Analytics Architecture: Data Mastery William Answers Your Questions

Hello and welcome. My name is Mark Horseman and I am a data evangelist with Dataver. We would like to thank you for joining the latest installment of the monthly data webinar series, Analytics Architecture with William Mcnite. Today we're going to have a little bit of fun and answer some questions uh from everybody that we've collected over a while and over by email as well uh in what we call data mastery. William answers your questions. Just a couple of points to get us started. Due to the large number of people that attend these sessions, you will be muted during the webinar. For questions, we will be collecting them by the Q&A section. If you would like to chat with us or chat with each other, we certainly encourage you to do so. And just to note, the Zoom chat defaults sent to just panelists, but you may absolutely switch that to network with everyone. To find and open the Q&A or the chat section, you'll see the icons for those features in the bottom middle of your screen. As always, we will send a follow-up email within a couple of business days containing links to the slides. And yes, we're recording the session and we'll send that recording as well as any additional information requested throughout the webinar. Now, let me introduce to you uh William Mcnite. William has advised many of the world's best known organizations. His strategies form the information management plan for leading companies in numerous industries. He is a prolific author and popular keynote speaker and trainer. He has performed dozens of benchmarks on leading database data lake streaming and data integration products. William is a global influencer in data warehousing and master data management and he leads Mcnite Consulting Group which has thrice placed on the incorporated 5000 list. And with that, I will pass the floor over to William to get started on this unique webinar. >> Thank you, Mark. And hang around, Mark, because you're involved this time well beyond that introduction. Thank you for the introduction and welcome everybody. So, as Mark alluded to in this series, we've received a lot of questions that I had just haven't been able to get to at the very end. Now, some of them, of course, they're timely to the topic of the day and that time has passed, but some of them I think you may still have have those things as questions. So, and I wanted everybody to hear the questions. Uh, they're they're that important. And furthermore, I said, send me your questions. I've been saying that for a couple months now. Send me your questions for this session. And I get a steady flow of questions uh to me regardless. and I've been sort of piling them up over the past month or two uh for this session. So hopefully some of you that have asked me those questions, you're back. Hopefully we get to your questions and I can't wait because there are some interesting questions that have come in. So I put them on slides as we'll go through. Uh but also I welcome your questions here today directly. So put them in the Q&A. Mark's going to be uh monitoring that for us and we'll just have that kind of session today. This is uh these are some of our some of our clients. But Mark um I think we're ready for our first question. >> All right, let's uh let's get to it. Whenever an upstream data pipeline breaks because a business user changed the schema without telling us, my team instantly gets blamed for the broken dashboards. How can we stop apologizing for system failures we didn't cause while still maintaining a collaborative partnership with the business? This is a hot start, William. I love this question. >> Yeah. And and and we got a I get a lot of questions about um the relationship between the business and and the technology team. Apparently, we haven't solved all that. So So here we are. This is a classic question. Um, now when when the customer, I'll call them a customer internally, is is blaming you for this, they are really, it's probably not personal. Okay, don't take it personally. I think it's more that they're reaching out and they're frustrated because they wanted to do something and they can't do it because they're they're being stopped by something. And if you're going to absorb the blame for this, then by all means, they're going to uh open fire in your direction. So, we need to start changing how we're reacting to these uh these outreaches. Instead of apologizing, which implies you're to blame, say something like, "Thank you for pointing that out." And maybe a little bit of the reasoning behind why it happened. the underlying schema was changed upstream without prior notification to our data integration team which broke the integration. That's what happened. We're working on fixing it now. Going forward to keep your reports live without interruption, we'll need any schema modifications routed through our change management checklist. Okay. Now, that implies Mark that you have a change management checklist. If if you are you cannot sit here and expect the users to do something the customers to do something if you haven't written it down made it a standard got your management to buy into it. >> Yeah. >> And and put that forward and say okay you know here's what we're following and it'll take a little time but here's what we're following. If you don't have that, it's just it's just an implied agreement between your within your brain. >> Yeah. >> And there's nothing for them to follow. So, yes, they're going to reach out and they're going to blame you and and this and that. So, create your policy, get your management buy in, start changing the direction of your response. >> Yeah. I I love your the point that you make when you answer this, too. It's like we don't have to apologize outright like saying I'm sorry for for the downtime or I apologize for the inconvenience. We're taking an ownership of the failure and and your your exact verbiage is right. Hey, it's like thank you for pointing that out. We're working to resolve the issue and and that's the royal we, right? The Wii could be anybody. Um and so like this can transcend to various different types of issues as well. And I I I really appreciate the point of view of taking um a serviceoriented uh mindset to this. The users don't know where to go when there's an issue. That's why they're contacting us. It doesn't hurt us to to quarterback the issue and and you know, run point and we'll say we'll update everybody as soon as we know more. We'll have a post-mortem and we'll get down into the problem. We'll we'll improve the the processes moving forward. Um, and not only um uh do you show some uh service level to users uh but you're committed to making the environment better even though it wasn't your fault in the first place. So there's a lot of uh um opportunity that happens when this kind of breakdown happens. >> There is. Looks like we have another question. >> We do similar something similar. Oh gosh, I love this one too. Our business stakeholders constantly try to tell our engineering team how to build the data lake or which AI tools we should purchase even though they don't understand the underlying architecture. How do we politely but firmly shut down this unsolicited technical advice without alienating our business partners? Sorry, I'm giggling because a story popped into my head. That's hilarious. But take it away, William. This is another theme theme of of the of interaction right between uh business stakeholders and and build teams, technology teams. There is a reason that big data is such a big industry. There is a reason that it is a top one or two or three spend in it today and that's because there's a lot to it. There's a lot to it. Uh but conceptually it can be rather simple. Well, you just want to collect the data and put it there for us to have access to. >> Conceptually it seems pretty simple. And this is what makes uh some business stakeholders feel like, oh, maybe maybe I'm an exp I read this article. Maybe I'm an expert. You know, I talked to this vendor. Maybe I'm an expert now. We ought to do this. Um, one thing I want you to work on is your credibility so that they understand that you have their interests at heart. You have the interests of the company in mind as you make your decisions about the tools that that we use internally. And there's a lot to it. Thank you for raising this tool. We'll add it to our architectural evaluation backlog. Mark, you know what I'm going to say? That implies that you have an architectural evaluation backlog. Yes. >> That you have a process that you follow. If you don't, if it's all willy-nilly ad hoc, okay, they're participating in that. >> So, what's your process? If you don't have a process, don't you can't blame them for saying use this tool because they're just they're just seeing what you're doing >> and doing the same thing. >> A process solves so much of this. Exactly. I love that >> it does. And and I I also want them to say we evaluate all vendor tools against our data governance, security and integration standards before making a procurement recommendation. And that should put them a little bit on their heels because you said a couple words in there that should trigger them. Security. >> Are did they think enough about security when they made their their recommendation? What about integration with everything else going on in the in the company? Did they think enough about that? Probably not. And they're and so, you know, if they're if they're cognizant, they'll think, oh, you know what? I don't really want to go out on that limb and and make this recommendation because I didn't think about security, data governance, integration. How about data quality? How about this? How about that? Oh, they're thinking of that. They're thinking of that. they've got the credibility for that. That's who I'm going to let make those decisions in the future. >> The the one thing that I've run into, William, and and maybe you can uh kind of get into this a bit, is when uh members of my executive, my seuite team, go away to to a conference and then they see a peer organization who's running all of this stuff that they've presented and and you know, when they're presenting it, they're making it look like everything is sunshine, unicorn, and rainbows, right? And then they're like, "Oh," and they said they'd share some of their code with us so we can just buy it and lift it and drop it. And the entire data and IT teams are rolling our eyes so hard that we we crank our necks. >> Yeah. Very similar. Is Isn't that Isn't that the truth? Yeah. Um, again though, you know, uh, I'm I'm glad they went to the conference and they learned some things, but >> hopefully they learned what they don't know. They've learned that there's more to learns. That's what I'm I'm always learning. There's more to learn. I'm always learning how little I know because there's so much more to learn. And hopefully they see that as well. >> Exactly. >> Well, we have another question, Mark. >> All right. Our data team is constantly treated like an ad hoc help desk, getting pinged late at night for emergency data polls that are rarely actual emergencies. We feel guilty saying no because we want to be seen as helpful. How do we establish firm boundaries around our queue without feeling like we're letting the company down? This is the milliondoll question right here. We >> don't want to let anybody down now, do we? Okay, this reminds me of back when I was running uh the data team and I inherited a situation where uh of course it was it was a 24 by7 uh kind of operation and somebody was on call all the time for a week and then we we handed off. I I took my turns on that as well. And I inherited a situation where the the the systems were going down all hours of the night, all hours of the day. And it was known that when it was your week, that was going to be a hell week. You were not going to get get good sleep and you were going to be grumpy and you were going to be uh very busy. So, what I set out to do was say, "Okay, hey, I know if we get a call at 3 in the morning, you're just going to patch it. You're just going to fix it. But once you get your sleep and you get up, you have got to make sure that never happens again. That that error never happens again. Go back in the code, make sure it is shored up so that it never happens again. And one by one, we're going to knock these down so that our call volume will go down to a much more manageable place. And it did. Now, this is a little bit different. The question comes from a a place of it sounds like somebody, you know, they're actually they're actually calling for help. Um, do you have, again, I'm going to get back to, do you have a policy about this? Do you have a policy that your management has bought into and is trying to put forward to the organization? An important element of this policy has got to be the severity. The severity of the problem. If it is a problem that you know really the company's going to hurt if if you don't get up at 3 in the morning and fix this thing and do this ad hoc request, whatever it may be. Okay, we got to acknowledge that. Hopefully, we learn from it and we try to fix the whole thing so that it doesn't happen again at that point. I want if you're on call, I want no calls. I want no calls in the night at least. Okay, number one. Um but for severity, um you've got one, two, three, and four classic, right? Number one, we get it. Okay, get on it. But for the other things, you need to be able to say, "Hey, I just saw your note." Just confirming, does this fall into a sub one outage, like a customer pipeline is down or there is an incorrect metric going to the board or something? If so, I'll jump right on it right now. If it's a sub three, I'll log it first thing in the morning so we can prioritize it in the daily queue. >> And so this al this all implies that you have you have priority. you have severities and you've you've lo you you've characterized it and your management has bought into that. So I'd say severity uh work look at your severity but also make sure that you are solving problems once and for all. >> Yeah, thousand%. when when I was struggling with this back in ye olden days and still to this day uh like even most recently um a lot of this comes down to um our executive teams and and management and leaders around the organization not understanding the lift it takes to ask for a brand new dashboard metric or report. Um, and so really this in my mind it it it really struck as an educational component. So this is what it means when you ask for something. It's not like we're just mashing some stuff together on a spreadsheet. I know you can do that. Uh, but there's processes we follow to ensure that the numbers and and and content that you're getting is accurate and meaningful. Uh, so we need to be able to uh support that when when you have issues. Now, if you want to get your own content faster, >> then let's spin up a project to do uh democratized reporting or self-service analytics or can you please give me some money for my data warehouse so that I can support you having your own PowerBI or click or Tableau or something. Um and and I've actually gotten funding for projects as as a leader at organizations that way. Uh so there's there's a lot of power there. There's that threshold between, okay, we're we're fixing it now ad hoc and oh, that's a project. Okay. And you have to have firm boundaries between the two. >> Yeah. >> Well, we looks like we got another question. >> Awesome. Awesome. All right. Uh when pitching data quality or master data management initiatives to leadership, we often soften our language using terms like we think or maybe because we're afraid of losing budget. How can we sound assertive and confident when presenting the critical necessity of data foundation projects to executives who only want to talk about Gen AI? >> Uh you could throw some other things in there too like data governance, right? uh foundational things that and data architecture foundational things that uh enable generative AI but it's not apparent maybe to to those who are in charge of budget you know that that is the case well here's my take stop framing data quality and master data management as data quality and master data management because you know what when you say this when I say this to an executive you know they're thinking in the back of their head, oh, oh, such and such employee just he he went off to a he went off to that conference. It was all techies and and he learned some new language and that he's excited about it'll it'll be here today, gone tomorrow, and here he is in my office talking about these fufu things, okay? >> They don't want to hear it. They don't want to hear that. They want they are responsible for what they're responsible for, the bottom line, you know, of the company. Usually when you get to budget level like this, there's some bottom line responsibility there. Sales, expenses, new customers, things like this. So you have to line up how these things are going to enable those things. And they should >> and they're not wrong. They should. >> U so do you feel confident that what you're proposing lines up with the objectives of them and and the entire company? If not, go do your homework and make sure that it is aligned. But if you're aligned, you should be confident. You know why? Because we we data people. We sit on the gold of the organization anymore. We know how to take companies forward in unique ways. It is our responsibility to to to assert this inside of our companies. It absolutely is. And as a matter of fact, if you don't and the company doesn't do anything, sits on their hands, doesn't do data quality or master data management, for example, uh down the road they're going to ask, well, why didn't we do this? Why didn't we needed that? Now we're in a hole. Why didn't we do it then? Uh you know, Joe, why didn't you tell us? You're you've been sent off to these conferences. you you know you've been softpedaling this stuff but you should have been in here you know telling us how it affected our bottom line instead of using that fufu language. So start positioning it as risk mitigation and the data fuel source for the exact generative AI initiatives that the executive leadership is obsessed with. I'll add one more thing here. Sometimes I won't get budget for data quality or even master data management in but it's it it's it's the right thing to do for a given business initiative that I have and so I won't use these terms. I'll use I'll just bake it in. Oh, we're doing targeted marketing and that's just part of it. That's just part of it. We're doing uh uh fraud management and data quality. It's just part of it. you know, can't have bad data as we're, you know, doing data quality, but I'm always going to do it with the long term in mind as well. I'm going to do it solid. I'm going to architect it so that it works for the future. It's not a oneandone for this project. >> Oh my gosh, William, I love your answer so much because it touches on all these foundational things that that we try and do at organizations. Really, we're not talking about data management specifically. We're tying data work to the strategic goals of the organization. We're when when people when we talk about master data management, our business is talking about customer 360. We're enabling customer 360 and and my biggest success in doing data quality is analyzing business impact. Talking to our executive team in terms of return on investment. If you want to go down this road and meet this strategic goal, then if we invest in a data quality program, we're going to be able to have this kind of return on investment. And um when I was in higher ed um as I was for a number of years, we did u mailing uh um uh campaigns to alumni. So, hey, you went to school. Would you like to donate back to school so we can set up a scholarship? um um and so on, right? Uh so we'd have these donation drives and like 30% of surface mail would get returned to sender. And so I just I went up and I said, "Hey, you know, we got this much money. Imagine if we cleaned up 2/3 of those surface mail addresses, we'd get 20% more returns. Um give me money for for tool, please. It's it's less than we would have made." Um, and and sometimes when we can talk in ROI like that, it it's a slam dunk for business leaders and and it's not even a question. >> And they say you don't need math after college. >> That's right. >> Okay, Mark. Uh, one more of these can questions and I'm I'm going to invite you to go to the Q&A uh here and see if we have any live questions. But >> uh we've we've got a couple uh cooking already in in in chat and Q&A. Uh but oh, I love this question, too. Let's see this one. Yeah. >> I love working at my boutique data consultancy. The culture is great and the work is meaningful, but inflation and life changes mean I need a 25% pay bump just to keep up. I know our margins are tight and I'm terrified that asking for market rate compensation will burn bridges or force me out. Is it greedy to push for a raise or is it time to leave my dream role for a big corporate enterprise stack that actually pays the bills? Okay. All right. Did my employee do this? No, I'm just kidding. Okay. So, first of all, I would say you don't love your job because pay is part of the job and 25% if you think you're underpaid by 25%. Uh, that's pretty significant. Now, what you have to look at is the overall picture of the job. Jobs pay us in different ways. Of course, there's the salary. There's bonus and all the money part. There's also, you know, the the benefits part. Okay, all good. But they also pay us in quality of work, work from home, the quality of our peers, the quality of and the compatibility of our boss and the interesting nature of the work that we do and the technology that we get to work with. All these things have to go together and only you only you can say what the right proportion of all that stuff is for you. if you're quote unquote the bread winner and blah blah blah and you really need to be making you know market in this case 25% more you really need this this is what you're saying then then you really need to look around for something more significant in that area and I'd be sorry to see some of these other things go by the wayside because that's important too but you as as they say you can't have it all so pick your battles and determine is that 25% gap important enough to go look now. There's no there's no expense to go look. Go look now. Go look right now. Start looking. Make sure you're saying 25%. Make sure of that because it's only 25%. If you can get an offer that actually is 25% more than what you're making now, not not theoretically you heard from this person, you heard from that person, or you went on this or that site and it looked like looked like the salary was 25% more than I'm making, same qual. It's it's you got to go get it. You got to pay the price and go get it. Now, there's nothing wrong with going and looking kind of at a low grade. I mean, you're you're you're covered right now. This is a good time. Go look. Now, that's a bad time in the market. Don't don't get me wrong, but it's a good time for you to go look and see what's out there. Start maybe five hours a week, four hours a week. Start looking. Start seeing what you can get moving in the right direction. And maybe you'll learn in that process. It's not 25%. Uh maybe you'll learn in that process, wow, it's way more than 25%. I can't wait to get out of here and get that extra money. Okay, >> but you got to pay the price and actually put in the time and get some offers going. Making sure that you know your proportion of that pie chart is covered fairly accurately around all the the ways that companies pay us. Uh so I suggest that you start to take a look. Uh, I don't suggest you are ready to go to your boss now and say, "Hey, I need I need that paybook. I I want to feel a little bit better about it before I go in with with I want to I want to bring some numbers." Okay, I want to I want to bring some some data, right? We're data people. I want to bring some data. You might at that point, we might be talking a month out. I love working here and I want to stay long term, but my family's financial needs have changed, blah blah blah. And here's the key question. What goals or milestones do I need to hit over the next 6 to 12 months to bring my compensation to that? Now, if you're in a big company that has tiers and and it may be difficult, but if you're in a smaller company, you say you are boutique data consultancy, they should be more flexible. Maybe there's a way. And it could be that your bosses are not trying to, you know, keep you down or anything. They're just taking things day by day and one day leads to another and sooner or later years pass and you're still paid what you're paid. and oops, I forgot. So, get a little more data from the market, go to your boss, uh, kind of in low key and say, "Hey, how do I get a pay?" And if that if that answer is no way you're getting there, that's crazy. >> Now, you know, now you got to put pedal down a little harder on the external things. >> Yeah. I I have nothing to add. You answered that so brilliantly. I and really there's so much at play here that's going to be personal uh to the to the person uh who's going through a situation like this. >> Yeah. >> Tough call. Tough call. >> It is a tough call, >> especially with today's job market. >> Yeah. >> Oh my gosh. So, somebody just put a question in chat that I love so much. Uh so, we'll do this one. Um much of the discussion around data mastery and ROI appears focused on private sector revenue and profitability. I call it money motivated. Um in local government success is more often measured through improved public services, operational efficiency, risk reduction, compliance, and better decision-m. How would you adapt your data mastery framework and organizational structure for a local government dealing with legacy systems, limited resources, data ownership spread across multiple departments and elected officials? So you are where you are with all that stuff in terms of the technology, your ability to deliver. You know, Mark was saying earlier about how difficult it is to get a question answered out of your architecture. Yeah, you are where you are with all that. Uh nowhere to go but up, right? It sounds like um so when I when I talk about this, yes, I talk about ROI all the time. I drive my public companies, my you know uh non-government companies towards that because yeah that's that's important but you you know you can kind of view that as a surrogate for goals of the company in your case the goals of your not company but corporation uh what would you call it a government entity um is not ROI it's not necessarily revenue although many of them are geared that way. So, I wouldn't dismiss it entirely, but you may have other goals that you're marching towards. Now, it's going to be a tad more difficult because it's it doesn't break down to math. But in in those cases, there is still a step-wise progression towards those goals. We had this much customer u um how how would you u um this much customer satisfaction as you mentioned this much uh utilization of the public utilities this much utilization of public transit whatever the case may be. So as you you'll want to see yourself marching up uh upwards in terms of those goals. Whoops. In terms of those goals. And so whatever the goals may be, you just substitute uh in in the ROI for that using those. Yeah. >> The reason I love this question, William, is because I I as you know, I've worked like almost 20 years in higher ed throughout my career and higher ed in Canada. Uh so we're not a money motivated group and and I have so many folks and friends that I know in government as well. Um uh and and really those departments are focused on an outcome. they're focused on a strategic objective and and when when working in private sector the questioner is absolutely right. We get stuck in these money motivated things. Um but it makes sense in that world like we can say you spent X dollars we can reduce that by Y dollars and so we can do this for a gain of zed dollars and and and those those business conversations make some flavor of sense. But when we were working in higher ed, uh there were a couple of things that we would run into. And and one of those was what we called uh uh student success. And what we really meant by that is are we doing everything we can to help a student graduate uh attend a convocation ceremony achieve their qualification. And so uh really the the data framework around that and and the the the data questions are hey how how is how are our students doing? Let's do some data mining. Uh do we have some indicators that a student might be struggling? Maybe we should reach out and promote some tutorial uh activity or some other learning resources to to specific student groups. Uh uh and so we get into those kind of questions. Really the ultimate answer is tied to a strategic objective. Now, in private sector, that could well be money motivated. But in public sector or nonprofits and and other uh things where you're not money motivated, you're still going to have a strategic objective that you can tie to and you can still show how data achieves that, right? >> It is there. It is there. Yeah. >> Yeah. >> Um great. Uh Mark, are there any other compelling questions uh for now or should we move on? We we do have one more in in in the Q&A. Uh how do we see the impact of using AI natural language queries uh for enterprise data warehouse data models? Uh the time we need to or used to spend on uh data models or EDW models uh uh now might not be justified. Uh I still vote for some basic relational model to support the business natural language queries. Okay. Okay. So, um uh first of all, how how do I see the the sort of the English interface now to our data warehouses uh evolving? Uh very strongly, as a matter of fact, very strongly. And I think that's an it's all enabled by AI. Um BI is dying as AI is rising too. Yeah. >> Yeah. And um and so as you as you state what you want from data that has to be interpreted and uh what an LLM can do is it can add in all the antonyms and synonyms and and just make sure that it's comprehensive in the query that it ultimately runs. So that makes in most cases that makes those queries fairly more superior and and and close to what a user really wants. So I think that that aspect of it is coming on really strong. Um however, I don't think that has implications on let's short change the data model as a result of it. the better the data model, the more that's going to work correctly. >> Yeah. >> And so, yes, you can get by with less of a data model and it would still work, but that's been true forever. That's been true for for BI forever. So, it's the same thing. It's just now it's AI instead of BI. So, so what are companies doing? They're they are shortch changing the data model. I see it all the time. um they're allocating their resources to places where they act they they feel like they need to have them. I'll put it that way. And that may or may not be to the data model, but I think they do that at peril. I think they short change their data model at peril. I think it comes back to bite them when they have a bad data model, when they have a subpar data model, they're missing elements, the relationships are wrong, and so on so forth. So I always say invest in that in the data model and this could be back to uh what I was talking about data quality and master data management earlier right um you want to do data modeling h that sounds very strange I'm not going to give you a budget for data modeling but I'm going to give you a budget for targeted marketing and fraud detection and things like that yeah okay well this is part of that great data modeling is part of that it does not take It doesn't take more budget, more energy, more effort to do the right thing. It just takes the knowledge and the focus and not only do you end up with a better result today, you end up with a better result long term. Yeah, the questioner added into uh into his question there think uh data bricks genie one and yeah 100% and ultimately how are your agents how are your AI solutions learning and being trained and what are you supporting them with to be trained in the first place I there's so much work happening now in context layers and semantics and knowledge graphing to make all this work. And you know what? That's powered off of a good data model. You're never going to be sad having a data model. That might be a tool for the engineering team >> as opposed to a communication tool uh to describe how data functions at your organization from conceptual to logical to physical. But I mean the AI is just another person quote unquote. Oh, I feel bad just saying calling AI a person by the way. U but it it still has to learn and do the job. People had to learn and do the job before. Just because it's shadow AI beast machine thing over there doesn't mean that it can't learn. And and then how do you prove that it was ever right in the first place? Could you imagine an AI that wasn't trained properly and provides a wrong answer and the erosion of trust that would happen at an organization if it was wrong? Oh, that'd be heartbreaking. especially if you spent a lot of time on it and executives would be upset. Um, yeah, I >> Oh, I have a good data model. I'm so sad. >> Nobody's ever said that. >> I've spent 12 years managing legacy databases at a slowm moving utility company and the bureaucracy is crushing my soul. Whoever wrote this, I love that by the way. My real passion is modern data architecture and launching my own consulting practice. But if I grind it out for eight more years until age 60, I lock in a fully vested pension and lifetime healthcare. Should I stay on golden handcuffs duty to guarantee my retirement or leap into the modern data market while I still have the drive? Well, he put this or he or she I'll say I'll just say he he put this right on the fulcrum, right? Oh, eight years to go. Uh so yeah, he makes it a difficult decision, which is why it's a question, right? So this is a very real question for people of my generation. Okay, so I'm going to guess he is 52 with some quick math there. Uh still has a passion for data. I can relate to that. Um but let me let me touch on the launching my own consulting practice part uh there uh at 52. What what is the the goal here with that? Um do you do you want constant employment or is it to taper things down a bit where if you're working you're working if you're not you're not? No, no right or wrong answers here by the way, but you need to develop that part a little bit more. Now, by consulting practice, do you mean you want to bring on people, you want to get projects, you want to have multiple people out on projects here and there, or it sounds more like you you want to do contracts, personal contracts um out there. Um, are you ready for all the things that go along with it? either way. Okay. The marketing, uh, the the the the the finance side of it, the taxation side of it. Um, hey, it's not rocket science, but it's a it's a grind, too. So, just just, you know, it I just want you to have your eyes open when it comes to the launching your own consulting practice part of this. as as far as the the bureaucracy that you're dealing with and the slowmoving utility company. Look, uh you have goals. You have goals there. And the goal could be met one or two ways, right? One way is yeah, you can just go along, get along and and crush your soul for the next eight years or whatever. Um, but the other way is to develop a plan to get the company out of the the crush that they're in right now and get the get the company get right where you are today to a place that you can be proud of to a place that is a modern-day architecture. And you do that one step at a time. It doesn't happen tomorrow. And and I don't want to hear, "Oh, I went in and I mentioned this and I got shot down, so whoops, no more." you okay? You you you gota you got to come back to it. You got to tell people like three times before they're going to, you know, even give you the time a day on something like this. You have to know that. So, I would say give give it a shot where you are today. Give it a maybe a year. Okay? You'll be 53. Uh not the oldest in the world. Okay? Uh you should still have your passion then. Um and and uh you know see if you can make make something better out of where you are right there because and here's the other thing. You want to you want to uh get your own consulting practice in modern day architecture but are you qualified are you qualified for modern day architecture? You haven't been doing it. You've been doing managing legacy databases out of slowmoving utility income. Is somebody out there going to hire you on a contract basis to do, you know, something more leading edge around AI and data lakes and streaming data and edge data and vector data? Are they going to? I don't know. I I I tend to think it's a very competitive market there. Um, and maybe not. So, the thing to do on that front is to, as I advised uh a few minutes ago, you know, start to spend some time to look around and see if there's something better for you. In your case, you're looking for a contract. You're not looking for an employer. Okay, fine. So, start looking for that contract. Usually, they're the the they're easier to get than employment. I say easier. I what I mean by that is uh the time lag is not as long. So yeah, start looking. Start looking and see. But you should have passion in your life. And you don't have it at work uh today. Um see if you can get a little bit of that going by trying to modernize the place you're at. But you can get passion in different ways. A life is a life is a life, right? You you got your work, you got your sleep, you got eight other hours a day to have that passion and make sure that you you do them you make the most of those of that time. I don't I don't care what it is. Go start playing pickle ball, uh gardening, uh uh uh uh kite flying, cooking, you know, pick it up. Pick up the other parts of your life so that you do bring some passion into your life because I think that's important as well. So, good luck. >> A lot of wisdom there, William. I love how you answer that. And somebody in chat uh um uh really um um um uh um empathized was with the story too. So uh they were saying their financial adviser would advise uh consulting on the side and keep the full-time gig. Um which yeah, that's cool and all. Have you ever like I' I've got a couple of friends who are getting close to retirement and what I find hilarious and this is like consistent and like they haven't talked to each other in like different friend groups. They know how many days are left. They have like the daytime timer counting down to when retirement is is. It's like, "Yep, there's uh there's uh 1587 days left." And it'll be like that high. >> Well, I wouldn't be surprised if this person uh isn't isn't doing things that way. You know, the other thing I want to mention is is this is this fully vested pension and lifetime healthcare deal. Is that all or nothing? you have to wait till 60 and then you get all or nothing. I got to believe that somewhere short of that maybe 55 or something that you can start some >> of it and maybe so if you quit then you still get some you know make look into that. Somebody in chat just says, "Mine's 1844." >> People know. >> I love that. This is a wonderful aspect of human nature when we get close to retirement, right? >> Hopefully, they have a passion in their life. Otherwise, >> with AI making inroads in so many areas of data and analytics, what human skills will still be in demand in the next five years? You know, it's a big spectrum of possibility there. There the we don't know. We're sitting on we're sitting on this this this this this thing that it's going to explode in in one way or the other. It has to. It's either going to bring us utopia or it's going to crush jobs and we're not going to do anything about it and life is going to go for us as a result of AI. We don't know what that's going to be. We don't know. We don't know if AI is sustainable in the way that it is put out there today. We don't know if it's too expensive. too expensive because every time we run a a query in chat, they're losing money because they're charging us so little for all the investment that they're making in in in AI. Um them and everybody the hundreds of billions of dollars spent in this circular manner. Where's the enterprise doing the spending for this? Ultimately, I got to believe that is required to to make this sustainable. And will it kick in to a level that's reasonable for that? Right now, it's not. Is it going to get there? Nobody knows. Um, and there's a lot of speculation on that. I'll leave that aside. But one thing that is certain to the question here is that some of the traditional business analyst roles uh some of the tra traditional data scientist roles I would say the DBA the dashboard developer static BI report writers the manual QA and test scriptors and legacy compliance managers. Yeah, those are less interesting today. So that would be those would be human skills that will not be in demand in the next five years. Now what will be in demand for sure is that we being uh great human conductors of autonomous AI agents. We being great >> humans in the loop of of AI and doing those things that are uniquely human, gathering uh business requirements and business contextualization, high level enterprise architecture, design, ethics, ethics, regulatory compliance and organizational change manager, change management. These are skills that that AI will not do. What about there's a there's a bunch of roles too that are emerging and um you may think about moving your career in one of these directions. AI explainer, AI ethicist, AI auditor, AI business strategist, uh AI trainer, >> uh and you know, people that can implement great humans in the loop uh around AI agents um and and things of this nature. So, anything that is quick and dirty and can be automated with with just a little bit of data, of course, that's that's gone. That's gone. You better look at your all of us need to look at our skill sets uh if we want to be working in five years. Okay? Be looking at our skill sets and and make and and thinking about AI in context of those skill sets and will they be in demand uh in the next five years. And I mentioned some things that won't and some things that will. Certainly, one thing that will that many of us are on the on the right side of history here is building AI. We're the builders. You know, we're not we're we're building the the systems that our enterprises demand with AI. Um and that will certainly be in demand. >> Well, and oh, I love your answer so much here too, William. Uh, one of the things that I've been crowing about for the last year or so is that human in the loop component. What are the guard rails? How are we defining how our organization is leveraging AI and where are the humans involved? And so there's like human in the loop, there's human on the loop, there's human out of the loop. A good friend of mine uh u um he his background before he got into data, he's got like a PhD in music theory. And so uh we get talking about orchestration. How do we orchestrate all of our AI agents to work harmonously together and produce a result? And and those are the kinds of things that we need to think about. What happens when an agent passes off a task to another agent, spins up another agent? >> Um uh what do the humans need to do with the results? How do the humans guarantee that the AI agents work together? Uh there's there's a lot of things happening in the space where the traditional uh analytics roles and and human roles are still there. They're just changing. Kind of like back in the 30s when when automobiles became popular and the nature of streets changed and we needed traffic lights so people didn't get run over. Maybe that's a terrible example. We should probably move on to the next question so we can brighten up our our webinar before the end. Will >> we'll see how bright this question is for us. >> A fellow data >> a fellow data analyst going through a tough personal situation calls me daily uh turning every conversation into a multi-our monologue about their life while ignoring my work and workload. I want to be supportive but it's completely destroying my daily productivity. How do I set limits on their venting without abandoning them during a crisis? Oh my gosh. >> The the the downside of being a nice guy, right? Yes. >> So now you're a therapist. Now you're a therapist. >> I I don't understand how you have allowed the conversations to become multi-our monologues. How have you not stopped them at some point? never enter one. I mean, okay, I'm compassionate. I want to help, but we're at work and I've got things I got to get done. And frankly, if I've got time for this uh for something, I I might want that time after work for something else. So, we're we're going to get to the heart of this. We're going to get there quicker. Never enter an open-ended call or conversation like this hoping that they're going to wrap it up on their own. This person has proved to you they're not going to wrap it up on their own. You have to set the boundary the second you answer. You know what's coming. Here comes Mary. >> You know what's coming. and and you better start looking busy because the next thing you that I want you to say is I've got a I've got uh I've got some I've got a deadline I'm working on. You are right. Certainly you are. We all are. I've got a deadline I'm working on. I've got five minutes. I've got a call in five minutes. I've and and maybe your call doesn't start till the top of the hour, but you need 10 minutes to prepare >> and you're going otherwise you're going to get to that call unprepared. It's not going to look well for you. >> That this is your this is your time. >> So set the boundary at the very beginning. I've got five minutes. Go. Let's hear it. I've got five minutes before they start venting. This sets clear expectations and puts a natural cap on the conversation. The other thing that I would do is I would start to interject into the conversation the idea of seeing a therapist or somebody else >> something. Yeah. >> You know what? A therapist could really help with that. This starts them to thinking two one of two things. Either they're going to think, "Wow, it he thinks I need a therapist. Wow, I did I didn't I didn't realize my problem's not that bad. Let me just back off. Um but more likely it it more likely it'll plant the seed that um yeah, maybe I do maybe I can't just lean on this person's shoulder. Um they got their work to do. Um maybe I do need to see a professional about this. So it's either yeah, it's real important. I need to see a professional or whoa, not so important. I need to back off. But either way, you get backed off. you get your time back and um and and and you you're still helping, but you know what? You can help in five minutes versus helping in in hours. >> Yeah. I I love your recommendations and like I I've often found myself saying, "Hey, I got a hard stop in at this time or in this many minutes." >> Um I think the the nature of how we support each other has changed significantly precoid versus postcoid. Uh so in the before times um uh we would like people would walk into your office or walk up to your cubicle and have that discussion in person or sit down in your guest chair and and hang out and that was a much different vibe than actively calling somebody on teams or zoom or engaging in a team's chat. I think the nature of that interaction has changed. Um, and so we have to be mindful about how we interact with folks cuz we we don't want to be negative. We still want to be supportive too, right? But yeah, I think this is a real challenge um in a lot of cases. >> All right, we have time for a couple more, I think, depending on how >> how how quick we are. Uh, what should your clients be asking about what they are not? Did I read that right? >> Okay, that's that's the question. Um, what are they not asking about? Probably that they should be. Maybe that's what what the question is here. Um, well, my clients are asking about a lot of things, but no one of them is asking about all the right things, right? So, um, and then none of them are asking about some things that they should be asking about. Uh and so kind of that gray area uh of stuff they all should be you all should be asking about is what about agent consolidation rationing rationalizing mounting technical debt. Mark, you were alluding to this a little bit before before the sprawl of agents gets out of hand >> and what is our plan for fastch changing enterprisewide data as we start to lay out AI agents across our enterprise and are we tackling the nonfunctional requirements that are always required even when it comes to AI agents. I see that >> happening all the time that we're deploying agents but oh what about what about data quality what about backup and recovery what about security you know we're just letting them those things go and we should not be doing that this is going to come back to bite us so it's not just full steam ahead on in development you're still in an organization that has requirements and so looking at the fuller picture is very important because we're doing a lot of AI agent deployment now And I I just want it to be right. But but here is a bigger question. This gets me thrown out of rooms. All right. No, I'm serious. Uh it's like we it's like the the the thing that should not be named. What if and I alluded to this before, but what if we're in an AI bubble? What if there are market ramifications to what we're doing? Remember I talked about the circular financing and all that. What if there are market ramifications to all that? Well, how's that going to affect us? What are we doing to to to uh wet a hedge against that possibility? Now, you may not think it's a possibility. That's fine. But acknowledge that. Acknowledge that risk position and that's okay. You can go forward with that. But I think it's it's a question worth asking. And that's what that's what this person's asking about. What are the questions worth asking? That is a question worth asking. And some of my answer is we got to spread our AI around, right? And we got to make sure that what we're doing with with AI makes sense to our bottom line and also makes sense if uh the this all reality this whole reality sinks in and the course of action taken by the the hyperscalers is to uh hyperscalers of LLM is to drive up those uh token costs like double triple etc to try to >> try to you know finally start making a profit. What if that were to happen? Just just make sure that we're doing what they are. >> Your token budget. >> Yeah. >> Could you imagine? >> Yeah. They're more expensive now than people. Is that is that right? Is that what we want to do? So anyway, got to ask that question. >> Yeah. 100%. And there was one one organization I was um uh working with on AI governance. I said to them, for every business case that we're applying AI to and you know, we we implement an AI solution, I want an ROI attached to it. I want some connection to the business outcomes attached to it, but we need to think about backing it out for when that ROI is no longer there. So, if the cost spirals out of control, how do we turn it off and go back to the way things were? Because that could be a very real situation. I We're at time. Oh my gosh. I thought we would have time for one more question. We could do this forever. Yeah. >> Well, thank you so much, William, uh for the wonderful uh uh uh casual event. Uh we've got so many more questions in the slide deck. We're going to send the slide deck out. Uh if something catches your eye, feel free to let us know. Uh and and we'll uh we'll see what we can do in the future. Uh for sure. Um, thank you to the community for all of your wonderful questions throughout the year. Uh, and uh, and sending us your thoughts and, uh, enga and engaging with us here on chat in Q&A today. Uh, William, any last thoughts before I hit the end webinar button. >> Um, keep asking questions. >> Yeah, good life advice. Keep asking questions. All right, everybody, have a wonderful rest of your day. Thanks again, William.