Infosys Manufacturing Tech Index - AI Pulse

Jeff Kavanaugh:

Today we are going to focus on the Manufacturing Tech Index with AI being that prominent tech, being that prominent discipline. And we're calling it the AI Pulse because it is focused on AI and where things are now.

This research is called the Manufacturing Tech Index and I wanted to give you some background. 650 executives from manufacturing companies from a variety of disciplines, a variety of sectors, automotive, aerospace, industrial and everything in between and logistics supporting them. The idea here is that every six months or so, we're going to have surveys, interviews, expert panels to dig into adoption, the challenges, investments in the industry. There will be some statistics, like an IndiSea index, but it'll also be a fair amount of where things are going and trends. Overall, think about this gap between intent and impact. Yes, companies have moved beyond experimenting with AI. They're embedding it into their strategy, and we'll look at the statistics. However, there's this gap between this strategic comparative and aggressive investment and actual execution. So execution capability is a constraint that we're seeing.

First point I'd like to highlight from the research that AI and manufacturing is focusing and needs to focus on execution because all the best strategy, we call it the binder, like in the old days, isn't going to make it work. It's necessary but not sufficient. So execution is a theme that we've seen come up. Getting into the data, the good news is three-fourths of companies are embedding or are prioritizing AI in their strategy.

You can see the mix is down to 2%, which ironically is interesting because a couple years ago in some other research we did on AI readiness, 2% of companies were ready. We had this five-dimension framework, and really only about 2% of companies were ready at that point. It's amazing to see the progress, but I think people are now trying to cross that chasm between all the pilots and the implementation. So integration into operations and all the complexity implications that that carries is what people are struggling with. Also, that manufacturers, the ones that do embed AI into strategy, they're launching more initiatives, launching more pilots, more proofs of concepts and more full implementations. That's good, but this launching of initiatives does not by itself guarantee success. The execution capability and it's not the same execution as perhaps that might have worked in earlier IT or even OT situations. There's a new version of it emerging, which we'll talk about in a moment. So, execution is important regardless of strategic intent.

The size, and this is interesting because a couple years ago it was all about concept and pilot, even last year, people were talking about on the verge of getting out of pilot purgatory. If you look at the size of this, over half is more than two million and realistically that two to three million seems to be a sweet spot. It's not this massive transformation that we may have seen with the ERP, but it's big enough to implement a use case across a unit or across a company. In fact, in our research we've seen that the feedback we're getting is this larger project can do more, this project size, but it also carries with it the consequences and complexities of integration, of data and governance and multiple stakeholders and adoption. So that is something that came through very strongly in our survey and our research.

Second big point. Just the intent alone isn't enough. You need the execution. One in five initiatives are starting to deliver value. You might say that's pretty low. Good news is it's much higher than before. Of AI-focused initiatives, before it was about can we prove the concept. That's where proof of concept, literally. Can we get an outcome? Can it work? Well, it's beyond that now to delivering value. So 20% is an improvement. That's the good news. At the same time, we're seeing workforce readiness.

That was one of the dimensions of our readiness research which we continued, continues to emerge. In fact, Jasmeet mentioned about talent. We could have an entire discussion about our research on talent, upskilling, readiness, all those areas. But just suffice it to say, it still remains as a constraint. And the leaders who in our surveys, in our one-on-one interviews, they've mentioned this repeatedly, that do I have that talent?

Do I hire that talent? Do I grow it? And you know the struggles with that, and they need a partner for it. Now, we said one in five projects started are delivering value. Let's say for those that have been implemented. You'd think all those are delivering value, right? Well, still under half, based upon the research feedback. And you can see the breakdown. So this uneven success rate is interesting when you dig a little deeper you see certain areas like customer service or marketing content seem to have higher ratios maybe because they're a little more mature. It's a little more difficult out in the shop floor. But the good news is the one theme has emerged and I'm not sure if it's a new term but we're coining it. Reinvestment agility. So imagine the agile methodology or approach applied to take the feedback and not just reinvest money or funds but resources, focus, processes, decision making. So that is something to think about in your own business. Are you agile at the way you incorporate, learn, and reinvest your time and your management attention?

Now, this is also something that's interesting. Based on what we said, you would think, well, the strategy of deeply embedding AI is the only way a project will be successful. We actually found that regardless of the broad strategy, we had three buckets. Maybe not laggards, but just not as strategic, somewhat significant, or very significant embedded in a pillar. The success rate was similar. And we found when we controlled for it, that it came down to how well you executed your projects. Did you have an execution capability or competence in your company? If you did, you tended to be successful. However, strategic intent still matters. You know why? To have a baseball analogy, you get more at bats when you launch more initiatives. You learn faster. And so while execution is critical and it's needed, by having this central pillar and investing more, you're getting more times of learning cycle, cycles of learning. So it is important to step in there and not just wait and be a fast follower because we've seen these nonlinear graphs, the longer you wait, the separation occurs. And I wouldn't call it a panic, but there's an angst that came through in our survey and our research. People know they needed to make the investments and move quickly, but there was a fear that if we move here, we should go there.

Are we off? And I think this reinvestment agility comes back as a good competency to develop and nurture.

Now, we looked at finance companies. While we didn't go as deep purely out of finance, these were all based upon lending in those arms of companies. We're seeing similar things emerge, similar ratios. There are some differences, especially with consumer data and the tech that's more consumer facing. What really was surprising of all three, and you might say that this isn't quite as strategic, but it is, is cybersecurity just emerged at the top of the two or three or four messages coming out. For manufacturers, cybersecurity and the related operations technology systems was the number one concern, number one implementation area. If you notice, these escalating threats, mentioned JLR, you know, last year is on everyone's mind, but there are many at the next level as well. Cybersecurity, all the other ones are important, but not just for the issues, but also just for the opportunities. And if you look in finance, it's similar, although there's also this whole fraud aspect because you're dealing with consumer data much more. The point is, it is similar. And you might think that, well, really is this a strategic topic? Jasmeet and I and others at Infosys, have a relationship with the National Association of Corporate Directors, NACD. Wonderful organization, literally at the board level. And in our work with them, we find, you might not think this, but cybersecurity is one of the most common areas where they're asking boards not just to focus, get certified in at a board level. You would think that's too much in the weeds, but it's that important to board members. And when you marry or pair cybersecurity with AI, that appears to be a real pressure point now at the senior most levels in an organization. So it is worth, if you're not already thinking about that, going an extra level down, even if cyber isn't your functional area. I'm certainly doing that as well. And you notice that it's only the Twin Towers there, data and cybersecurity. I guess using an automotive analogy, of a metaphor. Data is this fuel for AI, but cyber might be the brakes, because if you go too fast, cyber can cause all kinds of threats and issues and risks. We know the asymmetric nature of it. It was a couple years ago now, but the State Department in the US with some ministers from India as well delivering a briefing on where AI was going, and I was next to the head of cyber from MasterCard. It scared me, him going through the tens of thousands of threats every single day and how rapidly they evolve. And so if you think about that for your company, all these entry points, let alone the machines and the OT, all AI does is allow that to multiply. And I think that's something that we need to be a little paranoid about. That's, I think, some of the angst coming back from these good stewards of these company leaders. They're wanting to make sure that first do no harm, while we're trying to accomplish all these benefits. And then if you look at the finance firms, it's fairly similar, but uncertain ROI was also interesting how it popped up. Some of the path to value wasn't as clear. And of course, regulatory, because as you're dealing with finance, you're dealing with customer data, you're dealing with privacy issues, that's also a big deal.

To wrap up, or at least to summarize, manufacturers and finance companies are embedding AI. That's great. It's a strategy. More initiatives are still early than late, but there is progress, and cybersecurity being the top use case. You might ask, though, well, what does this mean? What are the implications? We do want to leave you with five thoughts here, that governance and operating models are important, because Jasmeet mentioned, trust.

In the new model we unveiled a couple days ago, AI trust is one of the six pillars or one of the six areas in this hexagon. They are data readiness and architecture. Again, if you can't scale, then you aren't really driving value at the enterprise level.

Cybersecurity foundations, we mentioned that. And also don't just think about IT and OT separately, but what's that bridge as information is passing data, it's passing back and forth? Value tracking, discipline execution, two sides of the same coin. If you execute and measure, you can determine where value is created and then you can also make better decisions. And strategic use of partners to scale. Our research found that three out of four companies are either partially or totally using partners, only 26% are doing this by themselves because they realize the clock is ticking. They need to take advantage of expertise where they can get it while they're developing their own. So shift from intent to impact. That might be a nice way of rolling this up. And think about this as a core operational capability and not a collection of experiments. And that allows all this human impact, human in the loop, and shaping to occur. Because without this, we're spending all of our time looking over our shoulder and fixing problems instead of looking forward to the opportunities. And again, if you would like to download or scan that, access this research, we'd love to share the other research with you. And we like to say at the Knowledge Institute, keep learning and keep sharing. Thank you.