Jonathan:
Systems need to upgrade to the AI world with today's special guest, Ryan Pollyniak.
Now Ryan, I'm really excited to have you on here because you come from the exact opposite background of me, which is you're an insider, you used to work at Microsoft, and now you work in Microsoft consulting and you really come from that perspective of taking systems and working with teams and helping them to come into the modern age. And one of the things I've noticed is that the larger an organization is and the longer they have been in business, the harder it is for them to make change.
My last job working for a company was at a large university, and they were using a database system from the 1980s. It was so old and so unwieldy and you had to really have one person who just did data entry. There was no such thing... It wasn't connected to any other system. It wasn't connected to their email system, so if you wanted to email someone, you'd open up the file and then look up their name and then manually type in the email that you saw in the file. And it was that real process, I was like, "Why don't we just choose hole punch cards? Let's go full old school."
But now with how fast things are changing, every company's feeling this pressure, the biggest FOMO I've seen in a long time, "Oh, we have to keep up with everyone else." So from your world with a lot of these companies are coming from legacy systems, what are the biggest challenges they're facing and the biggest stresses they're feeling with all this pressure of everyone needs everything to be AI right now?
Ryan:
Yeah. Hey, Jonathan, great question. Thanks for having me on. One point of clarification, I've been working for Microsoft Partners for the last 15 years, but not directly for Microsoft. Anyway, great question. So the biggest inhibitor to change for an organization with their legacy systems is business disruption cost, right? It's we've always done things a certain way. We have grown organically possibly or grown through acquisition, and maybe the Excel sheets and the legacy systems and the green screens that we used to use don't quite match what we do now, but it's just so much to get over that hump and change, let's just leave things as they are and continue to move along.
And while there is obviously some credit to that, there's always effort and cost to get over the hump, I think that the abilities that we have now with modern systems, with artificial intelligence, with cloud-based software as a service solutions that update themselves automatically, businesses are in a great position to gain so much and to mitigate so much risk by getting off of legacy systems where there's a huge cybersecurity risk, where there's potential risk of data loss. There's certainly some opportunity costs there by not achieving the efficiencies that you can achieve with modern system.
You nailed one thing in there that I think is what I hear more than anything, which is I've always done things a certain way. There's going to be tremendous pressure internally to keep things that way. And I would say my number one piece of advice would be to have executive leadership layer change management in place at an early stage, way before you go to try to make a system change and try to educate your employees and say, "Look, guys, we're on legacy software. We're doing things the old way that we've always done them and it's not helping our business. Here are the reasons we want to change and here are the things that it's going to do for you as a person and for us as an organization. And once we get over this hump, then here's what the future looks like."
Empowering your organization to understand why you're doing it rather than forcing it down their throats I think is a big thing from a change management perspective. And then working with a few different strategists, advisors, consultants leading up to that project to frame the blueprint. What do you want to do? What is your business looking like from a growth perspective and what are your pain point? And what are your foundational items that you want to tackle with this upgrade or with this system re-envisioning?
The last thing there is I'll just say that you typically want to stick to those foundational items initially. Going to be very tempting to try to boil the ocean so to speak, or do everything that your organization has ever wanted to do because this is our turn to change. It's more important right now to get your foundational systems in place, get your data, estate in order so that you can take advantage of things like AI and machine learning that require solid data and systems that are modern. Ask yourself, "What does my business really need? What have we been doing ad hoc for years and what has to change now and what do we want to do for stage two?" Build that foundation first.
Jonathan:
Do you advocate a slow transition, as in we're still using the old system and slowly move into the new one or old people are still using the old system as we're bringing new employees are in the new system, or you believe flip the switch? Like Friday night, we're using one system, Monday morning, we're using another one. Which way do you prefer to go?
Ryan:
So I would say that it is the latter, but with a caveat. Definitely you're not going to want to do things that are supposed to talk to each other. For instance, let's say I've had people come to me and say, "We want to change out our financial system. We're going to leave all of our operational side stuff in place and then eventually we'll bring that in as well." And while there are technically possible ways to do that, you end up with data discrepancies and you end up doing throwaway integrations.
To answer your question directly, Jonathan, certainly I would not recommend running two systems in parallel, especially financial systems of record, right? You're going to want to flip the switch on a Friday like you said, and go into the new system on a Monday. And while I do phasing functionality in, it's more for non-business critical, nice to have type functionality. Let's say that you've got your distributor, you've got to get sales orders and purchase orders and inventory management and financial management. Core to your business. That's got to go in one group, right? Typically, unless there's extenuating circumstances.
But maybe you've got something that you've been doing in Excel, depreciating your fixed assets or budgeting or something bit peripheral. Those are the things that you can take to the side and say, "Okay, we've been doing it this way a long time, not necessarily integrated to everything else that we do. We can take that and break it off as a secondary piece."
Now, there are opportunities to bring those things into a project, but a lot of that's going to depend on internal bandwidth. You never want to try to start one of these system transitions when you don't have the internal bandwidth, a lead dedicated that can do at least 50% of their day for the duration of the project to this project, an internal project manager, so to speak. Subject matter experts in each area who can devote a good amount of their time aside from their day job that they're used to doing to dedicate to this project.
It's the number one thing that I'll tell companies looking to make a transition, make sure you're ready or it won't work it. It's a two-way street. You've got to work with whomever is implementing the system for you and make sure that you've dedicated appropriate time. As in any relationship, you've got to put in what you want to get.
Jonathan:
A lot of boards of directors and companies are now saying to the CEO, "We want to implement AI this year." And then the CEO says, "Okay, what do you mean by that?" And they go, "We don't know. We just know we want it." And that to me is I can't think of a worse assignment. "It's just do it." "What do you want me to do?" "I don't know, but once you do it with it."
And for my place as an outside consultant specifically, I just focus on AI. A lot of people really have different ideas of what that means. So you could ask a hundred people, what do you think the best way to use AI for your company is and you'll get a hundred different answers. And there's so many things it can do. And most of the new stories and most of the news cycle is about really big hypey things that aren't useful for most businesses. Everyone wants to talk about AI to video. And they all say to me, "Oh, we're finally going to do AI to video." "So how many videos did you make last year? How many animated videos?" "Oh, zero."
If you're not doing it before, why would you do it now. If you were last year, great, then it makes sense, right? If you had an animated cartoon character in your commercials, okay. But for so many people, they're excited by things that their business doesn't do now. So they're thinking, oh, things we could do, which to me is always the rainbow or the grass is greener thing.
From your perspective, coming from an insider's perspective is what is the right approach to take? Because my approach is always, I say, "Let's first sit down before we touch anything and figure out where we can get the most bang for our buck. Where can we get the biggest movement of the needle and start there." A lot of people are excited by image generation and writing blog posts, and that's not where most companies revenue generation comes from. It's the exciting part. But I always look at efficiencies and from my experience, two of the biggest inefficiencies are how much time people spend, three, attending meetings they don't need to be at, answering emails or dealing with emails that don't need a response. And number three is looking for information in the system somewhere and they can't find it or it takes longer.
Even for a solo operation, even when I'm doing stuff all by myself, I'll lose files or videos I recorded a week earlier. No matter how well I name things to organize, then I'm like, "Which cloud server did I back it up to? Where did I put it?" So I know that's a big question, but I would love to hear your perspective on that overall let's build a plan before we start flipping switches.
Ryan:
Yeah, you nailed so many great things there. I think the first thing you said is something that I hear all the time, right? C-level executives, management want AI now, I'm not quite sure how just yet. And then how do you get to that point where you could start to take quick wins with it as well as set yourself up for long-term bigger wins.
I love what you said about what do you want to get out of AI. There are a few different things. You mentioned access to data. Rather than sifting through endless data repositories manually or with limited search capabilities, being able to ask an enterprise chatbot like Microsoft 365 Copilot, which will look at your Outlook and your SharePoint and every other Microsoft system you have and spit out meaningful results. How do I create a sales order in the system? How do I onboard a new customer?
Empowering employees to take tactical type actions like that is great now, and that's a low-hanging fruit, right? Finding answers, finding access to data, guiding users. How do I fix this printer? What's our policy on X, Y, or Z? Those are the things that I would tell the CEO in your narrative. Kind of start tactically with that access to information and empowering your employees to do the things they want.
The harder one to attain, but one that I would also give the CEO concrete steps towards doing would be the machine learning, greater crunching of numbers that comes from looking at your customer's transactional history over the last year, two years, three years, understanding how often they pay their bills on time, how often they pay their bills late, do they pay early? Are there certain customers that we need to change terms with? You need data that is in a digestible format, not data that's sitting in Excel and in SharePoint and on an old on-premise system with limited connectivity.
Aggregating that critical data and cleaning it up and making it usable by a machine learning algorithm is how you set the foundation now, Mr. CEO, for doing that later. And commonly it's the things that you mentioned that the tactical type stuff that you could do now. And then beyond that maybe it's in getting useful reporting together. And there are different people that need different levels of information depending on where they are in an organization.
So you've got your executive level who basically wants big aggregated data, summary type data, and you can get that typically with your dashboards and your financial reports, your KPI. But being able to then ask a question to that data, ask Power BI a question about, "Hey, what is our sales trend in region A for product B over the last six months," verbally, this natural language recognition that AI has and the ability to then query the data and answer it for you. If you're an executive, that's what you want, right? Let me ask a question about my data. I don't want to sort through endless reports.
If you're a layer down and you're in middle management, maybe it's creating that ad hoc report. Maybe it's using something like Copilot for Power BI where you can actually create a Power BI report with voice prompts. Now, this is in preview right now, but it's coming. It's right around the corner. I've seen it demoed. And using that ability. When I say ad hoc reported, that's when your boss comes to you and says, "Hey, give me a report for this trend of X, Y, or Z over the last year." And I don't have that report, right? So in old days, you're a SQL report writer or even modern times you're a Power BI report writer and you go do it. Well, now you're going to be able to as a middle manager create that ad hoc report very quickly. Or even if you're on the front lines and you're looking for one piece of detail, that step below management, two steps below the C-level where you're saying, "I need to find invoice 1, 2, 3, 4, 5, or all invoices with widget A on it." You can ask those type of questions as well.
So no matter what part of the organization you're in, getting that access to the data is going to be much, much easier. But I would tell the CEO in your narrative, before you take advantage of the machine learning and the more advanced sides of the analytics that are coming, and nobody can quite say all of what is coming yet or even a fraction, but nobody knows, get your data estate in order, get your systems in the cloud with a trusted platform, trusted provider, solid technology. There are several of them out there. Of course, I more live in the Microsoft world and certainly they've got it. They're poised to be the leader there. They very highly partnered with OpenAI, tons of investment there and tons of programs for concrete use cases.
So I think this will be especially important for companies who are industry-specific in terms of their requirements and their business systems very much, hey, I've got to have something for my business. Maybe I'm a flooring installer or a government contractor or a food manufacturer where maybe all the general high-powered systems don't meet my granular functional needs right out of the box. What I tend to see is let me go look at a small industry-specific player and focus on that feature functionality. And that is obviously important. You have to run your business. But you have to make sure you're partnering with companies that have vision and capabilities to take you to the next level of your journey with artificial intelligence as well. There's a technology layer to it for those companies that kind of goes beyond that.
I think you nailed it, right? How do you use AI tomorrow? That's tactically it's answering questions. If you're a customer service rep, I want to answer the question that you just asked me by maybe punching into Bing Enterprise or ChatGPT. The next layer of that is once you have your knowledge base up in the cloud and somebody asks a question through the chat and your chatbot can intelligently find that answer and spit it back to them, there's a bit more involvement there from a technology standpoint. It's not that tactical layer of typing it in and getting an answer, but you can use that stuff to great effect.
The premium-level bots, like the Copilot in Bing, the enterprise-level Copilot, you pay certain small number of dollars per month for is tremendously better in terms of providing results than the free one. And the amount of time that it should save your team if being used effectively will dwarf the amount of money it costs. So investing a little bit there and then encouraging your users to leverage this. When they're coming looking for answers, pointing them back to the AI and saying, "Hey, I put this into AI and here's what I got. You could have done this yourself quite easily." Those are some good ways to get started I think.
Jonathan:
So you brought at the end of something really that's important to me, which is that I always tell people, and I always believe, look at how do use AI to save time first before you look at I want to do a new cool thing, I want to add a new feature, I want to replace my customer support with a chatbot. First, let's make your team fat. Let's maximize efficiency. That's where the biggest opportunity is, I think, and the safest opportunity.
Whenever you go outside your area of expertise, you add in a new thing, you add a new functionality, people always come to with these really... There's like a new way to run a completely AI podcast. I was like, "No one's going to watch that." It's this, "Why would you want to do fake video, fake people, fake topics all written by ChatGPT? No one wants that." That's something new. We have no proof. Let put it this way, we have no proof that anyone wants it because it doesn't exist yet. It's a hypothetical.
So every time you enter a new market, there's risk. And every time you add a new function on, if you don't already have an expertise, then you can't error correct. So for example, if I have ChatGPT translate my book into French, if I don't speak French, I can't tell if it's a good job or not. I don't have the ability to error correct. So I always tell people before you get excited by that, stick to the things like you're talking about, organizing your data, better communication with your customers, the ability to do things faster because if you could just increase your team's efficiency. And this is how I talk about employees. I think of employees as if it's a full-time employee, I have 40 hours a week I can deploy. I have 40 hours of each employee, so every time I have more employees, you get another 40 hours. If I can get them to do tasks in less time, especially at scale, that's where the biggest numbers are made.
Even at my size company where I've always been 20 employees or less, every time I focus on processes or systems, revenue goes up more than any other thing I can work on. It's not the exciting stuff, it's not the fun stuff. It's always the mechanics, it's always systems. It's always if I can do this process a little bit faster. I'm working on a new process right now that I'm designing. Before I'll push it down to a team member, I have to master it myself so I can write down the steps so that I can give them exactly how it's going to work, and after I finish my testing. And I know that then can scale. I always think about scalability.
I want to dive in one thing you mentioned, which a lot of people may not understand because a lot of people hear terms like heuristics, machine learning, AI, and they get kind of used interchangeably. And then you start hearing about AGI and weak AI and strong AI. And the first thing I try to tell people is ChatGPT is not actually AI. It's not on the path to sentience. It is really a strong machine learning or a strong word guesser, but it sounds cool to call it AI. So that's where people start to get thrown off. So you talked a little bit about machine learning and when is the right time... That's more what I want to dive into. What's the right use case for machine learning versus AI versus an internal AI that has access to your data versus a public-facing AI?
Ryan:
Yeah. I think the answer to that is when is your data ready, when are your systems ready, and when do you have a clear-cut use case for what you want to do? Let's start easy with that. Office has great AI capabilities now with Copilot that your team could take advantage of using your internal data, but maybe not your P data or your CRM data. It's the data inside of Microsoft 365 that lives there already. So if I say, "Hey, schedule an appointment for me for Jonathan and Susie and Jake and Tom," and it goes intelligently, you give it some parameters between 2:00 and 3:00 on a Wednesday, and it will go find you a spot very quickly and allow you to create that meeting. That's something everybody does in every business, the most ever that I've ever talked to is try to schedule meetings. And that's a great example of what you mentioned, which is tactically doing tasks that I already do, which I totally agree with by the way, and making those tasks less time-consuming.
Another one that is really important for me is the ability to leverage the data that sits in teams after you have a meeting. And while my whole career I've been doing Microsoft Dynamics for 15 years now, I've been scribbling down notes, or as I became a bit more efficient, I would type notes into CRM as I'm talking to people and I'm catching up on notes and I'm trying to establish what my follow-ups are and I'm also wanting to pay attention to this conversation, not just be a scribe.
And so with teams now, you turn on the recording feature with Copilot, not only will it record the meeting and give you an archive of that, but it will transcribe the meeting for you and it will identify tasks for me that I have to follow up on based on natural language model. It predicted basically what I promised. You told customer that you would set a next meeting for next week and we'd do a demonstration in two weeks. And then it will automatically actually go and create activities for me to execute those follow-ups in the CRM system.
So for me, from an actual use case perspective, that kind of thing is tremendously useful for saving time, for making sure everything is captured properly. And leveraging your data to do that kind of thing is super important. To get to the bigger stuff, the cash flow analytics, the supply chain optimization, stock out warnings on the inventory levels, that's where you've got to have the data, the business systems data to back it up. And so for any CEO, that's my main message is getting things aggregated. If you have disparate systems, maybe you've grown through acquisition and you've got five, 10 different business entities on seven different systems. It happens all the time.
Creating a unified data model, data is stayed in a data warehouse of some sort and making sure that the data that's gone into it is clean and that you have a unified model for keeping your production data up with your current systems or with your new systems as well as with your legacy systems is going to enable that kind of more complex crunching of numbers. And so next year when you go to create your budget and you ask Dynamics 365 Copilot, "What should I plug into this budget cell for this year?" And it goes and looks at the last two or three years and says, "Your variance has been X, Y or Z, so maybe you need to put a little bit more budget here and a little bit less there," intelligently. That's great useful stuff. But it takes a cleansed, organized data typically sitting somewhere in the cloud and ready to use. So those are just a few thoughts on that. I don't know if I answered your question.
Jonathan:
No, I think that's a really good one that I want to bring this home on, which is that the biggest issue most small, medium, even large businesses have, it's not lack of data, it's lack of connecting the data, organizing the data like your CRM. The CRM emails, they're there, some people are using Salesforce, some people write their notes by hand and they'll put into the system. Sometimes you have a support quest doesn't get copied over, something gets lost.
It's not the lack of data. I think that's the most important thing for people to think about, which is how can I have a cohesive place where all the data's stored in a single structure so that I can connect it, whether it's an AI front-end interface or a dashboard, because most of the time we have my ad spend over here, then my sales tracking over here, and we have to merge the two pieces data even to know if you're running a profitable ad spend, let alone getting more complicated.
So I love what you're talking about. I think that this is important for people because a lot of times we just want to talk about the glamour stuff with AI, but the really important stuff is starting at the beginning. Like I mentioned at the beginning it's like, where's all my data, organize my data. And I want to see more AI tools moving in this direction. Everyone's trying to be a new ChatGPT. Everyone's trying to be a new chatbot. Everyone wants to be an image generator. It's like the really powerful tools. Everyone I talk to says I lose files all the time or I have to spend time looking for files. We all do. Or we're looking for a small piece of data and it's which sheet is it in, which Excel spreadsheet, which document is this? So finding data.
And I'm also interested, I think a lot about because now every meeting gets recorded and it's just now we just have more data. If we don't do anything with it or put it anywhere, it's actually makes the problem almost worse than better. So I love what you're talking about. I think this is really useful for people who've been listening or thinking about the right way to use AI in their business strategically rather than based on hype.
So this has been really great. Where can people find out more about what you do at Western Computer, more about the type of work you do and figure out maybe if working for you is just the right process for them, so this is the guy we want to help us get our system set up?
Ryan:
Yeah, absolutely. So westerncomputer.com is our website. We've got long history implementing systems, helping companies transform digitally, get off legacy systems and look towards the future. And part of that is the data like Jonathan mentioned and like we've talked about a lot.
And one last thought that I'll leave you with is the only way to have standardized data going forward that you can use is by having harmonized processes. And that's typically easier on the financial and inventory and operation side, because you have to run your business, you have to receive goods, you have to generate financials, than it is on the CRM side where you might have Don who's been there for 30 years and keeps everything on a notepad and you might have Jake who's brand new, puts everything you want in the CRM and everything in between. Unless you've got mandated, documented processes that are enforced, when you go to query what's my average lead time, how long is it taken me to get to an opportunity, how long is it taken me to close, you're not going to have that information if you don't have standardized processes as well.
So feel free to reach out. I'm on LinkedIn as well, MSDYN Solutions, and of course you can find my name Ryan Pollyniak, shoot me a message if you want. There's only a handful of Pollyniaks in this world and I know most of them, so you should be able to find me pretty easily.
Jonathan:
Amazing. And I'll put those links below the video and in the show notes for this episode. Thanks everyone for listening to another amazing episode of the Artificial Intelligence Podcast.
