Exploring AI Sales and Governance and Copilot Studio Glue with Bert Wijns

Mark Smith: 0:01
Welcome to the Co-Pilot Show where I interview Microsoft staff innovating with AI. I hope you will find this podcast educational and inspire you to do more with this great technology. Now let's get on with the show. Today's guest is from Duffel, flemish region in Belgium. He works at Microsoft as a Global solution strategy architect. I've had the pleasure over the last couple of years of being associated with various things he has done inside Microsoft. You can find links to his bio, social media etc. In the show notes for this episode. Welcome to the show, bert.

Bert Wijns: 0:39
Thanks for having me, mark, and I like the way you pronounce Duffel. Actually, not all people attempt that.

Mark Smith: 0:45
Yeah, so I've been to Belgium and like the old city, I suppose, where the train came in from London, Whereabouts is the region you're in relative to that train link into London?

Bert Wijns: 0:58
Yeah, I would say I live quite central in Belgium, in a small village close to Mechelen which is back in the middle of Antwerp and Brussels. I mean back from my constituency days. I had to go everywhere. So when we bought a house we just picked a very central place so I could go all over Belgium and surrounding countries.

Mark Smith: 1:18
Nice, nice. Tell me a bit about food, family and fun. What do you do when you're not doing your Microsoft day job?

Bert Wijns: 1:24
Yeah, time is always the thing, right, keep myself quite busy. So we got a three and a half year old son after very active boy keeps me busy as well and I it's gone to an age where it's actually fun, right, you can play soccer with him, see him cycling, you can teach him stuff and he picks it up. So he keeps me active as well. I would say fun. I'm quite into sports. There's a game of sports called padel here in Belgium. It just gained a lot of popularity. Okay, okay, it's popular over there. It comes out of Southern Europe. It's like a mix of squash and tennis.

Mark Smith: 2:01
Wow.

Bert Wijns: 2:02
Yeah.

Mark Smith: 2:03
Does it have a specialist court of squash and tennis? Wow yeah, does it have a specialist court? You know how squash is a specialist court. Is it like played on a tennis court or a purpose-built court?

Bert Wijns: 2:10
I would say it's purpose-built and it's like a squash course quite similar. But I do play that competitively as well. When I get into something, I go and take lessons. I improve, I do it three, four, five times a week. Nice I improve, I do it three, four, five times a week, nice, I'm quite competitive when it comes to that as well. Food I love Italian food. I mean I love Italy as a region. Been there on holiday many, many times. The pasta, the pizza, the wine, yeah, I just love it there in the atmosphere, and so any Italian food really hits the spot.

Mark Smith: 2:41
Nice. My daughter's three and a half years old, so I know the energy that you're talking about there. And interesting about Italy, because it's my wife and my favorite country in the world, just hands down. We love it, we love the culture, we love the food. I know you were just at European Summit recently and for next year's is in Vienna. So my wife and I and our two young'uns are going to come up and live in Italy for a month Nice and I will then do the Dynamics Minds conference at either start or end of that and then the European conference at the end of that. But we'll base ourselves in probably northern Italy as a good central. We love Lake Como and all that kind of region and through to venice and yeah, yeah, the whole region's amazing and we haven't spent any time in south italy so we need to go down, yeah, it's hard to get past tuscany.

Bert Wijns: 3:34
As soon as you get into tuscany, you fall in love. Yeah, and it's hard to go down south that's it.

Mark Smith: 3:39
That's it. That's it. So so true. Tell me, how did you end up working for Microsoft?

Bert Wijns: 3:46
A little bit of background. I mean, I've been working in the business application space for 17 years now.

Bert Wijns: 3:54
I think I spent quite a bit of time at SI Popsner, initially EDS, then it was acquired by HP and then I think it got renamed three or four times by HP and then I think it got renamed three or four times and joined the. I started out in NET and then very soon ran into the Dynamics product and joined the European practice, started traveling a lot, working on large engagements and at some point we ran into Microsoft Consulting Services in a couple of years and that's been my entry ticket back in 2014.

Bert Wijns: 4:26
I'm actually 10 years at Microsoft this year in November, so I joined that as an architect. It switched a bit, I guess, from doing implementation work to advising customers and partners on implementation and also the volume of customers and the challenges you get in front of you really increased, like in my seven years at HPE probably saw 14, 15, maybe 20 customers max, I think. In the five years which I was part of in consulting services in Microsoft there's been 50 customers and a really large enterprise and complex implementation. So it's a really steep learning curve as well. At some point then I started getting an interest into packaging up things, strategy go to market so I got more into corporals, then managing the power platform workloads for consulting services and then that took me to new people, new places and now I'm more working on the AI side, where we're looking at co-pilot products and a strategy and trying to get our sales team and Microsoft colleagues to collaborate the many co-pilot products we have out there as well.

Mark Smith: 5:38
This is interesting. So you're not in consulting. Now, right, you're not in consulting services. Yeah, you're in the main Microsoft.

Bert Wijns: 5:44
Yeah.

Mark Smith: 5:45
I call it the main Microsoft because you know a lot of people don't know that. There's the consulting arm of Microsoft, which is an own business in its own right. Then there's Microsoft, the mothership there, where you know all leadership and stuff kind of flows from. And then there's the subsidiary model. Is it still 14 subsidiaries around the world?

Bert Wijns: 6:04
I think it is yeah, I'm not sure how many it is exactly, because it changes quite a bit yeah in there, but I'm more at the product side. I would say not engineering exactly, but more around sales strategy right now yeah, so interesting time.

Mark Smith: 6:21
I yesterday I just wrote a 3 000 word blog post on so not a small one on practically using ai as somebody that's come from a power platform or business applications world. And how do I prepare myself, skill myself for the future world? And at at the moment, the internet is out, every person out there there's an element of snake oil being sold under the heading of AI, right, and there's so much theory like how do you use AI? And they go well, you need to automate. And I'm like, wow, you just threw out a phrase that's as big as the Titanic automation. What does that mean? Like that's tools, that's how you think about process. Like it's that's not definitive, that's not an action step to take. And then, when we look at it, selling, how do you sell ai? And and here's the other kind of thing and it'll be interesting to see your views on this is that to sell ai, and I'm not talking about selling AI software how do you have the conversation with a customer when they've come to you and said Microsoft, you're the AI company, we've already got M365 with you, we're already consuming Azure. We've already got a bunch of services you're now considered based on the most recent chart I've seen, pretty much got around 60% market share globally if you combine the OpenAI $10 million funding piece and there with the IP et cetera, 60 but and so microsoft is now in this market leading position.

Mark Smith: 8:00
For years I've always noticed microsoft has been in a market following position, and that's not a bad thing, and what I mean by that is that you had playstation we bought out xbox boom, we dominated. You had lotus 123. We bought out excel boom, we dominated. You had before windows nt, there was a netware novel netware I could hardly even remember of that product. Now right, and then boom, microsoft dominated. Own before Windows NT, there was a NetWare Novell NetWare. I can hardly even remember that product now right, and then boom, microsoft dominated it. He had AWS. We had Azure right, we win.

Mark Smith: 8:33
For the first time Microsoft has been in the market leading position, not the market following position. Ultimately, owning a category in those following positions are very powerful and I feel that there's a kind of crisis of identity with a lot of sellers in Microsoft around. We're now the thought leaders, not the thought followers. We're not challenging a brand, we're not challenging another product's position and saying, well, this is where we are better, we're leading in it, and so it's interesting that you're involved in the sales space of you know what? Has Microsoft got 70 different co-pilot applications now, or something as a number I heard recently? How do you get sellers to have those conversations with customers that ultimately will drive massive value for the customer and, of course, allow various SKUs for Microsoft to be sold?

Bert Wijns: 9:19
Yeah, I mean loaded question. First of all, I would say it's not easy. I always say we are a big ship and making a ship turn a big ship turn like that it takes a lot of time and structure. We have many co-pilot products out there, but I think we're also realizing that that is the case and we are taking action to merge some of that thinking. I think all customers are looking into AI. If they aren't, probably something wrong. I would say many of them have even jumped into proof of concepts, jumped straight away into building stuff without really thinking about what's the long-term vision. And you can almost replicate this thinking to Power Platform, where we had citizen development and everyone just jumped in and started building apps without thinking like what's going to be the result in one year from now.

Bert Wijns: 10:13
I'm having flashbacks Like I come out of that Biz Apps world with Dynamics in Parliament. I'm having constant flashbacks around. Maybe we need to set up a landing zone and a center of excellence and governance. Have I heard this before? So yeah, 100% within Microsoft. I think we have a journey to go around.

Bert Wijns: 10:33
How do we sell this, how do we sell AI and I wouldn't say it's always about selling but how do we bring an end-to-end AI story where we have all of these individual products and they all have their purpose, they all have their value, but where you see more and more is that there's overlap between some of these products and Microsoft is the only AI vendor out there which can provide AI products which you can buy with a license, like M2C5 Co-Pilot, which you can extend by adding extra data sources or adding plugins, like with Co-Pilot Studio. But then, if you want and you want to have access and configure all of the details, like choose your own model, fine tune it into the deals you can also build it with Pro Code. So we have that market leader position where we can offer all of that. Having said that, we always in Microsoft have our three big solution areas Azure, modern work and business applications and in each of these we have co-pilot and AI products, and what naturally happens is that people sell what they know. It's like you use what you know is that people sell what they know. It's like you use what you know and you go and push that to your customers.

Bert Wijns: 11:43
So what tends to happen is that a customer gets approached yeah, multiple sellers from Microsoft on AI products and they get a similar pitch and that is one thing we realized and we are trying to change as well maybe more bringing an AI end-to-end strategy story to a customer and really looking at okay, dear Mr Customer, you have a thousand employees spread around eight or nine business functions. We believe these three are probably most relevant for you to start doing things with AI and these are the top AI use cases we see in your industry in these business functions, and this is the return you're going to get, and it's not only going to be productivity, which is the easy return, right. I mean, you use AI, your 100 employees save two hours a day so they can either go and do other work or spend more time drinking coffee.

Mark Smith: 12:39
Yeah.

Bert Wijns: 12:39
But it's not going to impact the bottom line directly. So you also need to look at things like, let's say, I use cases in HR, in sales, where you also impact different business KPIs like new revenue, customer lifetime value, and you really need to bring more the AI story and journey from that perspective so a customer can translate it into a business case almost with a real business impact, versus selling products with features, individual products with features, and I think that is a big shift we'll see this year where we are pushing really to try to move away from that individual AI products sale. Having said that, it will take time. It will take time before we get there. I mean, customers always complain like, hey, I got this meeting with Microsoft and eight people show up and I just didn't want.

Mark Smith: 14:00
Yeah, at the end of the day as well, when it comes to product selling and when people, if we look at their base behavior, people are coin operated to a massive degree and what retires target, et cetera, gets the focus. And I don't necessarily get compensated for the one Microsoft story, even though I think Microsoft's biggest feature in the world and benefit to most organizations is the breadth of the software stack that they have that can talk together and having solutions for those areas and the value that can bring to an organization. I saw a report last year that said it was 6 000 ceos of companies who were interviewed around the world and about what was their number one priority for ai and out of it they saw was interesting. Over the course of the year the report was released twice. At the start of the year they saw the biggest use case was marketing. By the end of the year it had swung totally to. The most prominent use case was customer service. That's where they could move the dial the most and of course Microsoft has come out with, you know, contact center, ai and the whole AI and the customer service space. And when I look at the use cases where it's an audio conversation happening and being able to have that AI start helping the support agent with the right things based on what has been requested.

Mark Smith: 15:24
Understanding the nuance of language Did somebody just switch and start speaking or use a word from another language because English is their second language?

Mark Smith: 15:34
All those type of amazing things that Microsoft hadn't developed long before the Gen AI movement crashed on the shore, and that support scenario seems to be a big use case.

Mark Smith: 15:50
The other one that I always feel is a big use case is understanding the organizational knowledge, which I think you know. M365 is kind of set up to really do that, but organizations have, for, depending on their length of being in business, have collected data, and we, you know, I remember back BizApps days of talking about the four mega trends that were going to impact the world, and one was mobility, but the big one was big, what we call big data. I assume you remember that from back in your selling days of it and I feel that organizations got really good at collecting data, but not necessarily been able to use it to really gain insights, to drive value inside, not only for their customers but just for themselves, right, and so I see that as probably another big use case. What else, are you seeing as kind of those big things that can really impact a business by applying practically AI to the org?

Bert Wijns: 16:49
It's going so fast. Like, I mean, you touched on customer service. I remember doing a Dynamics implementation in 2010, I think, or 2011, for the Aqua Hotel group back then, and I remember implementing this, doing trainings, and I'm like, well, there's logic, we build actually. Well, you, you got an agent. Then all it really does is follow this decision tree and it then sends an email yeah and I'm like how wouldn't we be able to automate this anyway?

Bert Wijns: 17:17
and I think, yeah, I always said like, yeah, maybe one day. But if you look at the last two years actually one and a half years, I have to say if you see this technology which is is coming out and the speed it's coming, I think a lot of this information or knowledge worker jobs are really, yeah, I wouldn't say going to be replaced, but these people are going to be moved to different tasks. They're not going to spend time on the things, and contact center and AI is one area which I personally put a lot of focus on in the last year. But I think, if you look at any business function in a way like let's take sales, we have a lot of sellers and I work with a lot of sellers part of my role and I kept joking to them like, hey, I'm not sure you're going to have a job in a couple of years because this co-piloting, I mean you have a meeting with a customer. It picks up on the conversation, it automatically updates your CRM system. It sees if the customer says something and there's a hint in the history that it's going to give you a hint like maybe talk about this, maybe position this product.

Bert Wijns: 18:22
Another one is HR. I mean recruiters, recruiter, they're asking you certain questions, but they can probably scan your LinkedIn, your CV, to have certain knowledge about you, which a recruiter would take days to prepare for something you would have an interview. It would spot that you say certain things which don't align with the role. Any of these business functions has a massive potential for AI and it's just increasing and the speed is picking up, like the GPT-4.0 release, where now we have vision and speech in better as well, we have real-time. I think we're going to see it only accelerate and I think that's a challenge on its own as well. How do we keep up with this pace of innovation? And maybe in five years from now?

Mark Smith: 19:09
now the world will look totally different because of this ai technology what's top of mind for you right now and what you're doing in your role around copilot studio or azure ai studio, and I see those as a kind of three landscapes which address the three big parts of microsoft that you talk about and being able to go from a co-pilot being a totally sass solution and operating in line with your m365, what you're used to all the suite of products there co-pilot studio kind of seems to fit squarely in the space of being able to edit, introduce rag to the process.

Mark Smith: 19:48
You know, just like in where we had with Power Automate, we had Power Automate and then you had Logic Apps If you want to do some real heavy lifting, you can always go to Logic Apps and do that kind of heavy lifting. Is this the same thing with Copilot Studio and Azure AI Studio? That Copilot Studio is that, yeah, a bunch of let's say let's use the term business analysts that know what they're doing, can figure up, tailor it to the organizational needs, drive a lot of value, allow people to start practically using it in their roles. But if you wanted to go further, like we talked about you know working with which language model you want to use that type of thing, you go into the Azure AI Studio as a landscape, go to market with those three lenses. How do you think about that really? Because it's not necessary. I don't want you to talk about what you're doing or what you've done or what you're planning to do, but really, how are you kind of working that through?

Bert Wijns: 20:47
yeah, I mentioned flashbacks earlier, right, so I'm going to mention it again. Back in the days when we started with dynamics, people were building their own crm system with NET, java or whatever. So we're having this build versus buy.

Mark Smith: 21:03
A hundred percent. I used it this week, that concept.

Bert Wijns: 21:07
So in a way I'm having flashbacks, Like this is a similar thing. So many customers jump in straight away into building custom co-pilot. I mean mean, it's just like dynamics in the days with the first couple of releases. Something might have not been as mature or as many functionalities, so they jump into the custom build and start there. But then six months down the road they realize like hey, I might've just built M35 co-pilot custom and now I'm going to have to maintain this. Who am I going to put this into production? And maybe someone's going to ask something which will get us into the newspaper and hallucinate on us.

Bert Wijns: 21:46
So I think what we've seen is that customers need to take a step back, Like every customer and every business function is looking at AI and taking steps in there like don't prove concepts, build things, Maybe, have a couple of low-risk production use cases.

Bert Wijns: 22:02
They need to take a step back and think about what is our strategy. And I think overall in IT, many customers say overall they prefer to buy something versus build something, because it's a proven practice that you have less maintenance costs, that you get all the things out of the box and in the end you pay a license fee but it's going to be cheaper for you long term. Having said that, it could be that you have certain requirements which require you to go from the no-code solution, which is M65, code pilot, which looks at graph data and some knowledge sources, 365 co-pilot, which looks at graph data and some knowledge sources, into the low-code solution, where you can extend that standard co-pilot with adding different data sources, forcing it to look at some other things prior to going to the graph and back and then yeah maybe if you have an absolute need which is very specific, then you still need to go custom or a hybrid model where you pick and choose from both.

Bert Wijns: 22:59
But I think the big message we want to land with customers is definitely start by looking at the product you can buy off the shelf and then we allow you the choice. And, like I said earlier in the call, I think as a company we give you those three options and that is very unique. And also customers' requirements mature right. They can start with an idea where you can say, oh, there's an easy match to M65 Copilot. Oh, we need more data, okay, maybe let's add Copilot Studio. But then they might get a certain requirement which even might require to build something custom with Azure AI Studio as well. But we allow for that journey. But a lot of customers have jumped straight away into a choice without thinking about it, without having an end-to-end AI strategy, making decisions, governance. And when we get in front of customers, I always advise them like, take a step back. What do you really want to get out of AI? Where do you want to get the benefits? And then think about yeah, what patterns do you want to follow?

Mark Smith: 24:00
Yeah, I want to do another podcast with you because I feel like there's so much more to explore and people are struggling with what to focus on in this space. But one of the things I know that you've done you've built some apps on the side yourself over your career.  Tell us, tell us a bit about what the product is that you already have built and what problem did it solve. Tell us, tell us a bit about it. 

Bert Wijns: 24:23
yeah. So first of all, I'll explain a bit of context about how this is possible. When you work for microsoft, I think, yeah, time is the issue for everyone, right? You only have so many hours in the day and the other house you sleep what you eat. So I'm one of these people who wants to be very busy and keeps himself busy. And just when we had our first kid actually about three and a half years ago me and another colleague from Microsoft Belgium launched the God Power Accelerate and the idea behind it is yeah, we saw a massive growth of power platform and we also saw a lot of customers still had a lot of legacy IT application in Lotus Notes, microsoft Access, excel, sql, infopod, whatever and we figured like most likely customers are going to want to move away from that legacy application and get the benefits of modern IT, like AI features actually. But it's an expensive thing migrating legacy software to new platforms. So we figured couldn't we create something which would automate a bunch of these steps as part of that, like scanning the source, allowing you to do data modeling, auto-creating the data model, migrating the data, generating screens, forms, multi-driven apps, canvas apps, automating as much as the boring generating screens, forms, multi-driven apps, canvas apps, automating as much as the boring repetitive developer work in migrations, and that was the idea.

Bert Wijns: 25:46
And in Microsoft we have a concept called moonlighting, where you can start the company next to your Microsoft role. Of course, there's a bunch of rules about it. You cannot use your Microsoft plan Like, hey, I'm working on Microsoft, buy my software solution. There's a bunch of rules around it and it can compete with Microsoft. I'm not going to write something which migrates Power Platform to Salesforce apps or whatever, but it's actually an amazing learning experience.

Bert Wijns: 26:15
I mean, I'm almost 10 years at Microsoft and we've been running this company now for almost 40 years. I would probably state that I learned more doing power accelerate and having many hats, from customer service to the ceo strategy, marketing, anything in those four years, then in those 10 years, and microsoft, because there's just so much more happening. I mean you create a product you develop for three months. No one wants it. Now you throw it away and you start over with customers and partners and you have to go and find your first customer you don't have the Microsoft brand and find a crazy person who's willing to pay you money for what you've built. So that experience and the people you meet and and it is amazing, and I've used so much of that experience in my Microsoft role as well it's a lot of other things.

Mark Smith: 27:06
So let's just unpack this. Because you started a business, you build a product, you started a business and often when people come from IT, that's almost the easy piece the building the product, because it's tech, it's understandable. The hard piece is the marketing of it, the selling of it, and tell us a bit about what successes you've had so far with that product.

Bert Wijns: 27:29
Yeah, I would probably say we've had about 35 paying customers, which means we've probably done around 400 product pitches at least, I would say on Demos, probably more Demos, I would say. So we started small, focusing on, I think, excel. Actually, a screenshot to a PowerApp was our first use case. And funny enough, yeah, I have to tell the story Two and a half years later, satya, at the keynote MS Build, presented the exact same use case of taking a screenshot, moving it from Figma to a PowerApp. I would say that was a bit of a celebration moment, like hey, we beat you with two and a half years, satya, with this idea. So we started a small screenshot in Excel so we could add data.

Bert Wijns: 28:22
I think when things really go moving is when we realized like, yeah, we have to go and look at specific niches where people's pain is big enough that they're willing to pay money to make something happen. And I think Lotus Notes was probably one where you had these customers who had these notes environments around for 10 years and they've been trying to get rid of them for 15 years. So they were still paying on-premise licenses, they still had on-premise servers which weren't compliant anymore. We're getting into security. So I think a big realization for us was that you need to find a niche on the source systems and the reason why customers would pay money both to us as a software solution, to an SI partner to do the actual execution and the manual migration work, and then to Microsoft for the Power Platform licenses. So I think that was a big realization for us and then from there we accelerated MS Access.

Bert Wijns: 29:19
I think it's a really good channel for us as well. Things like 32-bit applications, old versions, security issues, hard-coded credentials in VBA scripts InfoPath is becoming a big channel for us as well right now, with some of the SharePoint servers deprecating still some on-premise SharePoint moving away from that. So, yeah, a lot of learnings around, I would say, picking the right source system and channel to migrate from it as a pain, but then also, yeah, getting that first pain customer. I think we had like 40 interactions before someone actually agreed to meet that person to actually pay for it, and then we learned so much from that first customer as well. We should have probably done it for free, I think, just because we learned so much from that as well.

Mark Smith: 30:08
So how do you find your customers?

Bert Wijns: 30:10
Marketing mostly happens in two ways, I would say. I mean, we looked at a lot of things like Google and all of these socials. Linkedin works quite well for us and it's more of a b2b platform. The other one is google and youtube. I would say, having recorded movies hitting a bit on the big search terms like notice notes to power apps, I think a Microsoft marketplace works quite well for us as well. I think packaged offerings there for the source technologies and then just mouth-to-mouth. Really, we've built relationships with the SI partners. Now We've done some work with us. They speak to another customer and he accidentally mentions one of these things. I think at the moment we're multitasking a lot. We have a full-time role, we have this company next to our role. We're not actively posting things on LinkedIn every day. It's actually quite busy as it is and we have a couple of engagements happening. So we're just progressing on that steady growth, I would say, with the time we have available.

Mark Smith: 31:18
That's epic. What was the product name? Again, just so people want to look it up.

Bert Wijns: 31:21
Power Accelerate and I'll share some links You'll provide the link.

Mark Smith: 31:24
We'll get it in the show notes. That is epic. As always, it's been great talking to you. Thank you for coming on the show.

Bert Wijns: 31:31
No worries, Thanks for having me, Mark. I really enjoyed this and happy to do this again. I mean, like I said, this AI and co-piloting is evolving so fast and maybe in six months from now I'm telling totally different things.

Mark Smith: 31:43
Yeah, three months Okay.

Bert Wijns: 31:45
Cheers Cool. Thanks, Mark.

Mark Smith: 31:48
Hey, thanks for listening. I'm your host, Mark Smith, otherwise known as the NZ365 guy. Is there a guest you would like to see on the show from Microsoft? Please message me on LinkedIn and I'll see what I can do. Final question for you how will you create with Copilot today, Ka kite?