Own the Unknown™ Interview: Jonathan Reichental

Sa. Hi, everybody. We are live. This is going to be tremendous fun. So Jonathan and I have gotten to know each other over each other over the years, and I think this is going to be a fabulous conversation. So if you are here, presumably you saw the event announcement, and even if you haven't, you'll come to know Jonathan overtime. So we won't pause now to do a lengthy introduction. We're going to jump right into it. I will say a couple of very brief preliminaries. Jonathan's an author of multiple books, also has courses. The topics we're going to touch upon, data governance is going to be a big one, certainly, but also smart cities. And we won't stop there. There'll be some other topics as well. Also, we're going to be doing one of these a month here at Further, so I encourage you to follow further. It's possible you're here because you know Jonathan. It's even possible you're here because you know me, but you want to follow Father. Follow Further rather, because even though it'd be lovely to have Jonathan as the guest every month, that's not the plan. We're going to have a different thought leader each month. So if you don't follow further, you might miss out. So data governance is a promised topic, as we said. But you know, I'm just so intrigued with Smart City. So since we're going to start there, for a lot of folks, they might think smart cities, are we going to talk about like, design and architecture? What does that have to do with data and AI? So maybe you could start there, Jonathan, and you can even mention a little bit about how you get involved in smart cities. But why is Smart Cities very much a data and AI topic?

Yeah, thanks for that and it's great to see you, Keith. I always love our conversations and congrats on all your success. I've been watching you closely these last few months, so thank you for the invite. But also, by the way, I'm very honored and humbled to be your first guest. So thank you for that. You know, I've been. I've been in the government space and the sort of tech or urban tech space probably now. Well, it's coming on about 14 years. A little bit of a surprise, you know, getting into this work, I got invited to be the CIO and CTO of Forest City, the city of Palo Alto in California, sometimes known as the birthplace of Silicon Valley. And I kind of fell in love with the work. I fell in love with public service. But it was also an opportunity to ask a question like, what could you do with technology to advance the goals of the city? Provide more efficient services, reduce overhead, you know, create better experiences for, for community, whether it's in public safety or transportation or energy management, education, healthcare. You know, this is such an incredibly big topic. Cities for humanity now are really at the center of our experience. You know, you think about, you know, originally, you know, if you go back not so far in history, we're kind of a rural species, right? We're agricultural and we move into the industrial revolution and that begins sort of our migration into cities. And today we're, you know, 60% of all humans on the planet. Out of 8 billion humans, we live in cities. And that number is climbing very fast. About 3 million people per week is the growth rate of cities. So for the rest of the 21st century, we're an urban species. And that makes this topic really important. Cities are probably not in, in where they want to be relative to, oh, traffic management, let's say. You know, if you live in a city, I think that resonates with you. So I think this is an incredible time to think about how we can use contemporary technologies and emerging technologies to advance the needs of all types of cities in the world. And, you know, by the way, a big point here is, you know, the solutions you apply are very dependent on the context of a particular city. Like, let me just put that another way. What we do here, let's say in Silicon Valley, excuse me, in San Francisco, where I am in the Bay Area, would be different to say Barcelona or Melbourne in Australia or even Neom or Riyadh in Saudi Arabia. You have to take into consideration culture, geography, politics, you know, finance, the money part of it, right? So we're looking at new tech. How can we move things forward? And, well, you asked me about AI and data in particular. I'm excited to talk about both. They're both massive topics. As you well know. Data is the most important asset in every organization today. I know this because I spent the last few years researching it and writing a book about it. So I have the evidence to support this notion. If you want to create more innovation, if you want to reduce risk, if you want to grow your business or you want to have more efficient services in government, you got to put a focus on data. And I know we're going to get into the details a little bit later. Now, AI is tremendously exciting for cities because it gives our urban communities for the first time this remarkable new set of capabilities that we've never had before. Yes, we can say the same thing about the private sector, but you're asking me about the public sector. And cities now for the first time have to learn about what this is and how can it do things like make administration more efficient, you know, give people easier access to the wide range of services they, they consume every day. How we can make, how can we make cities safer and produce, you know, non carbon, more non carbon energy and facilitate that transition in a climate crisis. Right. So what I like about both of these topics, AI and data, is they are so applicable across so many different problems. They're not, they're not narrow, these are big technologies. So I think that's my opening comments for you and let's see where this goes.

Yeah, yeah. Well, I have a number of follow up questions and you raise an interesting point about city government versus a corporate setting where I have lots to say about topics that cover both of those settings. Right. Data governance being one. But also we're not going to have done a good job if we don't at some point talk about how large language models have changed all of this. But for the moment, let's stay with cities because I have a couple of follow up questions for you. You know, I was, I think I might have even been reading your book in an airport or something. I just remember, I don't remember the specifics, but I recall having a spontaneous conversation with somebody and they actually had kind of a negative connotation that they associated with smart cities. And I can tell you specific, you can even guess what they were thinking. But you know, if you like spy shows or detective shows, London for instance, is famous for having extensive surveillance network. I mean, they say that you basically can't walk from one part of London to another without at some point being caught on a camera. And for them that was the dominant thing that came to mind. It's not, wasn't my initial reaction certainly, but I think some people have that. Could you speak a bit to some of the other technologies? Because we're talking about the pressure plates under the road, we're talking about utilities, we're just an enormous diversity of things. What kind of data are you gathering as a, you know, a data professional in a city?

Yeah, and I totally recognize this point so deeply. I mean I've been the, I've received the, if you like the bad side of that. People think maybe that, you know, when they learn that I'm in smart cities and I'm a, you know, a global thought leader on this, that I'm all about surveillance and about, you know, not having privacy. And, and by the way, let's put the record straight. It couldn't be. For me, this is the opposite of who I am. I'm entirely about privacy and people's rights and human rights. So to me it would be such a complete tension for me to have a sort of, to be advocating for the things that I don't believe in. Right. So from a personal perspective, I very much want to see technology protect rights and create the right levels of comfort for people that they know that their information's been protected and their rights are being amplified. Right. So in terms of just character, it's important to recognize that and, and truth. You know, I get to travel and work with cities all over the world. This is a great, the pleasure of my life. I'm so honored and humbled to have the ability to be invited to speak with governments in, in, in Europe and in South America and Australia and beyond

and remind me, and I can't believe I forgot this. We're talking about cities all over the world. We would love to know where you're joining us from. I usually say that in the opening couple of minutes, but my synapse didn't fire. So please do add in the chat if you haven't been on a streamyard call before, you might even be joining us from furthers YouTube channel right now. But all of those, all those comments get consolidated. So please tell us where you're joining from and we'll be able to see that. So sorry to interrupt you, Jonathan, but I want to make sure.

No, no, not a problem at all. Thank you for that. I'm always interested in that too. So in. To continue this point as I work with governments on all sorts of projects, whether it's again, transportation, energy, digitalization, sustainability and more because all these topics are under the umbrella of making cities better, which, you know, using tech, that's really what we mean by smart cities. It, it's very rare that ever, you know, the question of surveillance, you know, or, or you know, these, these, these questions that people are, are very concerned about come up. I, I don't find myself being that exposed to that. There, there's a, there's a role for that within, you know, public safety. And you know, I think people understand, you know, in, in, you know, if you're, if you're, if you're by yourself in a remote area and let's say you've, it's late at night and you've, you've walked out of a store or onto a street and it's dark, you probably would appreciate that the lights might come on automatically to, to, to, to bring light to your environment. Turns out that's a huge deterrent for, for, for crime. And, and so, you know, equally where having a camera in place and again in a parking lot or street where you can deter or have evidence that you can bring to, you know, the follow up where there was a crime, unfortunately that this can be, this can be incredibly effective. But what we're really focused on, although there's a role for that and maybe our public sector, excuse me, our public safety colleagues can talk to very well, you know, you get to choose. If you said to me, jonathan, we're going to spend the next 45 minutes talking about the role of data in, you know, optimizing, you know, our road network or, you know, reducing congestion, we could do that. Similarly, if we wanted to talk about the role of data in, let's say, something like the energy transformation that, that's underway, we could, we could do that too. So I like to think of, you know, data being first of all that engine of change, you know, the oil of the 21st century, that's often cited. If you take raw data, you process it, you apply it, much like we did with oil in the 20th century that drove, you know, the prosperity and production of that century. In the 21st century, data is increasingly filling that role, right? Driving productivity, driving efficiencies. If we apply it in all of the, you know, across the domains of, of an urban environment, we can get very, very interesting solutions. I mean, let me give in a specific cities and city leaders, elected leaders, have always struggled with timely, good information to make decisions. I mean, somebody who works in the data space and has written books on this, it all comes down to very often data driven decision making, right? How can you have the right data at the right time to make a good decision? This is what every leader wants. You know, 80% of leaders want this. They want to be a data driven business. Only about 30% actually have that capability today. So there's a great gap between the desire to be driven in your actions and being able to make better decisions and what capability you actually have. So governments want more timely information. You're really what I noticed, having worked in the private sector and then in the public sector, is that gap between how much good decision making was being made based on good data. Now, we've come a long way over the last decade. No doubt cities have come incredible at distance, but there's some ways to go too. I would sit in a council meeting on a Monday evening where an important decision had to be made by council members. They want to do the right thing, they want to vote on, on a subject. And of course the question comes up, you know, what data do we have to support this? Or what they, what does the data say about, you know, xyz? And so often the response was we don't have it right now, but we can get it for you. Which means you're deferring the decision or you're making a decision that you don't have the full context and sufficient data to make, which we don't like. And so what I saw later was as cities began to recognize, and in my case really double down or really focus on having timely at your fingertips, good current data, we could bring up that information pretty quickly. So if a question came then what data do we have on this topic? The, the analyst who is making the presentation or the city manager could, you know, in just a moment say, well actually look on the screen here. We're showing, you know, this trend over time and, and now we're better, better informed. So governments want that they're getting better at it and that's changing the ability to, for us to make more timely but also more informed, better quality decisions.

I still have a couple of follow ups, but I want to make a quick observation which I'm sure that you'll agree with. You know, of course I'm a, I'm a data scientist, you know, so I love my machine learning and AI models. But listening to you and I'm sure this will come up again when we talk about data governance, I'm all. There's one thing about the analytics maturity curve that sometimes concerns me and that is that it sometimes makes the latest, hottest new thing, gen AI being among them now somehow more important than the foundation stuff. But when I'm working with a client, if they don't have a strong BI team, I don't have the raw material that I need to do the machine learning and the AI model. So just a take this moment to praise those business intelligence teams out there because without that solid, accurate data, you can't proceed to the next step with the data engineers and then eventually the data scientists and so on. And I'm sure you see that perhaps in particular in city government where possibly they're playing a bit of, you know, catch up with the latest technologies because they probably don't have the budget that big Silicon Valley corporates, you know, have. So, so I'll ask you to comment on that. But also I also want to mention something else that, that, that we've talked before about smart cities and you believe that cities will become actually more pedestrian friendly with some of the things that are going on in AI. So you can comment briefly on my observation about business intelligence. But I want to call you out on a quote from about three years ago and I warned you that I would be doing this. About three years ago we were having a conversation much like this and you said that somebody that was a newborn or a toddler at that time three years ago. So I suppose they would be about the age of my best friend's young one, about four years old now, that possibly she won't learn to drive a car. And I wondered if you still believe that to be the case and why you think that the cities of the future will probably have more pedestrian activity and that we, we have an opportunity to avoid congestion and make greener spaces. Something that you spend a lot of time on in the book.

Yeah. Thank you. Well, let me see if I can answer the first comments you made around, you know, city preparedness to really leverage the power of data. Well, look, it's night and day today versus what it was 10 years ago. I can say that for a fact. We didn't have data. People like a chief data officer In a city 10 years ago, maybe 15 years ago, barely had data analysts. You know, that was a part of somebody's role and they didn't have good tools. Maybe, maybe spreadsheet was the, was the best they could, they could do. And by the way, I love spreadsheets, as you probably do. Still, still amazing. So yeah, like we have to. Cities have invested in talent. Cities have bought better tools. They are now increasingly having data strategies. They have made data more central to their operations. So the journey has been fantastic. It does lag the private sector. No question, no question about that. We shouldn't be hide from that fact. So yeah, as long as cities continue to recognize the value here and make it a priority, I think they're gonna, they're gonna make progress on it. When I think of my book, when I wrote, wrote about, you know, this topic, how, how can organizations take advantage of, of data? First of all, there was just sort of a lot of people who didn't recognize it as a major resource. And this was amplified by the way, in government. I mean, when you run a city specifically, you are creating data, you're collecting data, you're storing data at high volume, you know, in large amounts at high volume. In a Lot of variety and it's just a product of operations. What, what cities didn't do say, well, wow, we have this incredible resource, we'll just call it that. We have all this data. What can we do with it to better inform and better drive our efforts? And that's a mental sort of, you know, hop you've got to take, you got to go from you just data is part of what it means to run a business. To data now is a strategic asset for our business or our city. And so what I saw when I was helping to run a city was, is, was exactly that in the early days is no recognition that we're just sitting on this incredible abundant resource in a world by way of incredible constraints, which is what government is about. And we weren't doing anything about it. So we had to turn that corner and create those things. So the advice, I guess for technologists who work in government or vendors who serve government is there's still a massive opportunity there and we're making progress, but we got to, we got to kick it up a notch. Let me just be real, real clear on that. Now, the second part of your question. Yeah, it was about three years ago when I said, you know, a child born today will probably, there's a good chance they on average won't have to learn to drive. So I was making sort of a, assuming, let's say, person starts to drive at 17 or 18 years old. So about sort of an 18 year projection. And you're saying we're about threes and three years into it, so about 15 years to go. I still feel strongly that this is a good projection and I'm more convinced of it based on what's happening here in San Francisco. We now have a service, the Waymo service, that is completely autonomous. You know, even for somebody like myself who's deep in this topic, it's still a little spooky to go to San Francisco. I was there downtown yesterday actually for the whole day. And I had to go to several different meetings as I went to different meetings. You see a lot of these cars now, I don't know what the number is, but it's high. I mean, they're doing hundreds, if not thousands of rides now per month. And it's just the creepiest thing to see these white SUVs with people in the back and no driver. And it's just, they're driving around confidently. People used to say, well, yeah, the cars will drive in areas that are where there's, it's not too complicated or where there's no road work whatever. Those days are gone. And that was only like, that's only two or three years now. The. You can summon one of these Waymos, pick you up and drop you anywhere in the, in the city of San Francisco, we've just now had permission to expand that into the South Bay. They can't go on the highway, I'll say that. We still have some hops to go and, but they can come right down into the Peninsula here, into all, like Cupertino and Palo Alto and Menlo park and all the, all the cities you've heard about. So that's coming soon. Which means the, the, the Cape. The opportunity, for example, bringing people to the airport. I mean, that's a very popular thing here as it is in many cities now. You're, you're creating even more opportunity. The way the Waymo sort of leader CEO said it recently was, you know, use that expression, things go slow and then suddenly fast. Right. I know I messed that up, but that, you get the idea. It's like we're kind of going along and then suddenly this is a big deal and the number of autonomous trips is escalating very, very fast. Now you compound that with the fact. And you can just ask the car manufacturers this question. Who's buying the cars? Right. Who are buying cars today? Well, there's a whole demographic, right, that is, is not buying cars in any volume like they did just, you know, one generation ago. The, the, the, the new generation or generations, we've, we've several in the, in the population right now. As you know, they are much more inclined to live in an urban setting. They are going to do things like bicycles. Bicycles is a huge global phenomena. I consider bicycles part of the smart city movement. Light rail and walking.

To clarify. Yeah, something like city bikes in New York would be the kind of thing that you're speaking of. We're not necessarily talking about somebody that leaves home in a bike and gets to work that way, although possibly it could include that. It would also include somebody that maybe it's taking a subway and then biking six blocks from the subway station of the office. Those kinds of things.

Right. A big part of it, however, it does now include people buying their own bikes. Bikes is a, is a growth business, you know, and certainly, you know, I spend a lot of time in Europe in, in many major and, and secondary and third type cities and they're, they're dense on, on, on bikes. A great example of this, again is, if you want to understand it, look at the last decade In Paris. I mean, what an incredible transformation. Not, not a straight line, bumpy. You know, there's people on both sides of the, the argument there. But on balance, you know, a city that thought, you know, a decade ago that Parisians would never, would never bike because it wasn't seen as Parisian today have embraced fully bicycles. So you have not only the emergence of autonomous vehicles and that's going to next year, the year after by time we get to 20, 30, only know, five years away, that's accelerating in, in cities all over the world, all sorts of urban environments. And you have behaviors changing and you're seeing definitely the, the, the configuration, if you like, of transportation options completely change. I want to add one more point. I have to do justice to my friend and scientist Carlos Marino, who, who wrote the book and talks about and created the idea of the 15 Minute City. And I'll say straight off, you know, this is not about issues with restricting human movement or about, you know, surveillance culture. You know, if, you know Carlos, he's all about, you know, human rights and, and respecting privacy and all that, and that was never in his mind. I say that because this comes up a lot. People say, oh, Jonathan's talking about the 15 Minute City. This is, we're all going to be walled into little ghettos in our cities and this is going to be total control. It's the opposite. Nothing to do with that. And what 50 minute cities basically say is, you know, if you think about American cities and you, you know, you go to the mall, you go to a dentist, whatever, sometimes we're traveling long distances. If we put all the basic services close to where people are and we kind of replicate that, people don't have to move as much. They don't need to get into their car, they can walk, they can bicycle, they can get light rail or use scooters. Right? So yes, we are seeing the, this march towards pedestrianization and maybe to be completely. Well to, to, to kind of complete the, the little narrative on this is the amount of cities who are closing off areas to vehicles. Right? You've decided that we want to have a. Just a area where, where, where only people and bikes can go. And two quick examples of that. One would be New York City. You know, you're seeing Times Square become largely pedestrianized now. It's changed the dynamic entirely. You know, everything from much lower accidents.

We've.

They've reduced well over 30, 40% now of accidents that are vehicular in nature. The numbers are probably higher even today. This is a little bit of an old stat, better quality air, it's more, it's quieter. So there's lots of these incredible advantages. It also means that the city can put more focus on managing traffic around that area. And unintuitively, in fact, in New York City, the traffic improved when they stopped, you know, a lot of traffic in Times Square. And the other one, although I mentioned it, it's worth mentioning again because it's such an important city, is the city of Paris, right, Looking at major thoroughfares like the Champs Elysees and not entirely blocking its traffic, but making, you know, cars and motorcycles, for example, a very, very small part of that street. Putting the focus on trees, on people, on, on, on more natural aspects. And that's a multi year, multimillion dollar project that ultimately, over some period of time, I think it's about a decade or so, will transform the very nature of the center of Paris being largely pedestrianized and, and more beautiful, more quiet, more, you know, healthier and, and so those were two big examples. But we're seeing this replicated on every continent, almost every continent on the planet.

I'm going to want to switch gears to data governance now, but let me see if I can't attempt to tie in together what we were just talking about, because folks might be hearing pedestrians and city bike and things like that and say, and not quite making the connection with AI, but if you reflect on everything that you've just said, what you're basically saying is if technologies like Waymo, not only Waymo, certainly over time there will be competitors and so on. But if autonomous vehicles succeed, they're more readily available and they're cheaper, right? Then if you don't need your car as often, then you'll borrow your car rather than buying it because it's only an app away and you'll be pleasantly shopping and you're pedestrian only neighborhood. And then when you have to go home with the shopping bags or for whatever reason you need a little bit of help with your transportation, you would call the car and it would drop you at home, but at home you might very well not have a car in the garage. Did I get pretty much get where you were headed?

You did. You summarize it really nicely. It's already happening. And so let me quantify it because, you know, you're a data scientist and your audience probably likes this too. Let me quantify this. These are very, you know, generalized numbers. But basically when you own a car today in the United States, and you know, this can apply in other countries too, but you know, just bear in mind that the numbers are probably slightly different. If you own a car in the United States, it costs about a dollar per mile to, to drive your car. So I always kind of help my students think about this. If you got in your car to get, you know, a, a, let's say milk or, or some butter, right? And you know, depending on, let's say it was $5 for the product you were going to buy and you drove seven miles there and seven miles back, you spent 14 mile, $14 getting a five dollar product. And so you've actually spent $19 in whatever, 30 minutes. So we, we don't necessarily think about ownership like this. Why does it cost a dollar per mile, by the way? Well, gas, you know, maintenance, insurance, you know, wear and tear. So, you know, bake that together and you're basically paying that much. It's probably now even a little bit more. So if you could. So let's say you're going to get that gallon of milk again or 2 gallon, whatever it is, and you got a, a summoned, you know, autonomous vehicle to take you, you know, and you paid, you know, it has to be just less than $14 to make it make sense, right? And, and so in a yellow cab

years ago, before Uber and Lyft, you almost certainly would have been paying more than a dollar a mile, right?

Yeah, well, you won't.

So if you're saying is at some point we're gonna, we're gonna cross that threshold and we'll just naturally gravitate towards these other options and that makes a lot of sense.

It'll be cheaper. The last point I just wanted. Absolutely right. And I'll just add to this. When the founders of Uber were thinking about what they were going to do, and of course Uber and all its competitors are, they're a phenomena how they've been transformational, largely good, but again, I want to recognize the issues, but largely, you know, very effective in, in their service. They're still expensive. Even though in many places in the world, Uber is cheaper than a taxi. It's still expensive. I mean, it's still considered a premium service. Why? Because there's a human that has to drive the vehicle and that human needs a salary and it needs, that human needs, you know, some basic services. And, and so that adds sort of on top of the, you know, premium that Uber has to charge. So they make a profit and they can run a business. If you take the human out of that equation, the numbers drop very quickly and now it becomes the, the calculus is Entirely different. Right, where you might have said, well, I'm not going to get an Uber because it's expensive. Yeah, let's see if I can get a friend or maybe even drive myself and park my car. You know, the Uber, Uber equivalent is going to be significantly less. And, and you'll, you'll definitely take it because now it's only a matter of the electronics and the management of those vehicles. There's no, there's no human part to it.

Yeah, that, that for me is the fascinating, very pleasant irony of this is that AI is helping make cities more pedestrian friendly and more green. I just think that's, that's fascinating. But we have so much to talk about. We're at the halfway point, everybody. So I would challenge you that if you're not following further yet, open up another window and just do it now because otherwise, otherwise you might miss out on next month. And actually I'd like to announce that we actually have a speaker and a topic for next month. On October 16th we're going to be talking about MLOps with Kristin Carrer and she's going to be a fabulous guest. If you're not familiar with her, there's going to be a newsletter about her. And thank you for the lovely compliments, Jonathan, on the newsletter issue that we dedicated to you. So if you don't know that we have a LinkedIn newsletter now, further has other newsletters that you can access on the website and I urge you to look into those. But we have a LinkedIn newsletter and the LinkedIn newsletter is really about these thought leadership interviews. So we did, I thought, a very interesting article about Jonathan's work, including some of the topics we've talked about so far today and we'll have a follow up. Jonathan, I think you know this, but we're gonna, we're gonna take what we talked about today and summarize it in text. Right. So we're going to have a follow up newsletter as well. And then we're going to have a newsletter issue about Kristen that you can read before her interview next month on the 16th. So if you're not doing all that yet, please do it right now before you forget. So on to data governance, I want to refer to something that you mentioned in your book. And by the way, we keep on mentioning your book, but people are probably wondering just how big is this book with all these diverse topics? It's books, plural. Right. But Data Governance is the book that we're transitioning to now. And you also have a bunch of courses on LinkedIn learning, which if that's, if someone is a visual learner, then they should certainly seek that out. But from the book itself, and I thought this was fascinating, but it certainly resonated with me that even though everybody knows they need data governance, when you sit down and I'm talking specifically about you, but I've had similar experiences when you sit down with a customer and say, let's get our act together and really improve the job you're doing with data governance, that it's sometimes a tough sell because they know it's a risk, but maybe they just figure that it's not as important as other things anyway. Sometimes it's hard to get folks to focus and actually sit down to do it. And certainly our clients are facing some of the same things. I guess it's like spring cleaning. I don't know, maybe just psychologically, folks want to procrastinate on data governance. Kind of help us, help us make the case. Why do organizations have to take it more seriously? And maybe why are they not perceiving it the way that, you know, they could perceive it in a different way and it would make it more compelling, more urgent.

Yeah. Now, so happy that you've brought this up and, and we get to talk about it there. This is, you know, even though data governance has existed for some time, I almost feel it. I perceive it and others do as a new topic. And I think that's a good news story because that reflects the fact that it's getting more attention and more people are, are focusing on it and there are more vendors in the space, there's more tools. But at the core of the issue in my research, and I spent about a year, over a year talking to every type of industry around the world, is just a lack of understanding of what it is. Right. And why they need it. I think a lot of leaders understand data to some degree, right. They understand data strategy and data management. But you throw into the mix data governance and they're like, wait a minute, what, what is that part? You know, what does that part mean? And, and so maybe because this deserves a slightly longer answer, what I discovered in my work as I was focusing on data governance is there was an even bigger topic that wasn't getting enough attention. And this is the topic of data culture. Right? When I, when I think of when, when we should think of data culture, you're, you're really talking about, you know, culture is about behavior every day, how does a company behave every day, what does it do? That's its culture, right. And Sort of how does it behave relative to data? And so if you think of like a umbrella of data culture, what's within data culture is data management, right? Having the tools, data strategy, having a journey to go on. And then that third leg is data governance. How do you ensure you're taking advantage of data to do all the things you want to do in your business? And how are you managing risk? You know, so you don't, data doesn't create more problems than you know, you would like. So there's, there's, there's a sort of questions about first of what is this thing? And then what are you talking about when, when you say data culture? And if I ask a leader, can you, you know, give me your sense of what kind of data culture you have in your business? They didn't even know how to answer that question. Some of the ways you can answer that question is what priority does data take in everyday conversations? You know, what, what priority does it take in meetings? Do you have access to good quality, you know, data to make decisions when you need it? You know, is the quality of your, of your data within your organization where it needs to be? Right? Is it kept current? Are there gaps in it? Right? And then when you do things like pursue initiatives, are you ensuring compliance? Like you're in a healthcare business, you better be sure that you're managing all the aspects of data, data compliance and regulations. And so, so that's, that's an important thing. So then next point is real, real brief. So okay, yeah, I get to speak with a, with a leader says, okay, I, I can see why, you know, fostering a data culture can be good for us and I can see the role of, of, of data, data governance, but it's, it's going to be expensive and maybe we don't have the talent. Right? So then you get into those questions and, and so, well, look, in any organizational investment, if the, you know, if the product, the output is greater than the, the cost, well that seems like a good decision to make. And you have to be able to quantify that for, for a business. And then the, the talent and tools, well, they're coming. I mean we, we are and I'm a professor too, so we are generating more data scientists. We, we have had a shortage and, but that we're getting to a better place. More universities are offering programs, there's more certificates now. So we need just to acquire the talent. It doesn't mean necessarily hiring like net new people, but retraining people you have, that's always a Good management decision and then the tool sets. Right. Do we have tools for data governance automation to help us manage our data Data governance programs? And we see now clearly that there are more participants. And by the way, when there are more vendors in a space, it means that there's more demand, which means it's growing. Right. And over the next 18 to 24 months, data governance and the tool sets and the demand are going to go, are projected to go way, way, way up. So you really have to gauge in this sort of full, all hands on deck story and engagement with, with leaders. And the last point I would make is, okay, you might have a reluctant C suite or, you know, reluctant leaders that you've convinced me, let's try something with data governance. So yeah, you focus on maybe one small but important data set in supporting one small but important process in your business. Focus on that. What do you want to see as a better outcome? When you apply, you know, policies and, and, and talent and tools and you measure that and then when you see success, you communicate that success. Right? And, and this is again a gap. I see people, you know, they throw money at it and we, well, a year later, did anything happen? Well, we should be communicating that more often. You have to do that. You have to say the result of, you know, what we did with this particular, if you like, experiment with data governance is we were able to identify or convert a customer, convert a new buyer, you know, from what was taking 40 days now into 25 days. And you're like, oh, I see results. I actually see results. And that's really key here.

Oh, got a fun comment here. Retraining. Yes, please save.

Yes.

Okay, you know what, please go ahead and.

No, no, I want to hear what you. But I have to, I have to say this is something, I've been in the workforce over 30 years and I see this all the time. Leaders expect people to do things they haven't been trained for. And I just wanted to amplify Rosalind's point there. We need to give people the training and tools to be successful. I'm telling that to all leaders right now. We need to do that, please.

Keith and I want to thank Rosalind as well. And then we'll use the price of My Face Light Up. So I'll make my comment. But first, a quick bit of logistics. I'm going to start taking audience questions within the next five minutes or so. Please start populating the comments with those questions. And thankfully, and I couldn't be more pleased to tell you this, Jonathan and I have helped kind of organizing those questions. So don't worry about us. We'll keep talking about data governance, but start to queue up your questions and we will be talking about large language models before the end, so feel free to ask your questions about that. So the reason I lit up with Rosalind's comment is there's a thought leader that you may be familiar with, Joe Reese, who's written a very influential book on data engineering, and he and his co author described themselves as recovering data scientists. And it's kind of a joke, but of course, like with any joke, there's a lot of meaning in that. They found that if they went in as data scientists to work on engagement where there wasn't a good data culture or good data quality, they couldn't get things off the ground. So on a personal note here at Further, the little bit of a detail, many of us on the data science and AI team joined earlier this year, which is a whole separate, interesting story. But one of the reasons that we love being part of our family at Further is that the data and cloud engineering team is about twice the size of the data science team. And I think that's as it should be because as data scientists, we find ourselves on an engagement where data quality is an issue. We have numerous other practices within Further that can help us with that engagement, because when you're a data scientist, as a solopreneur, which I've been at various points in my career, if I go into a dysfunctional data governance situation, I might even have to tell the client, I'll see you next year. I'm not going to be able to do my job this year again. We're thrilled that at Further, we don't have that challenge because we have diverse skill sets and we have a large team. Let me, let me challenge you with something that might sometimes come up in conversation. Chatbots are one of the hot things this year, and we have a large language model team. We Even have a PhD in computational linguistics that serves on that team. So we have tremendous talent in that area. So if a client needs us to work on chatbots, we can work on chatbots. But having said that, there might be somebody in the C suite that's saying, oh, Keith and Jonathan, we care about data governance too, but if we could just get that chatbot done this year, let's do data governance next year, because the chatbot will presumably save or make us a lot of money and then we can deal with the data governance. It could be that you really can't get that benefit and postpone the data governance. And I don't know to what degree you relate to that scenario, but maybe you could speak to that exact thing. We'll push data governance into next year because there's something super exciting that we want to do now.

Yeah, yeah. This is real. This is real. I love how you set that up. If I think about this dilemma, you know, in a way you're kind of, you're saying, do we have to choose between successful AI deployments and data governance? I, I think the commonality there is data quality. Right. I does feel like that, you know, garbage in, garbage out. If you, if you're pointing, if you're generating an LNM from your internal data and you're pointing your chatbot at that and it's bad data, well, that's a bad business decision. There's reputational issues, there's brand issues, and there's a whole lot of risk associated with it. Funny that you should bring this up. This is in fact a actual engagement I'm working on. I won't give out any more details to respect. But you know, if you make the decision that we're going to put on our service a chatbot and they want to do that, but we're not sure that the documentation that we have that, you know, is on our servers for the last 15 years is good still or current, whatever, you know, you're kind of setting yourself up for maybe problems that you'd rather not have. So delaying that chatbot project might be, might be a good, good idea, you know, so, but look, leaders don't want to have to, you know, make those hard choices. We, we want good data and we want a chatbot, right? We want great AI and we want to move forward. And, and, and we, we, we, we, if we have to do data governance, let, let's do it. Look, one of the biggest items of recommendations I make to organizations who have recognized the value of data governance but, but are not acting on it because they think it's too big, too expensive. And I know this is kind of a low ball thing, but it's basically start small, do something right. So if you're, you're building out a chatbot, you know, start small, point it, point it at data that is in good quality that you have some governance around that's low risk. And now you're gaining a little bit of both. You might not get the big bang you want on day one. And depending on your business, I mean, if you're a big insurance company, you know, where people are relying on those conversations or a bank, you know, a financial organization small is not going to make it. You're going to have to do the investment in, in, in both. So, so, so there I think, I think that's a, that's a key area. You got, look, you got to have good data sources to have good, you know, output from, from these all, all, all types of generative AI products, particularly when you're pointing them at your own, you know, own data rather than ingesting from, you know, the big, large language models that are part of, you know, the big tech firms. Right, right now. So make that choice. I think the conclusion might be spend a bit of time on the, on the, on the quality data aspect as you begin to strategize and build out your plan for Chatbot. It might mean you don't, you can't deploy it next week, but you'll be in much better shape when you do eventually deploy. Deploy it.

You know, two things come to mind. One is I want to transition to analytics leadership, but staying on this theme of data governance, but getting into the Alphabet soup of CDO and CAO and so on. So that's where we're headed. But I want to make a quick observation first that you were talking quite a bit in the last few minutes about scaling up and scaling down. And I imagine that you'll agree with me here, there's a bit of an art to scaling up and scaling down. You know, if, if a client has an ambitious goal and you work it out and it's going to be, it's going to take a certain number of months and a certain number of dollars and they go, oh gosh, if that's the case, let's break it up into pieces. Let's do a certain amount now and a certain amount in the next calendar year. I think a lot of potential clients underestimate how tricky that is. You know, because there are certain things you can't do half of and there are other things where sometimes you can do a first half and a second half. Do you find that sometimes that clients come to you too late in the game because they basically have their list of to do's and they just kind of cut off like the first half. But it could be that without that the necessary skill sets and there's a lot of them, the data engineering and the data science and so on, that they've made a guess as to how to cut a project in half. But it's actually really hard to scale a project down. Sometimes there's a Lot of thought that has to go into it. Do you find that as well?

Oh, sure. Look, with 30 years of technology experience, one thing I can say for sure is we don't always get project management done well. At one point the failure rate of a project was well into the 70%, you know, number. And I mean it doesn't give anybody comfort. And, and that includes things like what you, what you're saying it's insufficiently scoped or it's scoped ambitiously or there's just, you know, the what you thought was going to happen versus what you planned on did not align. I'm sort of working a lot with government and I do work with private sector by the way, but I, a lot of my clients are government. I can say that we're, we're early in this, particularly as we start to plan around, you know, our own in house lnms. Right. So I haven't personally experienced this question, although from listening to you I can fully appreciate the, that it is an issue with, with a lot of organizations. You're deeper in it than, than I am. Yeah. And sort of, you know, again, I like to simplify things a lot and you know, can you, can you do data governance on half a data set? No.

Well, that's.

You're getting at.

Exactly. The issue that we're, we're talking about is that sometimes you have to do all the data governance. But you could because actually with the chatbot piece, and this is the kind of conversation that I have with clients all the time, is you might narrow the scope of the chatbot, that it's not going to cover all of your departments. You might roll it out in one department first. Yes, but. Or that it's restricted to certain queries. Certain queries go to the chatbot and certain queries go to a human, that kind of thing. But as you point out, you can't clean half the data set. I'll just clean the first half and clean the second half later. Half a clean data set is not a clean data set. You can't do the top rows and not do the bottom rows or something. Something crazy like that. We had a question actually that came in a little bit earlier that we didn't get around to, that our colleagues just alerted us to. So I'll go ahead and ask, and it may be pop up on screen here in a second too. How will the AU's AI act have an impact on ongoing innovation in AI and ML? And maybe you could also, although the, the person who asked didn't Tie in data governance specifically. I think we can talk about both of those AI and ML projects as well as the data governance that would underlie it.

Yeah, yeah. Well, I love this question. I'm going to give, start with, to the practical answer, like just what does it mean? And then second, what's my opinion? Because they're, they're not necessarily the same thing. So, you know, the, the European Union Artificial Intelligence act, which is passed and now in action is the boldest, most ambitious AI act that's ever been implemented. They had to do, they felt they had to do something. They wrote this thing up, they got it approved, it's now enforced and it does things like it does say they're, you know, it puts AI activities into different categories. So you have what's called extremely high risk, you know, high risk, less risk and without risk. And if you are hitting things like high risk, then there's a new level of attention it requires. For example, you might have to register that service. You know, if it has anything to do with, you know, facial recognition and it's, and it has a particular context. There is a EU system that you need to register with. You may need to get oversight on that. Things like manipulating people is prohibited. You know, some of these things, by the way, only relate to certain size organizations. For example, one level is if it costs over a hundred million dollars to implement and train your LLM, then you are subject to certain levels of scrutiny and adherence. Okay, so that gives you a certain feel about it and I think it's well intentioned and some of it makes a lot of sense. Now by the way, just to be complete on this, I'm in California here where there's a lot of activity on the regulation front. In fact, there's one law that's moving forward that it's got some, you know, legs to it. Elon Musk has sort of said he's going to support it, whereas Mark Zuckerberg is against it. You know, Google is against it. And you know, it sort of, it, it does reflect some level of the same characteristics that the EU regulation does. And I just want to correct something I said, I said about the 100 million that is on the California side, not the European Union. My apologies, I actually got that incorrect there. And, and so, you know, again it's, it's at one level it's kind of lightweight like your eu, excuse me, your AI shouldn't discriminate in its behavior. You know, one human versus another human seems pretty straightforward. Okay, so that's sort of the Nature of these, why they're happening. There's a lot of activity around the world on the regulatory front. I mean, there's people who believe AI to be so scary that we need to do something that if we're wait, if we wait now, we're in trouble later. So they're usually well intentioned. My personal opinion is this is very risky. We may do something here regulatory wise, which we regret because I don't know sometimes that we fully understand the extent of the decisions we're making. What we don't want to do is create limitations to innovation artificially. Right. Well, the American economy and so many economies in the world are powered by innovation. And we live in an era. The 21st century is about progress. It's about, you know, quite accelerated movement into the future, often producing better services and, and features in our cities which I think we all would, would appreciate. So I think that the riskiness makes me a little bit us, you know, kind of reserved about any of this regulation. I have to say I haven't read every word of every proposed or passed acts, you know, recently and I feel like I probably should understand it better. So I respect anybody who's maybe screaming at the screen right now saying, well, you should read it and know that it's not that risky. But I don't know, you know, reading, understanding the analysis, analysts approach and hearing different perspectives that I have heard that reduces or eliminates fully my concern that there's some risk here. So the best case scenario is it does protect people and, and we do the right thing. The worst case scenario, which is more, you know, has a, I think greater likelihood. It does stifle innovation, as the question asks, and even causes us, you know, competitive problems, you know, versus China or you know, Europe versus the US or California versus New York. You know, what I'm seeing right now, for example, in the US is that each state is coming up with their own AI regulation. Colorado has AI regulation that's passed now. California is about to pass it. It looks that way.

Well, it's becoming almost like a full time job to keep up with that piece. So we actually have at further, we have two individuals that have joined our team. Their titles are AI Strategist. And this is one of many topics that they follow closely so they focus in particular on that aspect of it. And we just had like 10 people, you know, go through our training on this, but it's really difficult to keep up. And the fragmenting that you describe is going to make that a challenge for sure. I want to tell people that they have their last chance to possibly ask a question. I'm not even sure we'd have time for it, but very important for everybody to know, including a reminder to you, Jonathan, that the beautiful thing about LinkedIn Live is it lives on in the feed. So sometimes the most interesting comments and questions are in the live feed after a talk. So I don't know if you have a few minutes free after we hang up here, but I always make an effort to do that. But sometimes there's amazing comments. Like 8 hours from now somebody will write a comment that's maybe didn't catch us when we started that lives on the west coast or even further west than that. But I have one final question for you. It's a huge question, so it could end up end up being addressed in the chat. The Alphabet soup of Chief Analytics Officer, Chief Data Officer, Chief analytics and Data Officer. Now I tend to see them as different and maybe data reporting up through it because I think of that as more data governance and CAO may be reporting up to strategy, for instance, since it's really about ROI and some of these other things. But yet there are one person who's both Chief Data and Analytics officer and now there's caio. And our AI strategists, for instance, are not unlike a fractional C AIO actually. So just a final comment on a huge topic. What structure do you find works best?

It's all over the place, isn't it? That's what we see. Look, I'll go out on a limb here and say I really don't think there should be a caio. Like you shouldn't have a chief level role on a tech. You know, that's a role best, you know, underneath the strategy person or under the Chief Technology Officer. I find that that role will exist in some organizations for some time, but it'll, it'll get like away or it'll be absorbed into, into another role. The, the. I think the most important point I can make here because you, I really have like 30 seconds to say it is the hierarchy should represent what your priorities are. If you consider technology and data to be sort of a back office overhead, you're probably going to report it into the cfo. If you consider things like data and, and data in this case to be of strategic importance. And I hope you do. The reporting line should be into the, the Chief Strategy Officer or even directly into the CEO. So I think that's one distinction, you know, maybe a straightforward one that, that I, that I would make. There's Lots of nuance.

That's the cdo. You're saying the CDO you would suggest not report to the cio, but possibly to strategy or cto.

Yeah, yeah. Depending on the degree of, of how central it is to your market ambitions. You know, it's a, it's a such an important question, though. Your structure is vital and that's why you brought it up. I, I think you almost need to look at organization by organization. And so I'm reluctant to sort of say this is the way you should do it. So, you know, let me say this

then, because we're already a minute late. I will challenge everybody to put a comment and question for Jonathan and I because again, we're going to sign off here in just a moment, but the conversation lives on in LinkedIn and my colleague Julia. And thank you so much, Julia, that we actually have a real focus on privacy and some of these issues that further. And we actually have a newsletter dedicated specifically to that subject. So there's a URL there. But again, this lives on in LinkedIn, which is fabulous. So I challenge you to comment a bit on how you structure CDO and CAO at your organization. Do you have a cdo? Do you have a cao? Okay, and thank you for holding up the Smart Cities, but you really have a whole chapter dedicated to this structure of business and cd you have a whole chapter on CDO specifically and who they should report to and so on in your data governance book. So I invite people to see that. So let's continue the conversation asynchronously on LinkedIn, everybody. And I want to thank Jonathan. I want to encourage you again to follow further so that you can see us next time on October 16th. And on October 22nd we have one of our bi monthly webinars, so be looking for that and lots of stuff going on. So the best way to keep up with it once more is the Further page of Follow Further and the Further newsletter. In fact, be looking for a newsletter issue about today's interview where we address some of those topics in writing. Thanks again, Jonathan. I had a lot of fun and apologies everybody for a couple of minutes, but we were very enthusiastic about our topic today and I hope the conversation continues. Thanks everybody.

Thank you, Keith. Awesome.