A Deep Dive into the Future of AI and Work

Mark Smith: 0:01
Welcome to the Ecosystem Show. We're thrilled to have you with us here. We challenge traditional mindsets and explore innovative approaches to maximizing the value of your software estate. We don't expect you to agree with everything. Challenge us, share your thoughts and let's grow together. Now let's dive in it's showtime. Welcome back to the Ecosystem Show.

Mark Smith: 0:25
Today, in a rare occurrence, I am the only one on the show, so I've never done this before. Although I've recorded over 600 podcasts where I generally interview a guest, I've never been in a situation where I am just doing a solo podcast, just with me, and it's going to be a bit of a surprise to the other crew. Will, anna, andrew and Chris are all out and about. They're either traveling internationally or they're at the Scottish Summit in Aberdeen presenting, so they don't know that I'm recording this and really, who knows what I'm going to cover and whether it's going to be interesting and whether this is a model that works. But I just didn't want to skip an episode because that's a bit of my um my, I don't know what it is, but the way my brain works is that I don't like to miss things. I like to be scheduled, I like to produce results consistently and if people are waiting on an episode to drop. I want to make sure we're going to drop an episode, but what I wanted to talk about is a couple of things, just news, top of mind.

Mark Smith: 1:34
At the moment, dynamics Minds have announced their call for speakers, so I think it's a great opportunity to speak at that event. Honestly, it's the most epic event I ever attended. Well, started this year this is for 2025, of course so I highly recommend you submit your sessions. There's a range of tracks that are covered and it's such a great opportunity and it's the only conference that I've been to that has a very festival feel. So it's hard work during the day conference sessions, that type of thing but then at night, oh my gosh, they know how to party and you know whether it's bands, live music, entertainment, magic, as you can imagine, in Slovenia. This is just an experience that, if you're in the BizApps community or really the wider community of AI now is becoming, you know, a big topic of discussion across all our ecosystems. You want to attend this event. Um, I'm about to book my ticket ready for next year. In fact, I'm going to take my whole family this time. I'm going to stay in italy for about six weeks and catch up with uh friends in the community over that time and then hopefully also present at the conference in Vienna the European Power Platform Conference about four weeks later. So it's going to be a fun year in 2025. I want to discuss something got to do with the work that I'm doing at the moment, which is why I believe Microsoft is uniquely advantaged in the enterprise when it comes to AI, and I think that a lot of people are perhaps missing the actual impact of what this means.

Mark Smith: 3:24
Now, when I started my career, we often would use a term this is back in the XRM days of build versus buy. When a company is faced with a decision around building an application or a solution, they could either go out and buy an off-the-shelf solution what we'd call a COTS commercially off-the-shelf solution, what we call a COTS commercially off-the-shelf solution or they could build their own. Now, the advantages and disadvantages that we used to talk about back then was that if it was commercially off-the-shelf, it's probably at a cheaper price point because it's sold en masse to a lot of organizations, but the drawback of it was it's probably not 100% fit with your organizational needs. And then we would have another story that we would tell, which is you could go and build it from scratch, right, Rather than the buy off the shelf, you could build it from scratch, you know. You could choose your programming language of choice, assemble a team and you could build it. The problem is you're then responsible for managing that code over time the technical debt that all would incur, the time it takes to build it. It'll be highly tailored to your organization. But there would be an associated cost right of engineering and building that solution out and then maintaining that code over time. And what we would say is well, you could have the best of both worlds. You could get a commercially off the shelf solution with Dynamics 365, and you can tailor that via customization and extend it, um, with pro code as necessary, and so you kind of get this ability to create a hundred percent fit, but with a solution that microsoft would continue to upgrade. So you'd take all the benefits.

Mark Smith: 5:00
And 21 years ago, me as a customer that's what sold me on what was called mscrm back then is that all the tailoring I did to meet my organization's requirements? I was working for a company called eagle technologies back then. I could ford grade them when microsoft updated the product, and that's you know. I fell in love with the tech. Um, from that point, why am I telling you the story? Because, if we bring it to today, I think there's a concept of the one Microsoft story. And why is it compelling when you look at an ecosystem from the one Microsoft story and unfortunately this is a mantra that Microsoft has internally, but I don't think there's any teeth in it and what I mean by that is because employees' compensation models are not tied to the one Microsoft story. Sellers sell what their targets are, what their quotas are, what they need to retire, and so therefore, often the one Microsoft story doesn't come to bear inside a customer's account because they have to sell whatever they are a seller of.

Mark Smith: 6:09
If you're a seller of Copilot, you're only going to sell Copilot. You're probably not necessarily going to think about hang on, microsoft. Fabric should be a key part of this mix. If I don't have a requirement target to retire Fabric, or if I don't have a requirement target to sell the Power Platform, am I going to focus Fabric? Or if I don't have a requirement target to sell the power platform, am I going to focus on that? If I'm an Azure seller and I'm going to sell Azure AI Studio or the Azure AI Cognitive Services. Am I going to be too focused on Copilot and M365,? Right, and I think that because there's not this one Microsoft story that is told end-to-end with customers.

Mark Smith: 6:47
One customers get confused. You know, I could have three different sellers from Microsoft come to me. One might be selling Copilot for M365. One might be selling, hey, all the benefits of the Power Platform and Copilot Studio. And the third might be coming to me and going, hey, listen, we can build custom AI solutions on Azure and allow you to infinitely scale. And the customer, they're just trying to make sense of this AI world that we're currently in, let alone which product?

Mark Smith: 7:17
And here's one of the other things that I noticed with Microsoft they are very product-centric in their education. So everything on Microsoft Learn, as you can imagine, is very much how to use products. When you talk to sellers, they're very product-driven and the thing is there's a gap I am seeing in business where, before the product story, they need to just get their strategy story for AI sorted out. They need to understand why do we want to have AI as part of our organizational mix? And I feel that piece is not emphasized enough. Those business discussions, those thought leadership discussions, those discussions around how technology can be an enabler now in an AI world.

Mark Smith: 8:08
And I think part of that discussion needs to be, when I go back to the one Microsoft story again is that Microsoft has the ability to sell you an off-the-shelf solution with Copilot, a solution that is highly configurable, allowing you to bring in other data sets, connectors, etc. And even tailor that off-the-shelf solution using Copilot Studio in a relatively low-code format. And then you have the ability to extend out infinitely with what we can do in Azure and all the AI services offered there. You introduce, then, microsoft Fabric into that mix and get your data estate sorted. You have a compelling one Microsoft story and why I think Microsoft has this massive advantage because of M365. I think something like 82% and I don't even know where I got that stat, so don't quote me, it needs to be validated of businesses in the world are on M365.

Mark Smith: 9:04
Of businesses in the world are on M365. And if you look at the core day-to-day data of those organizations, particularly the unstructured data, they're most likely, or quite commonly, sitting in SharePoint and in their email systems, right, which is all on M365. And then you add the other plethora of apps that sit in the M365 pool, whether it's Word, Excel, powerpoint, power BI, etc. You've all of a sudden got this data estate that's sitting on Microsoft right. That Microsoft, of course, making their tools. They have roadmaps to light up more and more of those data sets. Now, of course, organizations have then other data, right, they have data sitting in SAP and Oracle Systems and ServiceNow and Pega, just to name a few. They've got Blue Prism and you know, uipath and other automation tools running across their estate.

Mark Smith: 10:00
And what I've noticed Salesforce included no-transcript to the day-to-day users of AI in that organization. Because if I, as a user, want to engage with HR and the HR AI agent do, I have to go and find where can I get access to that AI tool. And then SAP comes out with their AI tool and ServiceNow comes out with their AI tool and what you have is every single piece of software out there Salesforce, pega, everyone's got their AI tool and all I see that doing is creating a massive degree of confusion for the day-to-day worker in the organization, because now they have to know where the starter resides and which AI tool to use to do this, to do that, to do something else. And six weeks ago, satya spoke about the wave two of Copilot and he opened with this it's going to be the UI for AI, and I think that statement, that such a simple statement, was lost in the noise, because UI for AI is something that's going to be critical, going forward to remove confusion and to drive adoption of AI, and that is I.

Mark Smith: 11:58
As a user inside my company shouldn't have to know which AI system to use or which AI tool to use from all the different products that we have running inside our organization. Most companies will have over 200 to 300 different software applications inside their organization. Does that mean we now have to have 200 to 300 different AI applications that interface to those data sets? No, I think, with BizChat and Copilot pages and what Microsoft is doing with UI for AI will allow you to have a single pane of glass for your AI experience and all those data sets inside the organization, whether it be SAP, servicenow or you name it right whatever HR system they have in place workday, et cetera would be able to be surfaced within this UI for AI of co-pilot, and I don't need to know where that data resides in the organization. As long as the data estate and infrastructure has been lit up, I can use a single pane of glass UI experience rather than 300 new AI applications based on every vendor rolling out their unique experience. And this is why I think the one Microsoft story is such a compelling story, because you allow the ingestion of all your organization data, the entire data estate and, as long as you've done the work around security and RBAC, you know role-based access is applied and you know the data access is meeting the governance and the compliance rules, you create a very powerful, compelling experience for large and enterprise businesses across the globe, and so I think that's the story that Microsoft needs to really double down on the UI for AI a single pane of glass. Yes, we've got multiple different ways of doing UI, whether it's even you know, sorry doing AI, whether it's AI for sales, whether, you know, co-pilot for sales or co-pilot for finance or co-pilot for service. All those are great, and if I'm only in that application, yes, I can kind of if that's my only job, but we know employees have got HR requests to make. I think that we're going to move to a world where this co-pilot UI for AI is going to become extremely prevalent and allow the cut through of noise. Now I find it interesting that OpenAI came out with Canvas as part of their model just probably three or four weeks later, and when you listen to the interviews done by them, they've been working on this for 18 months. So my little conspiracy theorist type, you know. Did Microsoft do a deal to get the jump with Copilot Pages and BizChat out to market before OpenAI did with their Canvas? Openai's Canvas experience or ChatGPT Canvas? I don't know, but it's interesting that this is a way of thinking that's moving forward.

Mark Smith: 15:02
The co-collaboration on prompting within the organization. I'm seeing some great stuff bubble up around running prompt-a-thons inside organization, running multiple workshops and innovation challenges and hack-a-thons and things like that. So the business can really start once again from those subject matter, experts come up with ideas, and this is what I really think needs to change from a strategic direction across organizations Rather than going. What are the workloads that we can use AI across? I think they need to go. How do we increase the digital literacy, the AI literacy of every employee in our organization by 30%? Because if we rise everybody's understanding not only understanding but practical application of AI therein is going to lie all the use cases that you could ever want to backlog a mile, 100 miles long. Right, because if you enable everybody to change their mindset and think with how can AI enable what I do? There will be an unlimited supply of ideas that are coming from people that affect their job every day.

Mark Smith: 16:18
Now, one of the things that has plagued me for the last two years is am I doing enough with AI not to get left behind Because I see so many online? I see so many new AI products. Every minute there's a new AI tool popping up and that's great, that's innovative and it's good to see people developing new tools and ideas. But my concern for me is like I think AI is more than just tools, right, it's like how do I become good at what I do? How do I scale what I do by using AI myself and even my observation of many consultants. They talk about AI in a theory, they do it as a demo, but they're not intricately using it as everyday part of their job. I would hardly pass an hour without prompting multiple times nowadays in an hour and interfacing with some form of AI. My entire podcast has been lit up with AI for the last 18 months, in that the transcripts are now all done via AI. I didn't offer transcripts before because it was too cost prohibitive. Now, all the transcripts are done via AI and are becoming more accurate all the time. I use it in every part of my workflow that I do every single day.

Mark Smith: 17:38
Even coming up with the outline for what I was going to discuss today, I ran through my thinking first and got a critique. I use AI as a sparring partner with me to find out the gaps in my thinking. What am I not thinking about? What should I think about more? And I feel there's so many people in the consulting space that know AI in theory but not in practice, and I think we need to build it into every part of our lives. So, for me, I don't think about using electricity. It's intrinsic to my life. Right, everything that brings this message to you right now is done by electricity, or the need of electricity to make it possible, and I think AI needs to be the same thing in our lives. And so I've been doing a lot of study in the past weeks and it's got to do with a lot of the work that I'm doing for other companies around the area of agents and agentification.

Mark Smith: 18:30
And what does it mean to be agentic? Look up that term. Look up what agentic means, because this is the world that we're moving into in 2025, and so I'm like, as we hear about agents and we particularly hear from corporations talking about agents often the marketing message is way ahead of the reality, and I feel we're in that little bubble at the moment of the reality of agents is not where the marketing message is telling it is at right, and so when I think of agents, I kind of look at it from three lenses. I think it from rules-based agents. So if this, then that type of experience, an agent that has a body of knowledge and it's very much configured by a person or a group of people to do, you know, an if-then-that type statement. Then I see semi-autonomous agents right where the human's still in the loop type statement. Then I see semi-autonomous agents, right where the human's still in the loop right, we still might need to do final approvals.

Mark Smith: 19:33
And then there's this future idea where, which, where this term agentic comes into play, where we look at a world that is fully autonomous. The agent not only can follow rule-based patterns, but as new knowledge is presented, as it learns from interactions it has, it becomes more powerful, it has more understanding of how to make decisions, and I mean more powerful in a very positive perspective, because it allows for greater efficiency. And so, between that delta of agents and where the marketing is telling us agency, I'm like what can I do today to prepare myself for a very agent-centric world? And then imagine an agent that is fully trained, fully autonomous and runs brilliantly. And OpenAI came out with a new API, which is this idea of swarming, which I could say okay, now scale that across 100 instances of that agent, 1,000 instances of that agent, a million instances of that agent.

Mark Smith: 20:41
All of a sudden, the amount of scale that I can produce me as an individual becomes mind-blowing. So when I look at a 40-hour work week, which I've grown up with from the Industrial Revolution, I see my output could potentially be 500 hours a week with agents assisting me, right. So it really transforms what is possible. And so what this led me to do is that I've started a process in that every day, I try to write one SOP, one standard operating procedure for my life, and when I say my life, my life includes work and private life, right? And so I'm documenting I'm just using Word, but I'm documenting every detailed step, an SOP of how I do something. And I'm. I'm looking at every possible area of my life that I do something as a repeat pattern, over and over and over again, and I'm writing a document SOP and I learned about SOPs many years ago, before I got into tech.

Mark Smith: 21:45
Over 30 years ago, I was in the medical industry and I met this gentleman Ray Avery his name was. In fact, I went to North Africa with him and with the Fred Hollows Foundation and set up a micro laboratory in a country called Eritrea, which is between the Sudan and Ethiopia, after they had been at war with Ethiopia. This is Eritrea for 30 years. Amazing experience. It's written on my consciousness forever.

Mark Smith: 22:13
But one of the things that he had was this concept of Kaizen, which comes from Japanese culture, and ultimately, we've heard Toyota espouse this quite a bit, and the whole idea of standard operating procedures is that in Japanese culture, as I understand it, if there is a mistake made, an accident happens in business, they don't ask who did it, whose fault is it, which is very much a Western culture that we see. What they do is they ask this how did the process fail? So, rather than blame people, they blame process. And so SOP standard operating procedures means we need to improve our SOP over time to make it more reflective of what we're doing and the outcomes that we're driving. And so my model of developing these SOPs I am then taking the SOP and I'm giving it to another person that I employ and they go and they follow the SOP and they follow it consistently over time and every Friday they have to report back to me on any changes that they've made to that SOP, any improvements that they feel need to be changed over time. And then every three months, I'm setting myself a routine to use the latest LLMs and I hand my SOP back into them with a query of where can I improve this? How can I make this more efficient? How can I make this more cost effective, more time effective? And so, at least four times a year I'm having an LLM reason over my thinking, but the rest of the time, for three months at a time, I'm having a real person carry out those SOPs, allowing me to scale, Because I know that within 12 to 24 months, I'm going to be able to hand that SOP to my agent, my personal agent, which will then be able to fully handle it autonomously. But I don't want to be playing catch up when agents are ready. I want to have got my ducks in a row and created that scale. So my goal now for me is can I do 500 hours a week productivity, and in that I'm reducing my actual working hours, my physical work hours, to only four days a week, and my goal over the next four years is to drop one day a week of actual the stuff that makes me money type work off, while increasing the revenue I'm producing by using technology and specifically AI. So, surprisingly, I've talked for a bit of time. I've got one last thing I want to cover.

Mark Smith: 24:49
This week there was a white paper released by Anthropic CEO Dario and it was called Machines of Loving Grace how AI Could Transfer the World for Better. And what I loved about this is that his lens is so aligned with the way I think of AI. I don't think of a dystopian world of AI. I think of a very positive world and a very utopian world that we're going to move into a world of AI. I think a very positive world and a very utopian world that we're going to move into, a world of abundance and unbelievable opportunity. But I also balance with that.

Mark Smith: 25:24
At the Power Platform Conference, the lock note on the final day a speaker talked about, the third largest GDP economy in the world is that of cybercrime, and so I'm not saying that people will use AI for nefarious means. They absolutely will. And I think of all those companies that haven't updated their software, haven't modernized their software. The risk profile that we now stand in to anything that's internet facing, facing that's hooking back into an organization's data, is going to be extremely high when you have got ai, that is going to go out and find those holes and potentially be exploited. But, going back to this white paper how ai could transform the world for the better, I highly recommend go google it, take a look at it.

Mark Smith: 26:11
But he has kind of five key areas that he looks at. One is biology and health, and so I read the entire white paper. I then ingested it into uh, an ai, and what I did is that said, let's start with first of all. The first thing he talks about is biology and health, and my prompting is listen, don't give me all the, the medical data that has been influenced by pharmaceutical revenues. As to what is medical advice?

Mark Smith: 26:42
Things like bmi, right, bmi was never meant to be body mass index, a individual measurement of one's overweightness. It was meant to measure a population's overweightness. It was never taken into context of a diverse population mix, but it still held on to most health apps I look at. You know I use MyFitnessPal and even now, bmi is the big measurement of health. Other tools that I use my scales, my digital scales, et cetera it's all BMI, bmi. Yet that's such a false measurement of health. You know that we should have today.

Mark Smith: 27:19
Anyhow, I took this biology and health section and I created this you know engagement with AI and and looked at the in context of the way I look at health and that you know, could we live to 150 years of age If we could eliminate all disease? If we could? Uh, you know, could we live to 150 years of age If we could eliminate all disease, if we could? You know, and would we want to right? It's interesting quandary.

Mark Smith: 27:45
I do think we are going to be faced in the very near future with a concept of a mortality and that is the ability to actually, from a medical perspective, to live much longer than our current lifespan enables in a modern world. And so I looked at this biology and health from how does it apply to me? And I gave the supplements that I take. I've been intermittent fasting now for over six months. I do long-term fasts twice a year. I've just finished a seven-day water only fast and so with that, with the supplements and stuff and what I'm doing, my body composition and health and adding in, you know, routines around. You know, because of my age, you know, mobility becomes, can become an issue in older age. So things like working in yoga into my pattern, working into high intensity training, interval interval training, as well as resistance training into my lifestyle, so that for me, if I keep my body in health optimal state, I'm probably going to be able to see a lot of this amazing future.

Mark Smith: 28:45
And then he moves into an area number two of neuroscience and the mind and the ability of AI and what that future could look like. In working with mental health and this whole area of neuroscience. It's becoming more prevalent and as the brain is understood more with AI, the capabilities of humans and the ability for deep thought and intellectual development really becomes unlimited. You know, when we look at the inventions of the last century or so, what happens when neuroscience advances to allow those innovations to happen at scale. Then he talks about economic diversity and poverty and what are we going to do around making sure that this wealth, economic possibility, is evenly distributed, you know, those that in extreme poverty don't move into, you know, get left behind, so to speak.

Mark Smith: 29:40
How do we think of of it in that context? And I think it means that things and I you know, like the free market and the concept of the capitalist type market, I think it needs to be balanced with fairness for everybody that might not be at the same stage of economic power that we have. I think it changes and that leads into the fourth area, which is peace and governance. And what does that mean when you create a governing model that's not influenced by large corporations that can buy off the politicians that we see all over the world, not just in the US, but all over the world, and what that means for peace. I talk about peace in a world that is full of more war, conflict than I've ever seen in my lifetime. And I wonder how can things change? I know it won't change necessarily rapidly, but what does that future state look like? And then, finally, work and meeting. What happens when we don't have to work as much but can make so much more? So I I think we go into a very unique world in the future, uh, and so I recommend take a look at this. Let it be a thought challenge to you.

Mark Smith: 30:47
And with that I've just completed my first 30 odd minute solo spiel. Oh, I'm keen to get your feedback. Hit me up on LinkedIn if you find it valuable. I know the others are going to be surprised when I drop. Hey, here it is that this podcast happened with them out here. I find I enjoy it so much more when they're here. I look forward to the next episode with them all back in the house. And what are your thoughts? Thanks for tuning into the Ecosystem Show. We hope you found today's discussion insightful and thought-provoking and maybe you had a laugh or two. Remember your feedback and challenges help us all grow, so don't hesitate to share your perspective. Stay connected with us for more innovative ideas and strategies to enhance your software estate. Until next time, keep pushing the boundaries and creating value. See you on the next episode.