Despite the vast investments in digital transformation, worldwide $3.9 trillion by 2027, according to IDC’s forecasts – up from $2.5 trillion in 2024 – persistent challenges remain; high failure rates continue to plague enterprises. Only 26% of enterprises reached their digital transformation objectives in 2023.
A study by McKinsey found the cost of failures to be high: $2.3 trillion wasted on stalled efforts and 17% of business technology projects posing an existential risk to organizations.
Despite the attention, AI initiatives aren’t fairing much better, with 42% of companies projected to abandon most of their AI initiatives this year, according to a study from S&P Global. That’s up from 17% in 2024. Deloitte finds that 70% to 85% of GenAI projects fail to meet return on investment targets, with 30% abandoned post-pilot due to poor data quality.
Recently, we reached out to Tim Crawford, a CIO strategic advisor at his firm AVOA, to discuss the nature of digital transformation and the nature of transformation failures, as well as the mindset it takes to succeed.
Crawford has extensive experience in business and technology. He has served as CIO and in other senior IT leadership roles with global organizations such as Konica Minolta/All Covered, Stanford University, Knight-Ridder, Philips Electronics and National Semiconductor.
Digital CxO: When discussing digital transformation, you prefer the term “business transformation.” Why is that?
Crawford: Everybody talks about digital transformation, or has been talking about digital transformation, and part of the problem there is that it’s become just digital for the sake of digital in many cases. For many organizations, it came down to spending millions to upgrade and improve their digital estate. That includes their systems, network architecture and applications. The problem is that all that investment was very much technology-focused, but not very focused on business outcomes.
At the end of this, organizations risk ending up in a situation where the CEO or CFO asks the CIO: “Okay, great, we just spent ten million on this digital transformation. And they’ll put digital transformation in air quotes. But what did we get for it? Did we expand our revenue? Did we reach into new markets? What did all this investment get us? The short answer too often is “nothing.”
In a few cases, CIOs have even lost their jobs in this process. This made “digital transformation,” at the end of the day, dead on arrival. Because “digital” transformation is not what we should have focused on in the first place.
Digital CxO: What does this mean for CIOs going forward, and how should they be sure to make this shift in mindset?
Crawford: I’m a big proponent of the CIO as a business leader first. Who just so happens to have responsibility for technology. When you have that mindset, you think about transformation, starting with the question: How will we transform our business?
That translates to: “How are we going to expand our revenue or reduce our costs, or how are we going to differentiate our business?”
That’s the actual transformation and why one brings innovation into the mix. It’s a business catalyst. It’s not the flip side: Do the transformation, and then these business benefits will happen. Instead, you focus on these business outcomes and decide what technology shifts will get you there. It’s a different mindset, and when you get into that business transformation mindset, it helps you focus on what’s important and discard what’s not essential. It aligns with those aspects beyond technology, simply because things need to change beyond technology. These aspects could be culture, or they could be your processes, or they could be the go-to-market product mix. It could be any number of things if it somehow transforms the business.
Digital CxO: Why is forgetting to focus on the business need such a common and persistent challenge in business technology?
Crawford: It’s a great question. There is this mindset of it being shooting fish in a barrel. If you bring technology and innovation to the mix, there must be some positive outcome that comes from it. The short answer is generally yes. But when you’re talking about big numbers and you’re talking about value, remember, value is an equation. Many people want to talk about it, like we spent ten million to do this digital transformation project as if that means anything about its outcome. Who cares? The cost isn’t the critical piece. The price is one component of an equation, and that equation is the value. So what’s the outcome for that ten million? If I spent $10 million and received $100 million in additional revenue, that’s a good investment. Suppose I spent $10 million and got nothing? That was not so good. I think this is one of the pieces we learned in the .com era, and we’re learning again through digital transformation. That we must center on value, and it’s not technology value.
Digital CxO: How do business-technology executives reorient themselves toward business value?
Crawford: It seems like common sense, but I guess we get enamored by the technology and go down these side quests that don’t produce value. You could go back to the dot com era and look at the digital transformation from that era and have the same conversation. Look at GenAI today, and this will vary because every single vertical is different, and every enterprise is unique,
There have been a lot of investment miscues with GenAI, but there are great opportunities, too.
One opportunity is for organizations that need to summarize incredibly complicated data. HR is another excellent area for GenAI applications, but organizations must use it correctly and put in the right guard rails. If not, you could fall off the rails pretty quickly. There are also use cases in legal and other areas where a lot of data needs to be summarized. Customer experience is another opportunity for GenAI, especially when moving from chatbots to agents to agentic AI and automation, especially when agents call other agents. There is a lot of opportunity.
Digital CxO: I think there’s a lot of misunderstanding here. What is the difference between traditional robotic process automation and “agentic AI?” Is it the intelligence or the “agency” to adjust to new and unforeseen conditions and to know when to hand tasks off to another agent?
Crawford: There’s a level of intelligence, a level of automation, and a level of complexity to it. With RPA (robotic process automation), the intention was, in general, to take existing processes and automate them as much as possible. When the RPA sees this, do this. When it sees the other thing, take this specific action. But agents don’t necessarily know what’s coming at them, and they can make decisions on the fly within particular parameters. Not being able to do this is one of the problems that chatbots struggle with. With chatbots, the different kinds of questions a user would ask had to be anticipated beforehand. The idea with agents is that those questions don’t have to be expected. They can be asked and intelligently parsed through what they are asked. They can then decide as to how to manage the request. If that agent can’t answer it, they may reach out to another agent that likely can.
The other piece of this is you have to put in work, and sometimes it is a considerable amount of work, ahead of time to get the value out of AI agents. That’s something that people are starting to come to terms with now.
Digital CxO: A lot of groundwork is necessary, such as readying the data and the infrastructure. And I think the laggards there are learning this now.
Crawford: Yes, because the reality is, and that’s a good point, George – the fact is that enterprise data is not in a state that is easily consumable by people or by technology. There’s still a lot of prep work to get that data in a state where you can leverage it to its maximum potential. We still have this data problem. Data is the gold. I wouldn’t go so far as to say data is the new oil. That has some bad connotations and confusing aspects, but data is the root of success.
AI does have a lot of potential across the board for enterprises. However, it needs to be used appropriately. Just like we think about digital transformation, we go back to e-commerce, and in the .com days, we have to find the appropriate place to use it and then put the work in to get the value out of it right. It’s not just something you drop in and go, you know, automagically, here we go. It’s not that simple. That’s one of the pieces that enterprises, especially the lighthouse enterprises, started working on in the very early days.
Digital CxO: This has been great. As you can see, are there any downsides to the rapid acceleration of AI in the enterprise?
Crawford: We’re still learning how to use this. And it’s not just the upside. We must think about the downside, too. So, the downside comes in multiple factors. So, on one end, you’ve got the downside of, okay, am I exposing data that I otherwise shouldn’t be exposing? Okay? So we get into the whole data leakage issue, policy, governance, and the rest have to come into play in ways that we’ve never had to manage as complicatedly as we have in the past. We don’t necessarily have the expertise within our organization, so we might not know any better.
The second aspect is that you and I have access to the same tools that every enterprise on the planet has – and so do the adversaries. The adversaries are also looking at ways to use this technology to their advantage. And guess what? This isn’t just Tom, Dick and Harry sitting in a closet somewhere. These are nation-state players. Now, you have an enterprise going up against another country pouring millions of dollars into advancements in this space. This is the other piece that we must be mindful of. When this innovation comes out, there’s always a downside.
This is why we also need to think about opportunity and risk. If you’re a board member, you think about opportunity and risk across the entire enterprise.
We, as CIOs and CISOs, must think about the same thing. We must think about how our organization takes the opportunity that AI brings, but hedge that opportunity against the risk.
