CIO Playbook for Enterprise AI | CXOTalk #810

CIO Playbook for Enterprise AI | CXOTalk #810

Show Video

Today on Episode #810 of CXOTalk,  we're discussing enterprise AI with   two of the top CIO advisors in the business.  Welcome to Tim Crawford and Isaac Sacolick.  Boards and CEOs expect CIOs to have a strategy,  to have a defined vision in terms of where   the organization can take advantage of these  technologies. There's a lot of pressure top-down,   but there's also a lot of bottom-up activity, and  that's both people who are very excited about what   they can do with AI and spending a lot of their  time testing out these different tools, and those   that are really fearful, scared about their jobs.  What are they going to be doing as a marketer   or a software developer, as prompting becomes  kind of commonplace in their technology acumen?  What should CIOs be doing in terms  of championing these initiatives   and bringing the pieces together? That's the $64,000 question that   every CIO I work with is asking. Where does  it fit in? Where does it not fit in? How do   we start to navigate some of the things that  Isaac had mentioned starting with just the   innovation but not ignoring the people element  to it as well because there is a lot of concern   about job loss and "this is going to replace me." Most of it inappropriately thinking along those   lines. I don't think it'll have as dramatic  an impact as people might think. But let's  

not forget; we have a lot of history that shows  that there's change that comes with innovation.  The real issue for CIOs is how to navigate  through that to ensure that you are embracing   generative AI and AI. Granted, AI has been around  for decades. Generative AI, a little more recent.   But at the same time, making sure that you  put up the guardrails, where appropriate,   and that you don't over-rotate, which I think  we'll talk about what that really means.  Then, at the same time, be a leader. Be  a leader to your organization both up the  

chain of command as well as parallel to your  position and within your own organization.   Help them understand and educate them  on what those opportunities are today.  Frankly, that'll change over  time. As we get to know these  

technologies more and more, it will change. We have to be adaptable. We have to be at   the front end of the line and help guide  that process as leaders. That's probably   the single biggest opportunity for CIOs is to be  at the front of the line and help provide that   guidance in a balanced way. I agree with you, Tim.  You think about every introduction  of a new technology, analytics,   or AI capability now. We're moving up stack.  We're taking things that we do today. We figure  

out a workflow around it, a way of working. All of a sudden, a new technology comes into   play and says, "You know what? I don't  need to do keyword searching anymore.   I'm doing prompting. And I can continue to  re-prompt," and that's changing the workflow.   It's changing the skill set of people and  how they're interacting with these tools.  That breeds some who just lose a ton of their  time because they're just experimenting and   playing with the latest gadget that's come out.  Others are just fearful about what this all means.  "I'm a developer. I know how to code.  What's Copilot going to mean for me? I  

already know how to do development very well." The same thing on the marketing end in terms   of building campaigns or writing content. That's what's happening in the trenches,   and when you talk to what CIOs should be  doing, the first thing that I think about   is coming up with the types of problems  that you want people to be focused on.   Where is this really going to have impact  in your organization? How do you register   the types of value that a particular AI is  bringing to an operation? How do expose that?  How do you make sure that when  people are using a particular tool,   they're using data in a sanctioned way? They're  using the right tools with the right data?  That's the leadership role, bringing it all  together and saying, "We're focused here. Here   is our short-term vision. Here is the strategy  around it. Here is the registry that we want  

to use to capture ideas. And here is how you  provide feedback to us about what's working."  How is this different than any other  technology initiative? What Isaac was   just describing sounds like plain vanilla  leadership for technology. We get lost in   the details. We forget the big picture. What's  our strategy? How is any of this different?  If you look through the big tectonic shifts in  technology, whether it was going from mainframes   to distributed computing or the Internet  or cloud, all of those kind of sat behind   the scenes within IT, at least initially,  and then eventually would get out to the   public domain. One of the things we've seen with  generative AI is it started in the public domain,   and we have ChatGPT to thank for that. Now, everybody and their brother has access,  

and most people today, the average  person has used ChatGPT in some way,   shape, or form. They've been exposed to this new  innovation and seen what it potentially could do.  Now, that's coming from a position of just  experimentation. But that does filter into   driving what enterprises then have to look  at and have to think about and have to do. 

Meaning, I used it to build this content.  Isaac talked about marketing. Content   creation is one of the three big areas  that generative AI is being used for today.  The content creation piece is what common folks  are using generative AI for. Then they think,   "Okay, well, I can do this in my personal life.  How do I do it in my professional life?" to make   that end of the spectrum easier to manipulate  and work through. You have that coming about,   which I think has just accelerated the adoption  and the interest in generative AI specifically. 

Some have said it's having its iPhone moment  because we saw that when the iPhone came out,   and there was just this mass adoption because  people realized, "Oh, my gosh. I can take these   20 or 30 different functions, and it's  all within one device now." Generative   AI is kind of having that same moment. But we're going to learn. We're going to   learn where it does fit and where it doesn't  fit, as we go through the upcoming months.  We're now seeing more companies looking at  how to leverage large language models behind   the firewall, looking at their information,  starting to look at everything from textbooks   in higher ed to financial information that's  coming in the form of news and other content,   and saying, "How do we make people smarter with  this content?" and also use the capability of   natural language querying and prompting so  that that skillset goes down, that we don't   necessarily have to have experts in every field. We have to have people who have to know how to   ask the right question and being able to  challenge and interpret results. I think  

that's the bigger challenge here when you look  at AI versus some of the other technologies.  Everyone that's listening to this has heard  the term "hallucinations." Generative AI will   hallucinate. It'll make things up. I had someone just recently say,   "Generative AI and ChatGPT can be the  best liar because it builds confidence."  There's actually a story that's out there  now talking about an attorney that used   generative AI to create a brief and then  kind of double-down on that. Generative AI   ended up making up the cases [laughter] and  the details in the cases that didn't exist. 

Is that a bad thing? Well, maybe. But maybe  not because it's optimized for confidence,   and so building toward confidence. But as Isaac said, there is this big   question mark as to when you do a search,  you can tie it back to the source. And so,  

provenance of data and trust in data is  really important. But with generative AI,   because of the nature of how it works,  there's this breaking point within the   algorithm itself and how it works that we have  to kind of get over and start to understand.  We saw this in cloud with ephemeral workloads,  workloads that come and go, instantaneously   almost. And so, I think we have to go through  the same type of process of learning a different   way to gain confidence in results in data, and we  haven't gotten there yet. We've got a ways to go.  Please subscribe to our newsletter and  subscribe to our YouTube channel. Just   go to We have amazing shows coming  up, live shows where you can ask your questions,  

and truly, genuinely, you are part of the  conversation. So, check out  Where are CIOs in terms of understanding the  opportunities, as well as the various risks,   as well as the learning curve that Tim was  just talking about? Overall, what's the   state of the CIO when it comes to these issues? There's a lot of learning happening at too slow of   a pace is how I would categorize this. Technology  and the capabilities are moving too fast.  The employees are experimenting, and the CEOs  are trying to nudge the organization to be at   the forefront, sensing that it's going to change  business models, it's going to change customer   experiences. They're hoping that it makes their  workflow and their employee environments a little   bit more efficient than they were in the past. CIOs really need to have a documented vision   and strategy around this. I've done some writing  around what goes into that: getting into a sense  

of what problems to focus on, what areas of the  business are going to have a material impact by   experimenting with that. What's the goal around  an LLM capability that maybe they're trying to   build up over a horizon two and a horizon three? I think there's a lot of learning, but the risk   here for CIOs is to get something out there  on paper and start communicating, letting your   business partners know that you are going to be  the center point of putting a strategy together.  There's a risk of shadow AI when a general  manager says, "I need to go figure this   out," and decides to go off and sign up a new  platform or test a new set of capabilities.  Even if it's wrong, it's going to evolve. What  we are talking about now as a set of capabilities   has changed since two, three months ago. Being able to do blue sky planning with  

business leaders, with technologists and data  scientists on a very frequent basis to say, "Is   our strategy aligned or do we need to pivot or do  we need to add to it?" I think that's really the   goal for a CIO now is to continually do that over  the course of how this technology is changing.  Tim, Isaac just said that the learning that  CIOs are undertaking right now is, frankly, too   slow. Do you agree with that? And the important  question, therefore, is what should CIOs be doing?  Generative AI doesn't work in a vacuum,  and the remit for the CIO is not just about   generative AI. They're still navigating. CIOs  are still navigating through trying to figure   out how to make things like hybrid-work  work, how to make remote-work work,   how to think about technical debt, how to  deal with cybersecurity and ransomware.  All of these other big, huge, monumental issues  still exist for the CIO. And so, this is just  

one more thing, one really big thing that has  popped up and now is on top of everything else.  He's right. The learning is moving  too slowly. But at the same time,   this is where I think it's important to  leverage your peers, leverage trusted sources.  It doesn't have to be just the two of us, but  it could be the two of us. It could be others.   There are some really good sources out there  that are actually looking at the bigger picture   and starting to navigate and starting to see  some of these challenges that are coming up.  Then make sure that you're putting  the right guardrails in place to   at least bide you some time. Don't hinder  innovation. Don't hinder experimentation. 

I've seen everything from trying to block  access to ChatGPT to blocking every .ai   domain (by institutions). I don't think that's  necessarily the right way to go about it.  I understand why those individual companies do  that. But I think it's important to find ways   to expand your organization in terms of how it  thinks about experimentation so it's not just   the CIO learning, but it's their executive  team, it's their junior folks on the team   bringing fresh and new ideas to the table. It's about leveraging your organization,  

and I don't think we do enough of that  through the course of being an IT leader.   But this is a good time to explore that and  generative AI is a great place to test it out.  How much emphasis right now should  CIOs be placing on enterprise AI,   generative AI? How much resources compared to  all the other things that are on the plate?   Let's make this really practical right now. I would say a lot and the reason for that  

is because we know that all of these big  spaces, these big boulders that we have   to contend with as IT leaders, they all are  going to rely more and more on data. And so,   when you think about data and you think about  the mounds of data that you have access to today,   let alone what you have access to externally,  you have to find a better way, a smarter   way to be able to gain insights from that data. This is where things like generative AI can really   help move that forward. We saw that recently with  networking companies that are starting to move   natural language processing into the process. Isaac mentioned Copilot and code development.   But again, let's balance that, too. Code  development is a great example of ensuring   that you're not leaking intellectual property  or you're not inadvertently bringing someone   else's intellectual property into your code base. Those are examples where you have to find ways  

to leverage this in a myriad of different ways. I  think it's a total game changer for the enterprise   across the suite of pillars within the company.  But you have to make sure that you're using it   in the right way so that you don't end up leaking  confidential information or intellectual property   or, vice versa, bringing that in inadvertently. CIOs are looking at their business strategies,  

looking at where they are in the technical  modernization, where they are in terms of   security. There are a lot of companies  out there that are still catching up.  When you're still catching up, you look at a shiny  object like AI, and your data isn't in order,   you don't have a way of communicating with your  employees in terms of what experiments they should   be working on what experiments they shouldn't  be working on. You have known security issues.  I look at the entire corpus of what is the CIO  driving at and say, "Okay, where are the greatest   opportunities and risks?" For some industries and  some companies, AI is going to be in the top three   right now, and it should be in the top three. For a lot of companies, the CIO's job is to  

articulate back to boards and back into their  executive group, "Look. It should be in the top   three. But we have no data governance strategy.  We're still struggling with hybrid work," some   of the examples that you shared there. We need to catch up on that, I think,   regardless of where you are. I think when we talk  about CIOs being too slow, I think the real issue  

here is the CIOs need to come up with a point of  view and a documented plan and strategy around it.  It doesn't necessarily have to be elaborate or  consume 20%, 30%, or 50% of their resources,   but it needs to say, "Look. Here is an area of  the organization where we're going to experiment   with these tools in these areas. If you have your  own area that you want to experiment, come to us.   Here's how you register an idea with us so that we  can get the right data and tools in front of you."  Not being able to do basic portfolio-like  activities around where you're experimenting,   I think, is a key area. Then being able to take  those and marketing them back to the business  

sponsors and the executives saying, "Look.  This is where we're experimenting. This is   where we're starting from. These are the impacts  that we want to see. And here are some of the   learning activities that we're doing so that  we can see what's happening in our industry or   maybe learn from some other industries." I always say CIOs have to be one of the   champions of being lifelong learners, and so  going out there and seeing what other companies   are doing is part of this equation as well. I say something very similar. I say it takes  

a village. As a CIO, it takes a village. You  can't do it by yourself. Your CIO network   is immensely important, especially now with  these new innovations that are coming about.  But here's the risk that I'll also put  on the table if you don't follow Isaac's   advice of putting that vision in place—you  mentioned that earlier—and really thinking   about what's your plan of attack. I'm already  seeing this play out in organizations where   the edicts are starting to be discussed. What I mean by that is you get the CEO,   you get the board, the board, edicts  from the board for technology. That's  

really bad. We saw this with cloud. We're  still dealing with it with cloud today.  But the last thing you want is a governance body  – whether that's the ELT (executive leadership   team) or the board of directors – dictating what  technology should be used. The CIO needs to be,   absolutely needs to be, in front of that and  driving that conversation. If they're not   driving that conversation, I think that opens  a whole other conversation about where they   are and where the organization is today.  But I would encourage every CIO to get in   front of this and have that plan of attack and  communicate it and socialize it with the ELT.  But how do we get there?  Isaac, how do we get there?  CIOs have to get used to putting out  something with incomplete information and   where the information is going to change. This is  actually one of the easier areas because of that. 

If Microsoft or Google or somebody else comes out  with a big capability that's another game-changer,   guess what, board, guess what ELT, the rules  of the games have changed again, and we can   go modify our strategy. This is actually one  of the easier areas for CIOs to put something   out there that may end up being wrong. I think the problem here is that CIOs are   afraid to be leaders around it, to go and  bring all the general managers in place,   to bring in the functional leaders  in place, and be that coordinator of   where the focus should be. I think that's the  skillset, that's the job here that's missing.  The other side of it, again, is controlled  experimentation. It's hard enough to go to a   team of 100 Ph.D. data scientists and ask them to  have a disciplined, scientific process about how  

they're focusing on which problems to focus on,  when to pivot, when is it delivering value. It's   a constant churn to bring their models from early  stages to something that's ready for production.  Now you're doing it with the entire  organization, like everything is out there. So,  

is it worth experimenting in marketing around  content creation? Well, if you only put out   two blog posts a month, maybe not. Or maybe give a  simple tool that's licensed that they can go use.  I think the other thing we haven't covered  here, Tim, is just about every platform that   any enterprise is using is probably embedding a  language model capability built into it. I'm not   going to name names, but every one of them has it. It's a limited capability because it's often   siloed to the data that it has access  to. But it's an easy area to experiment,   whether it's customer service or CRM or coding or  marketing automation. Go tell me what you learned  

by experimenting with the technologies that we  already have that have AI capabilities built into   it. Then go back to your data strategies. Most of our data strategies have focused   on structured information. How do we get  our rows and columns in an organized way,   well-documented, in a catalog so people  can do BI and machine learning off of it?  We have ignored our unstructured data because it's  been hard. Now we have tons of this information.   Now the question is, now that there is an  LM—and maybe it's a little hard to use it today,   but in a year or two, it's going to be easier  and easier—where are we going to get strategic   value by bringing in content and making it  easier for people to ask questions around it?  We actually have a question from Arsalan Khan, who  always asks these great questions on CXOTalk. Tim,   maybe I'll direct this one to you to start.  He says, "As IT becomes a commodity—" That's  

an interesting point in and of itself that we  could argue about. Is AI becoming a commodity?   "—and AI becomes used more and more in different  areas of IT, will we have an IT department that   just has a far diminished role and really is  only asked to come into play when there's a   functional business need from a user?" Really, he's bringing into question   the very role and existence of the IT  department. Tim, any thoughts on that one?  If that is the mindset of the organization, full  stop, I think that's just the beginning of the   end for that company. I strongly believe  that IT and technology is the ultimate  

differentiator between competitive companies,  and it gets down to data and how you use data.  I don't see the value in IT as a cost center  or as just a response to a specific business   need. But in fact, if you were to play this out,  I actually think the better organizations (broader   organizations, not just IT) is where the CEO and  the CIO have an incredibly tight relationship.  Now, that doesn't happen very often.  That's actually fairly rare today. But   in those organizations, what happens within IT  is pretty magical, and it's very well engrained   in the business, the business strategy, and  the direction that the company is taking,   especially within their industry and starts to  impact even economies based on their success.  I don't look at it as IT is potentially heading  toward this kind of somewhat nonexistence—I know   that wasn't the word that was used—toward  this diminished capacity. I actually see  

it going the other direction, and I think  that we already have good examples of that   in some large enterprises today. We just need  to expand that leadership capability and learn.  This is a bigger question than this show  about generative AI, but CIOs need to learn   how to be better leaders. I'm going to say  it. That's one of the biggest challenges I   think that we have is that, historically,  IT has actually been struggling with this. 

I think the rubber is hitting the road now  and we've got to step up to the challenge.   This is why you're starting to see some  CIOs get sidestepped and get diminished   in their capability because they're not able  to step up to the challenges that companies   need them to today. And we see other roles that  are starting to come into play to augment that.  I do think there is a bigger opportunity for  those that are willing to A) understand it,   B) learn from it, and C) really  have the courage to go after it.  We're using the word "CIO" here, but I think the  real magic also happens when the CIO has a team   of leaders underneath them. You talk about where  CIOs are spending their time to really understand   the impacts and to market the solutions. They're out there customer-facing. They're   out there leadership-facing. They need to change  the mindset and culture. That means they need  

really good lieutenants who are learning how you  connect problem and opportunity to solutions.  The people doing this are what I call digital  trailblazers, and these are people that,   depending on the state of technology, are  moving up stack, are defining solutions at a   much higher level than we have ever done before. We're not defining storage solutions anymore.   We're using cloud capabilities to do that. What AI is going to be doing over the next   three years is maybe we're doing a lot less coding  and a lot more prompting for code solutions, code   examples, to help me document the code, help me  build an API out that's a little bit more robust.  

We're still required to move up stack, continue  to bring problem and opportunity into solutions.  There really is one major change as these  technologies get easier to use, and that is,   we're taking capabilities that we used to have  to go to IT for and now they're becoming business   capabilities. We're not just doing spreadsheets  and Word documents and Office 365. We're now doing   analytics. We're doing small AI out there. We're  doing code development with low-code platforms.   We're taking things that we didn't have enough  people in IT to do over long periods of time,   and we're bringing it out to our  marketers, our finance people,   and our operations, and saying, "You have  a ton more capability to go do over here." 

I don't think IT is going anywhere any  time soon. I think the challenge is   learning these new capabilities, and it's hard. We're not just coding anymore. We're connecting   an idea to dozens if not hundreds of different  ways to stitch a solution. And now we're trying   to figure out what's a good or optimal way to  start experimenting and then seeing which are   the solutions that will help us get to market  with a technology and a capability a lot faster.  We have a question from Twitter. The question is,  "What about the data privacy? How do you prevent   or mitigate the leak or misuse of intellectual  property and confidential information?"  This is a risk. This is a concern. It's already  happening. We're already seeing examples of this. 

I just wrote a blog post about an example of  IROs (investor relations officers) that are   using ChatGPT with pre-published financial  information to build their press releases,   their financial releases. That's a concern. Then, of course,   you see it in Copilot with code creation. I think the biggest opportunity here—and it's   going to sound like I'm kicking the can down the  road a little bit, but I don't mean it that way—is   education. We cannot put the technology guardrails  in place to cover every permutation that we ever  

are going to run across without inhibiting  the ability to experiment and really explore.  This is where technology does not replace  the human. We need the humans to use that   gray matter between the ears and think  about what they're doing. And we need to   educate them and help them understand where  the risks are so that they can make good,   educated, valued decisions from their seat. I think that's incredibly important. I think   too often we think of IT people (or others think  of IT people) as machines that go to the freezer   and get the box. We've way over-rotated on  that, and we need to come back to "these   are people." These are people that have a  brain, that can think, and we need to give  

them the trust and build that education so that  they can help make great business decisions,   whether they're the most senior person in the  IT org or the most junior person in the IT org.  Those organizations that have gone down that path  have had wild success at limiting the risks. Also,   these people start to become identifiers to  say, "You know what? I recognize this issue.  

Maybe we should talk about that,"  and educate the rest of the team.  They become lampposts out in the distance, too.  I think that's another piece that we often don't   think about is the power of the organization.  To me, that's where the guardrails start. 

How are attitudes in the C-suite,  in general, evolving towards AI,   towards the consumption of AI, and the  adoption of AI? Where are we today, Isaac?  We have a spectrum. I think some believe  it's science fiction and we're ready for   Star Trek – you know, "Computer, help me  with this problem." – and expect business   leaders and IT leaders to collaborate to get  an answer around that. I think you have a lot   of naysayers who would prefer keeping business  as usual, prefer using the tools as they exist. 

I see that as a spectrum that we see with  every new capability. The main difference   here is it's moving a lot faster and a  lot more people are experimenting with it,   so it needs to come together much faster. I think how we consume generative AI will   evolve. Isaac talked a lot about building  LLMs within your organization. That's  

really challenging and really expensive to do. One organization recently mentioned—this is a   vendor—that they're paying $0.5 million a year for  junior folks. And the more senior folks, they're   paying $1 million a year. Now that's an expensive  individual, and if you go back to when we were  

trying to put data scientists within our IT org,  we were having trouble finding that skillset.  I think there's that. There's also the actual cost  of running these models in the cloud and paying   for that. I think it's just too complicated  for most organizations to do it themselves.  I do think that the way we consume this will  be through those enterprise apps that we   already are using because they understand  the data, they characterize the data,   they can map the data and market, and they can  help put the right guardrails in place. It's  

a much easier lift for the average enterprise. There will still be corner cases where folks are   building their own LLMs and managing that.  But those will be corner cases and very,   very specific. It will not be the masses. I agree with you. It's really expensive.  

The skillset to work with LLMs is hard to get. This is really about, number one, getting your   data understood so that when we start embedding  them in LLMs, you know what that data looks like,   and then training and educating the employees to  the executives about planning for the art of the   future where you can ask those questions and there  may be an LLM in the middle that's being able to   answer them. It's not really necessarily  about going out and building it yourself.  The other thing we haven't talked about,  Michael, and we don't have time today,   but I would encourage folks to just be aware  of the whole regulatory space. There is a ton  

of regulation that is on the books today  as well as in process at a state, federal,   and global level that companies have to contend  with, and they have to be aware of. It's only   going to get messier as we go through time. Again, the problem gets to be harder over   time (not easier) and more complicated. I  think these are things that (going back to   your point about education) we have to learn.  We have to spend time and learn and learn fast.  A couple of comments from Twitter. Doug Gillette  on LinkedIn says that the answer to the question   about the relationship between AI and employees,  he says, "You don't ban AI but monitoring   employees is where many should start." I know that  can potentially be also a provocative statement. 

Then Arsalan Khan comes back on Twitter,  and he says, "Do CIOs and CXOs, in general,   actually understand the value, the financial  value of their data and where to therefore make   investments in data?" Very quickly, any thoughts  on this data evaluation and investment question?  To the first comment about monitoring employee  behavior, I think that's a massive hot button   and something that most organizations should  absolutely stay away from. There are specific   ways that IT orgs are doing that, but they're  incredibly transparent about how they're doing   it. It comes back to trust in your employees. Then to the second question about the guardrails.   I do think that there is more conversation to be  had there and more understanding to be had there.  We're in the very early innings of this.  But the game is going very, very fast. 

I would add risk and opportunity sides.  On the risk side, we talk about chief   privacy officers and chief data officers. We're now at a point where every business leader   needs to understand the basics of this. Where  are experiments happening? What data is being   used? Is it being used in an appropriate way? I think we're taking that knowledge and that   responsibility and saying, "We need to have a  lot more people with that level of knowledge   to make better decisions around them." Then on the flip side, they have to take  

a stronger role in saying, "We're going to  experiment chasing after value and chasing   after experiments here. Here's what the  value looks like when we want to have   developers do something or have a test out  in the field with some IoT technologies."  Being able to capture that in a way that  you can go and say, "We're heading down   a path that's starting to deliver value,"  because if we end up in a situation and said,   "What's this added data source worth to us? What's  the ROI around an experiment?" we're not going to   be able to calculate it upfront. We're only going  to be able to say that the experiments we're doing   week-to-week, month-to-month are leading us  down toward this value equation that we think   will ultimately deliver financial returns. It's important to understand something that   Isaac mentioned, and I want to underscore  this because it's really important to focus   on this. Focus on value not cost – value, not  cost. The reason I say that is, quite often,   people are focused on, "Oh, well, the cost to  do something is X," but they don't necessarily   understand what the upside of doing that is. To the question about, "Do we understand the  

value of data?" No, we don't. We're learning  that as we go. I think that will get exposed   as we go through time and more experimentation. Deep Khandelwal says, "How can you bridge the   knowledge gap in employees to help them get  ready for AI and new technologies?" Isaac?  I want to market teams that are having  success. I want to be able to bring that  

to the forefront and show what they're  doing, how they're going about doing it.  I'm thinking of the old school town hall picture,  but we need to be able to show people examples.   They went out there. They went above their  job. They went after something that had a  

real problem, an opportunity statement around it. It was defined upfront. They went through two or   three pivots in their experimentation. And here's  something. Maybe it's not ready for production,   but here's something that proves that  they're heading in the right direction. 

I think CIOs, with their CHROs, with their CMOs,  need to have a program about bringing these   success stories across the organization. We've talked about that for a long time,   for decades, and I think we've kind of lost  the marketing ability. But I agree with   Isaac. We need to bring that back, finding those  experiments that are successful and being able   to support the teams that have discovered that. But also, let's support the people who failed.   They tried, they failed, and they moved on to  the next thing. What did they learn from it?  Let's not just focus on the successes.  Let's also remember the failures or  

things that we learned from, not the successes. We have another quick question from LinkedIn.   This is from Kash Mehdi who says, "Since data  is so critical for organizations worldwide,   how does one go about creating a data balance  sheet? Are you familiar with any particular   frameworks or just guidance?" Data balance sheet. There are a couple of books around that   in terms of turning data into a product  exercise. I think, for most organizations,   trying to keep this simple and starting with  something that most tools already have today,   measuring data quality, measure data  utilization, measure the outputs from your data   experiments. Where is it leading toward value? You get into quality. You get into being able   to load in more data sets efficiently. You're  starting to impact the gears that are going to  

start impacting results that come after that. The other things you have to think about,   especially as you're working at a global level,  there are tools that are just emerging onto the   market that are starting to consider this. But you  also have to think about regulatory and governance   and privacy at a global level because  that's getting incredibly more complicated,   and these tools are what's going to  be required to help you through that. 

This is from Sylvia Macia. She says, "What  is your opinion as to the state of entities   using their own internal transformation manager  such as portfolio program or project managers   as ambassadors or, as McKinsey calls them,  analytics translators to help bridge those   knowledge gaps between the business functions  and the CIO, CTO, and IT organizations?"  There are a lot of people with different  titles that have that business-IT   alignment and business-to-IT translation.  Whatever the titles being used inside,   I think the most important skill set there  is really a product management skill set.  Who is the customer? What value, what problem  are we solving for them? How are they working   today? What are their current tools around  it? Why is this strategically important?  Every organization has different skill sets  around it or different titles around it. But I   think the most important thing is to think about a  customer focus first and strategic intent second.  The last point from Doug Gillette  who just wants to add this,   he says that it's a complicated tradeoff between  AI adoption and monitoring employees. He says,  

"Do you give employees free rein and, therefore,  refrain from monitoring their activities on AI to   maintain the trust? On the other hand,  monitoring provides an opportunity to   create policies and governance and address  issues that might otherwise go unnoticed."  Very quickly, Tim, you made a statement  about the monitoring versus adoption.   Thoughts on that? It's a tough problem. There's no one answer to that question. Each   organization is going to be different. You have to  think about the culture. You have to think about   what you're trying to do, the sensitivity of the  data, sensitivity of the risk that comes with it.  There are a lot of factors that go into the  answer to that, so I wouldn't say that you   live at the extremes of whether you monitor or  don't monitor. I think it's a more complicated  

conversation to have, unfortunately. Isaac, Kash Mehdi comes back and says,   "How do CIOs view chief data officers?"  I love this. We're getting provocative. I   like that. Not as an adversary, as an adversary?  If yes or no, how do you empower them as a CIO?  You're getting into the entire organizational  structures. Does the CDO report to the CIO?  

Are they peers? What are they really  chartered with? Are they more data   governance focused or do they have both  data and analytics responsibilities, even   though they just have a data officer category? I think it really comes down to sitting down in   rooms (or virtual rooms) and coming up with  what the objectives are, agreeing with them,   and then drawing that line and saying,  "Regardless of the reporting structure,   here's where the CIO's organization is going to  lead things and take ownership of, and here's   where the CDO is going to take ownership of." If you try to make too many generalizations   around that, you end up with putting people  in boxes, which I don't think actually works.  Tim, do you want to have a last crack at  this? Then you're going to have the last word.  I've seen pretty much the spectrum, as Isaac  has. I've seen the extremes of the CDO is a   fill-in for the CIO that's incapable of taking on  some of these more business-centric pieces. Okay,   there I said it. Unfortunately,  

that's the reality of the world we live in. And  so, now we're starting to see chief AI officer.  I think the thing that people forget is  this title is getting watered down to   the point that it's irrelevant. A chief officer role has a very   strategic and important role amongst the executive  leadership team – or it should. Unfortunately,  

we don't necessarily play by those  rules evenly. I think there's some   conversation we have to get back to doing that. With that, we are out of time. A huge thank you   to Tim Crawford and Isaac Sacolick. They are  two of the very best CIO advisors out there.   Gentlemen, thank you so much for being  here. I'm very grateful to you both.  Thanks, Michael. Thank you, Michael.  A huge thank you to our audience. You guys  are intelligent, bright, and sophisticated,  

and you make CXOTalk. You guys are awesome. Now before you go, please subscribe to our   newsletter, and subscribe to our YouTube channel.  Just go to We have amazing shows   coming up, live shows where you can ask your  questions, and truly, genuinely, you are part   of the conversation. So, check out Thanks so much, everybody. Have a great day,   and we'll see you again next week. And I should mention, next week, we have   the president of the consumer group at T-Mobile  on CXOTalk talking about customer experience. The   president of the consumer group at T-Mobile, you  know he's pretty high up there, so check it out. 

Thanks so much, everybody, and we'll  see you again next time. Take care.

2023-11-17 16:21

Show Video

Other news