娇色导航

Welcome to 娇色导航Leadership Live. I'm Lee Renick, executive director of 娇色导航communities for CIO, and I

am totally thrilled to welcome Meredith Whalen, chief research officer at IDC. I actually am getting goosebumps introducing you. Meredith, thank you so much for being here. Your research is so impactful to the CIOs and tech leaders.

You know, they have the opportunity to to research it and read it and listen to you speaking about it.

So could you please introduce yourself and tell us a little bit about your amazing role at IDC? Yeah. Thank you Lee. And it's an honor to be here.

So I am Chief Research officer at IDC, which means I'm responsible for all of our analysts and the research that we develop at IDC in order to help our clients that are navigate the business in the technology world.

And we put out some amazing research papers. But yesterday you presented, on the, on the main stage. It was a keynote address to our attendees here at the conference. I watched it, the presentation was phenomenal. Lots of really great insights.

So could you tell me a little bit about your presentation yesterday and some of the research findings? Yeah, absolutely.

So, you know, a year ago when we all came together at the 娇色导航100, you know, we were about nine months into generative AI and we were all, you know, dreaming of all the possibilities and what we could do.

And a year later, you know, when we look at our data, we've seen that organizations on average, have done about 37 proof of concepts, and they've only moved about five into production.

And so it's been a big year of experimentation. But now the question is, you know, what is it going to take to move from experimentation really into adoption of these use cases? And so that's yesterday.

What I was speaking about is, you know, the main areas that we see are AI strategy, having a unified governance model, and then also getting over the technology costs associated with this and figuring their way through that so they can present a positive business case to their executive team.

And that is so bang on with the CIOs I've been interviewing here. You know, at, at at our conference, those are some of these things that they're talking about, especially the governance issue.

And a lot of sales are telling me they've, you know, used generative, generative AI, excuse me, to build productivity internally with their teams. So we're hearing a lot of that around maybe Copilot or, you know, ChatGPT or utilizing that in a gated space. Right.

But they're not really extending that out to necessarily the consumer aspect. Unless maybe they're like a retail, they have a chat bot and they're using it in that way. So I'm hearing a lot of that about it. just that well, I'm well with the research. Yeah.

And so what what happened was there was this big scramble to be able to identify genius cases and to experiment with it.

And so the challenge organizations have found is that, you know, when you look function by function, you know, every function may have said, you know, there's 4 or 5 or six of them that we identified.

But holistically, when you look across the organization, you may be looking at like 80 different gene AI initiatives that have been prioritized.

And so for an organization to be able to figure out which ones are really going to have an impact on the business, that now is the time to bring together all of those and create an enterprise use case roadmap for AI use cases.

And in doing that, what we're encouraging leaders to do is kind of look through and say, like, what really are going to be the super use cases, you know, which ones are really going to give us the biggest business outcomes for the amount that we're investing in that use case, which ones are going to help us be more resilient and which ones are going to overall, support our health as an organization, whether that's being innovative, adaptable or meeting our sustainable growth goals.

And I think that's the lens that now needs to be applied. And it's really needs to be a partnership between business and IT. That's also something that we've seen in the past year.

You know, our data shows almost 40% of organizations said, look, we didn't work closely with business in it. And as a result, that's why we're having a harder time moving these use cases into production, because there wasn't probably that same type of vetting. And no problem.

I mean, that's a lot of what experimentation is about. But now it's time to get organized around this partner with the business, create that enterprise use case roadmap. That is really great here. And I had those discussions today.

I interviewed for other CIOs, and that's exactly what they were talking about. The ones that had successfully moved those use cases forward.

They looked at all of them that they had done and really narrowed it down to the top 3 or 4.

One said of the best ones that would move the business forward, create that productivity internally, which was massive, you know, like taking 400,000 hours of work that people would have done down to, you know, a certain amount, very much smaller amount.

And then, you know, letting the board know, letting everyone know that we've created this productivity and then focusing on that and launching that out to the, the customers and, and getting their sort of feedback on how that is evolving for them.

So I've really been having those types of conversations as well. So that's really tying into sort of building that AI tech foundation.

So you know, what are what are some of the thoughts here you could share with the 娇色导航listening in on that. Yeah. So again, I think something that we've started to get greater clarity in the past year is this big.

Do I build or buy or compose my own model. And so, you know, first and foremost, you're usually going to try to buy the technology.

but when you look at it, there's not necessarily all your use cases being addressed through the package solutions that are out there today. More and more will be over the future. But the reality is, is that most of these cases, you have to do some type of composing.

You have to take a, you know, commercially available or, open source lab and then do inference on it or some form of grounding with your own data set.

And that's what most of them are finding our data showing, like very few are actually building their own. Lem. and so I think that's some of the clarity that's developed.

I think a year ago people thought they had to do that, and now they're recognizing that there's other paths to get there.

And so what that leaves a lot of organizations have to recognize, like if we're going to have to do this composing, it's going to rely on our data.

And so this is getting back to the data issue again, and that organizations may have older data that they're uncovering that they're putting in there and realizing we haven't updated these documents in years, or that they have incomplete data sets. And so that's a big focus.

And one of the things that we've been recommending is, like, as an investment, we tend to spend a lot of money, you know, in on what we call the data plane, which is really, you know, where we store our data and we organize our data and we manage our data.

And now we're starting to invest in really all these AI models and to do the grounding, to do the fine tuning.

But we don't really invest in like the data engineering, the exercise that takes the data and gets it, puts it in context and gets it ready for those models, because that's where the governance is going to happen as well.

And so that's one of our key recommendations is don't overlook that layer as we're all starting to jump up and look at that next layer around the AI models. Yeah. And certainly the discussions I've been having and at this conference has all been around data.

I mean, as one 娇色导航said to me at one of the conferences I was hosting, you know, garbage in, garbage out, and, you know, the CIOs today were talking about, well, we have data, you know, during Covid, we put it in the cloud. Right?

Then we found a that was costly in some instances. So now and then we had to figure out how to use edge to cloud computing, bringing the most important value to the edge.

And now we're saying, well, maybe we should bring some back on Premier, have a combo of it, right, and really ensure that that data that we need the most is really clean.

And going into that gen AI instance where it will be the most productive and they'll get the most value from the data. And that really is pointing to this future state which will be really fit for purpose infrastructure.

And it's true, because right now, as companies are experimenting with their proof of concept, sure, it's fine. They're using it in the cloud, but they're now going to have to be much more thoughtful.

Like, you know, as the AI PC comes out and you can run the workload. on the AI on the PC, is that where are you going to put it? Are you going to put it on Prem? Are you going to put it in the cloud?

And so you they'll start to figure that out based on the use case.

And that's going to be an important part of the technology investment plan. So you can run those costs effectively. Thank you for sharing that. All right. So let's talk about securing and governing the AI enterprise.

So you know a lot of the sales today we're talking about security security with AI. And but also the governance aspect of it. Right.

So they're really focused on that. We talk to a lot of the CIOs about their role in the leadership role of, you know, being an educator with the board and the C-suite. and so maybe you could talk a little bit about that.

And then this aspect of like, you know, the governing aspect of AI, the AI enterprise. Yeah.

And again, if I go back to when we were together a year ago, we were really talking about how do I experiment with guardrails, right. Because we knew we had to experiment and we had to put some governance guardrails in place.

And I think organizations did a fine job doing that. But in doing that, the governance that's in place is kind of fragmented. Right?

So again, now it's time. It's time to go back and look holistically and say, what does it take to build a unified AI governance model?

And when I look at some of what got put in place, it's pretty focused around the technology, governance, you know, particularly around the data governance or the model governance. Again, perfectly fine.

You needed to put those guardrails in place, but now it's time to take a step back and say, how does this fit into our strategy?

Because once you identify, you know what your strategy is and how governance fits in, then you can start to make decisions as an organization about what you're going to do and what you're not going to do.

Once you define what responsible AI is, you can figure out what your commitments are, what will be the organizational commitments that you'll make.

And so we're really recommending that the organizations think about that first. And then you can zoom in and look at the technology. In this time when you put your governance in place, remember it's the data, it's the model, it's the applications and the infrastructure.

And think about your governance holistically across what really is an AI operate technology operating model.

And then you can put in place, you know, your your processes to assess the risks, to implement and support the governance. Report it all out. But it's really important to tie it to the strategy. That's really interesting.

I just interviewed a 娇色导航who was talking about how they how he works with his C-suite, the CEO and the board of directors, and they have specific agenda items around this and every board meeting, in every agenda.

And before that, they really have an educational partnership with the CEO. to allow that person to understand what they're doing, understand the risk, understand, you know, the profitability of what they might be doing and how they've tested it with the group. Right.

So I think there's there's this nice sort of feeling for me of the agile work, sort of right with this, but then bringing it all together and having that governance model that they're building up, but also ensuring that they're CC C-suite, excuse me, their board of directors really understands the process.

They're going through a great I think that's the right way to do it. All right. So let's talk about creating business value.

I mean, you've talked a lot about the enterprise and that this is going to be kind of wrapped around the whole enterprise as part of the technology.

So any any tips or advice for those CIOs out there listening in around creating that business value across, you know, right across the business? Yeah.

So I think the good news for CIOs is like, you're going to be the one leading this. I mean, this is all all signs are pointing to the CIO. Is going to be the one who sees this across the organization is very technology heavy, exercise in the organization.

So it's great for our CIOs. And really we tried to articulate that there's really kind of two paths to move around this, like AI adoption model.

You know, one, one company will leave this experimentation phase start to move into production in 2025 and then really move to this whole I feel business model, you know, in the next several years, once you hit that, I mean, I think not just us, but the rest of the industry is really expecting those business outcomes that go vertical in terms of what you get for a payback.

The other one is if you take too long and you don't really move out of the experimentation phase in 25, and it takes longer to achieve that. And maybe you don't even really transform the entire business.

I think the delta you're going to see between those two paths is going to be pretty significant, and it really is going to have a business impact.

It's going to show up in the productivity gains, in the ability to how fast you can innovate, how close you can build a relationship with your customers. And it's going to be significant and it's going to show up in the financials as well.

So there's a lot at stake. It's really important to start to continue to move this journey forward. Yeah.

And you had a really good slide in your in your presentation just around those curves. The people that are adopting it and moving forward and looking at that governance and creating that productivity.

And then the ones that might take longer, there was a real gap in overall, you know, how they would, increase increased revenue for the business and just increase the business output. Right, exactly.

In your competitive advantage. Yeah. You're going to have. Yeah. All right. Well, this is where I need you to bring out a crystal ball. Not really, because you're a researcher. So you have a research team.

So it is not a crystal ball, but maybe you could just talk about you have a little bit already around that, you know, advancing business.

But any other sort of predictions around future growth concerns for AI. yeah. I'd love to learn more about that from you. Yeah.

And so, you know, when we think about what's going to happen with the workflows, the processes and the employees, I think that's going to be pretty interesting to watch take place.

Like you're going to have, I either acting as an assistant, an advisor or an agent or all three. Right.

And so when you start thinking about that, that's going to change the actual business processes, and then that's going to change the applications and how you as an employee engage with those applications that you use today. And ultimately it's going to change what you do at work.

And it's going to be really important for organizations to think about the people aspect here and what is going to be the change management plan that you're going to put in place in order to bring the organization along.

If we don't, you know, people aren't going to adopt these technologies and we're not going to see the business outcomes that we're all expecting to happen. Yeah.

And when I, I hosted a deep dive panel session today on AI at the conference and some of the insurance providers were really embracing that idea of what do our people do for us?

How can we get them to, you know, adopt what we're doing, test what we're doing, give us the proper feedback and not be worried about losing their job, you know, and really embrace it because we need them still for these, like, we're shrinking down activity time for them so that they can actually focus on things that we, you know, really it was took too long or it was difficult to focus on.

So we're hearing a lot about that piece around people and really then mapping out and utilizing people skills in a way that maybe they don't have time to do in their regular roles, but also adopting that new process that might be in place, that prompt engineering piece of it learning, teaching all of it.

So, that's what I'm hearing.

Thank you for letting me share that. Yeah. And well, and it's it's gets back to why the governance has to be part of your strategy. Because once you decide what you're going to do for strategy, then you make your commitments.

And if one of your commitments is we're going to have to reskill the workforce, then you can start putting that plan in place for them.

But if you don't really have that overall strategy and vision for where you're headed with the company, then you won't make that decision. Well, it has been totally fantastic to have you here today. Thank you so much.

It's my last interview of being here for the last few days. I appreciate it so much.

I always appreciate your work and your insights, and having the opportunity to see you on stage and present those, what we will do is we'll make sure we include a link to your presentation, if that's okay with you.

Share it with the CIOs and tech leaders. Listening. Thank you so much. Meredith. Yeah. Thank you. I appreciate it.