think that's not necessarily true. It's a repository of good ideas and I've got to work out if they're relevant or not.

Mark Smith: 18:17
So my reason for reading his book books when I read them and using this power process is that I could see where AI was going, because I've only read them in the last 18 months and I have been collecting knowledge for probably 15 years, and what I mean by that an archive of all my emails, every receipt, every medical checkup, every manual for any tech or tools I get. I digitize and put it in there, and I've just been collecting, and my two main tools, of course, has been the archive function on email. That's pretty obvious. It's easy to create a PST file and extract that off, and I've never gone back yet, but I'm hoping that one day that the AI will be able to go take a look at that and perhaps understand me better, or what do I want to get out of it. But then the other one I've always used is Evernote and I've got, I think, 10,000 artifacts now sitting in Evernote and 10,000 artifacts now sitting in Evernote. And then, of course, the OneDrive storage, which is now where I'm running my new power system within OneDrive and feeding that in. So AI has been.

Mark Smith: 19:27
I always felt that in the future, tech would get to somewhere, and why I did this is that of all that history that I've been in tech, I would be able to derive greater value in the future when the tooling allowed me to extract that In the area of AI. Now, with consulting, how do you see that changes? And I'm not really looking for so much. Yes, it's going to allow us to build new software, but how does it apply to a technologist in a day-to-day basis? How can they use AI? A technologist in a day-to-day basis, like, how can they use AI? And really I'm asking how you're using AI in your day-to-day roles. That is an enabler for you.

Douglas Daley: 20:07
So I've got two key perspectives on this. One I'm using AI so I can get time back for deep and creative thought, because I think one of the biggest risks of AI is that we stop producing interesting, novel content. You know some of the estimates are, by 2030, such a large percentage of all the content that's publicly available will have been generated largely by AI that you can't even train AI on that information. So one for me, it's about creating time to do really meaningful work, and meaningful in the sense of that. It's well thought out, it's creative, it's innovative, it's pushing the bar forward somewhat. It's taking time for mastery. Two, I'm kind of using it the way you were just talking about it. It's how do I take a whole lot of knowledge and and express that knowledge in a context? So I'll give you an example.

Douglas Daley: 21:09
One of the little pet projects I've had on the side at the moment is taking everything that's on Microsoft's documentation of biz apps, so particularly the documentation around configuration of FinOps, for example, being able to extract that, put it into a knowledge graph and then relate it back to some of those things I was talking about in my para model around a process model and a set of capabilities so that I can quickly go. Hey, I'm going to run a workshop and it's going to talk about the source to pay process. I want to specifically focus on guided buying and all the design decisions and configuration decisions that the customer would need to make in relation to setting up the procurement and sourcing functionality in FinOps for a guided buying process and the entire buying channel, which is right through to payment. In essence and there's a few interesting things there I think people are getting pretty familiar now with RAG retrieval, augmented generation how do you get a document to be turned into something that AI can kind of work out? Oh, these bits of the document are relevant to the question that's been asked and spit you back something.

Douglas Daley: 22:31
Where that becomes more interesting is when you start to take a bit more control. And so, one, if you model it in a graph to begin with and you're defining the relationships between data in a more structured way, which can lead to a more usable output when you need precision in consulting. Two, if you start to combine it with tools like Langchain or Semantic Kernel from Microsoft or frameworks I should call them then you can actually start to build quite a structured pipeline of question and answer with models to get really well created outputs from them that are quite sophisticated. More than a simple, you know, a simple prompt might give you with a document that's, you know, attached to a GPT-4 model.

Douglas Daley: 23:21
And so I'm starting to use AI and it's actually been a reason to develop deeper technical skills or muscle that I haven't used in the last 15 years, because the interesting thing is, when you go just a bit beyond what's almost turnkey and you start to apply that technology to the exact problem that you spoke about, there's all this stuff that we've gathered over time. How can we use that competitive advantage to get rid of some of the time-consuming work that we might be doing today? But you know, mark, there's also another really interesting point there. I spoke about this idea of diverse community of practice.

Douglas Daley: 24:06
For me, the greatest thing ai can do for a lot of consultants right now is, if you ask it some questions, it can give you a response that's formulated from diverse knowledge yeah you know, it's automatically starting to point you in the direction of maybe some of the things you might want to learn yeah, one of my common prompts is show me where I've got gaps in my thinking.

Mark Smith: 24:30
So I'll give it a bunch of stuff, yeah, that I've created, and say show me where I've got gaps. And of course it's filling that gap from that diverse knowledge. And it's interesting you talk about that chaining, because that's what I've got heavily into lately is a multi-prompt that is all automated. Yeah, so it has output that is then fed back into input to another query and refined down. So we're talking about five or six prompts. But I've set up as a automation because they're going to use the model over and over again. And it was interesting what you said there about the problem is, you know, in the near future there will be no new data to train on.

Mark Smith: 25:10
I find that, like earlier today, I was putting together a presentation and I asked, you know, ai for some advice around the content of a particular slide and it took a direction that was just not where I wanted to go. That was just not where I wanted to go. I wanted to refocus it on a specific area and I feel that, although AI is creating content in that situation, for me it's highly curated, because I go back and go. Yeah, sorry, that's not what I meant. I'd much rather want to focus on these three key areas. And can you expand those? And so obviously it goes out and it goes.

Mark Smith: 25:47
Okay, these concepts are what I need to highlight on, and they become the three or four pillars of whatever my slide was going to be about. So, although AI is creating that content, I think a lot of it is still going to be massively human assisted, because its direction was not the direction I needed to communicate with this customer. It was different. That I knew just from the conversations and the engagements that I have had. So I wonder if that knowledge will still be highly valuable and that it's still going to be human, curated, if you like. Yeah, I hope so.

Douglas Daley: 26:22
I'm certainly not an expert on model training, but I think the key thing is that we just have new ideas still and we're still putting enough effort into creating new ideas and expressing new ideas. Have you got a favorite AI-based tool at the moment?

Mark Smith: 26:39
Yeah, well, as in, gpt has always been my favorite. As in, I haven't been like I've got Claude. As in right now, if I open my browser I've got on my first time opening a browser I have six different AIs come up that I'm paying for, and so if you look from a text perspective, it's really ChatGPT and Claude are the two that I go to, and then I have Perplexity as well that I pay for, which gives me much more if I'm wanting to make sure I can solidly reference, because I've found sometimes I'll ask for a quote from somebody and it will make up quotes and attribute them to somebody and I'm like where's that from? And then it will go. So I had one. It said it was Brad Smith. The chief legal counsel of Microsoft said the quote. When I drilled into it, it said no, no, no. Chief legal counsel of microsoft said the quote. When I drilled into it, said no, no, no, it was such here that said the quote. When I drilled into it further, it said no, it was bill gates that said the quote, and what I found it was a fabrication of three snippets of three different things that they had said, that it had put then into a quote and just, you know, gave it one of them so much more.

Mark Smith: 27:45
Anytime I have that, I absolutely fact check. You know, gave it one of them, so I'm much more. Anytime I have that I absolutely fact check. You know that it is. If I'm going to give statistical data, I fact check that I can tie it to a body of work that I can reference, with the url, for example. But then on the image, I have paid for mid journey for ages. I have um, I've stable, um, stable diffusion. I've installed full stable diffusion on my desktop. So we're talking about 10 gigs odd of that LLM running locally and then the other one that I have which I've just been running recently and I'll see if I can do this without interrupting our session. My probably favorite image generator at the moment is our Lambda chart, as I have that Ideogram. Napkin AI is the other ones that I have automatically open. So have you looked at Napkin AI? I haven't looked at.

Douglas Daley: 28:42
Apkin no.

Mark Smith: 28:43
Wow, you want to check it out so you can create your body of work and you want a really it out as in. So it will. You can create your body of work and you want a really good illustration for it and it will generate not a, an image like a photo or anything like that. It'll generate a csv diagram of your concept and it's totally free at the moment but the quality of it to illustrate your concept? So, let let's say I used the other day I wanted to illustrate RBAC, right, role-based access control, and the diagram it came up with. Just I knew that when I explain this to other people they're going to go, oh yeah, I get it. Like the diagram makes sense. And so for finding that for technical illustration, such a powerful tool, such a powerful tool.

Douglas Daley: 29:28
Yeah, I have to have a look at that. One of my favorites at the moment is notion and notion ai oh man, I'm just deciding whether I shut down notion.

Mark Smith: 29:35
I'm just, I haven't, I just it feels cluttered to me it feels cluttered.

Douglas Daley: 29:38
I think it's about again. I've used para as a structure for notion um, along with those layers of business design, and and I've built it into a database model of capabilities and processes that I can attach everything to. But the thing that I love about it and I think this could be really a revolutionary tool for people that want to become business-centric technologists is you can ask it to describe a business process and you can ask it to tell you what are the controls like workflow, approvals or reports in that process. But then you can say, hey, could you draw a flow chart of that process? Is that right? And it will generate. There's something called Mermaid, which essentially lets you, in almost a markdown style syntax, describe process flow diagrams and some other uml style diagrams as well, and so it'll draw a diagram of the process, and the time that that takes in client delivery is a lot of time massive, and so that's part of notion.

Mark Smith: 30:40
I'm going to crack that up after this and and and have a go with it, because one of the things I haven't touched and I've got a full version of Notion I haven't actually touched the AI functionality in it, just haven't gone there, so it's the only AI that I pay for is Notion AI?

Douglas Daley: 30:56
No way, because I've been lucky through work to have access to. Certainly, ibm has its own suite of tools for this. But yeah, notion AI is the only AI I pay for, and I think it's quite a phenomenal tool for both building the knowledge base of a business-centric technologist, but also developing all of those consulting artifacts that you'd use to deliver a business-centric, technology-enabled program of work.

Mark Smith: 31:28
Incredible Well we're at time. It's been an interesting discussion. Thank you, douglas, for coming on the show. Well, thanks for having me.

Douglas Daley: 31:36
Mark.

Mark Smith: 31:38
Hey, thanks for listening. I'm your host business application MVP Mark Smith, otherwise known as the NZ365 guy. If there's a guest you'd like to see on the show, please message me on LinkedIn. If you want to be a supporter of the show, please check out buymeacoffeecom. Forward slash NZ365 guy. Stay safe out there and shoot for the stars.