Why Technology Transformations Fail
David Williamson explains why major technology transformations often fall short before the technology itself becomes the problem. He shares how stronger change management, clearer communication, active leadership involvement, and early end-user participation help organizations turn new systems into measurable business value.
Successful transformations depend on bringing people into the process early, communicating what is changing and why, and giving leaders and end users meaningful opportunities to shape, test, and adopt the new way of working.
Episode Chapters & Full Transcript
Select any chapter or transcript timestamp to begin watching from that exact point in the episode.
Full Transcript
Welcome to Conversations with CISOs, Security Leaders and Technology Executives, where we sit down with the leaders shaping cybersecurity and enterprise technology. Today I'm joined by David Williamson to discuss why technology transformations fail. David, thank you for joining me today. To get us started, could you introduce yourself and tell us a little bit about your background and your current role?
Yeah, sure. Thank you for having me, Sullivan. So Dave Williamson, I spent my career in life sciences as a technology leader implementing systems, technologies of various kinds to drive consistency, efficiency, and measurable business value.
Alright. So, over the course of your career you've led several major technology transformations. In your experience, what causes them to fall short most often?
I think one of the largest problems you have with any sort of technology initiative, transformation initiative is change management and ensuring that you have the right involvement from the people, that you bring them along the journey of understanding why are we taking this transformation on, allowing them to have a voice in how the transformation occurs. And then ensuring that they're set up to be successful with the new set of tools and or systems that they'll be using.
Okay. So what's the biggest mistake you've seen organizations make before a transformation even begins?
Well, I think it's sort of human nature to make assumptions around the complexity of the system to use and the complexity to implement. But I think a lot of that goes to we work a special way for certain reasons and we're not very open to changing how we work and that often sacrifices a lot of the benefits the technology or systems could give you because now you're adapting the system to your current way of working versus taking advantage of what the system could give you. So I kinda end on that note of not an openness to change how one works.
So how can technology executives, when they embark on these transformations, what are some strategies they can use to get their team or even the organization as a whole to be more open to changing the way they work to adopt these new technologies?
I think to me one of the best tools is communication. So engaging the larger end user community. So different communication pathways. One's for the end users, one for the executives and the leadership, maybe another for middle management of what are we doing, what will the benefits to the business be and how are we going to bring everyone along? I mean, just like you hear in the current AI, you know, AI's taking jobs. I'm not convinced that's actually true in aggregate, but yeah, you have to address those concerns up front. So communicating and giving a broad set of regular updates to the population impacted by the change or the transformation.
Okay. And how do you make sure that leadership is aligned and that the rest of the organization is aligned within the culture to adopt the new technologies?
Yes, that's that's a difficult one ensuring the alignment. Often you have various folks at all levels who believe they understand what the change is and how it's gonna look and that doesn't necessarily line up with what actually is being implemented. And certainly I've had experiences at the executive level where different executives have made assumptions about how the transformation is gonna occur. And again communication and demonstrations, and certainly for the end users getting them engaged in the testing and having a hands-on experience with the new systems and technology so that they can learn it firsthand versus trying to I don't know interpret whatever slides or other materials you're sharing with them.
So what are some of the most common misconceptions or expectations that leadership has about technology transformations that you've seen?
I would break it into a few a few points. The first being say I'm doing ERP system, I work in commercial, it's not gonna impact me much. So I don't need to spend that much attention to it, even though it's a large endeavor for the corporation and it probably does impact you quite a bit more than you are assuming. I think the other is I pay attention to some of the communications, I interpret them a particular way, and then my mind is kind of set on what I believe is going to happen and not continually reinforcing or challenging, is that actually true? As we know, you know, change is easier and less costly up front. Often you get into user acceptance testing and it's like, well, it's not working the way I thought I understood you to tell me it would. I think those are the two main categories. It's it's not going to impact me much. I'm I shouldn't be that concerned, or I don't need to be that concerned. Or I was engaged up front. I understand how it's going to work and I don't continue to challenge that viewpoint is accurate.
So what can leaders do to address those things up front to make sure that those misconceptions are not problematic down the road?
They need to be involved or have what you call a trusted lieutenant that's very involved. They need to ensure that their teams are involved and if we have these one on one sessions or user acceptance testing that their folks are engaged in those activities and that they're having their own discussions, meetings amongst themselves, describing what are their concerns or what have they seen that you know, positive and negative, out of those sessions. I don't think you can be too involved in these types of transformations because if it is successful, it will significantly change how you're doing your day-to-day work.
So what does a successful technology transformation look like? And what do successful transformations have in common across the organizations you've worked with?
You've engaged with the people. The people, again, the end users and the leadership, have a solid understanding of what's occurring. They've participated in trainings, the testing. Almost always, you'll have some sort of I don't know, missteps which you're able to address quickly and you're able to respond in an aggressive fashion to feedback once you are live, and you are able to start seeing the change in the business metrics that you justified the project on. And yeah some you hear the horror stories of we struggled for a week, two weeks, etc. To continue the business as usual or other sort of dramatic problems. So the avoidance of that certainly and the just measuring the usage, at least for the initial few days, that people are actually trying to use the system and trying to use it as intended.
So in addition to the usage metrics, are there any other metrics or KPIs that organizations should track to measure whether a transformation has been successful or not?
Yeah, absolutely. You know, we should justify the project or the program based upon strategic business objectives. In, say, inventory reduction or automation of highly error-prone processes or other savings metrics that you can measure. So certainly you want to be predefine those up front and then be measuring them along the way. I think it's very difficult to justify and or get the right engagement if it's just, I'm switching from one brand to another brand. That those are hard. I mean sometimes upgrades are like that, but a true transformation should be focused on driving business value.
Okay. And for the transformations that don't go well, at what point do you, as a leader, recognize that a transformation isn't working and you start to change course and how you approach that?
in my experience it's pretty quickly, say two to three weeks, that you realize things are not working well at all, and you need to intervene up front and organize the right people to address what are the issues to understand what is what is preventing the benefits or the usage of the system as you want it. How can we address that? And what skills help are we gonna need as an organization to do that?
To piggyback off that, what are some of the most common examples you've seen of a technology transformation needing to be reworked? And what are the strategies, the normal strategies that you might take once you get to that point that you need to shift?
Well the majority of these solutions are cross-functional, multi-step processes. You do your part, now I do my part. So often the understanding of what my part of the process is, how that impacts and who it impacts downstream. So that could be addressed by training or, you know, other forms to build that knowledge up. Often the data that's driving the process had inaccuracies. And instead of me stopping and fixing the data before I proceed, I'm often feeling that I am rushed and I need to just move ahead and I'll fix the data later. So capturing those opportunities and that may need to be addressed by other people because of the capacity of the folks in doing the actual day-to-day work. Those are very common. The other phenomenon, having worked in operations, manufacturing and the warehouse, is that the system works very well in a conference room, but when I go out on the floor or I'm on a forklift, now you're asking me to perform a task. Well, I have something like a tablet and a barcode scanner, and I only have two hands. And so you often run into situations where the usage of the system isn't as easy as it was in say an office setting. So that then you need to rework things to I don't know, make it usable is maybe the best way to say it.
Right, that makes sense. So, as you know and everyone listening knows, the most popular transformation today is with AI. Lessons from past technology transformations, particularly with cloud transformation, from ten or fifteen years ago. What are some lessons from that organizations can take and apply today as they adopt AI?
Yeah, I think last year in twenty five we had a lot of organizations rushing to implement AI. They had some ideas around like headcount reductions or other savings that they were gonna achieve. And I think a lot of the older technology lessons, which even we've discussed in this conversation, were being relearned. And your analogy to cloud also has a lot of merit to it. When cloud got say introduced, many organizations rushed to it. They didn't have the proper monitoring or continual feedback around it and you wind up with cost structures that are actually much worse than they were pre-cloud. So we've seen this also with AI, where you've heard in the news of several large organizations and I've used my whole AI budget and it's April. You know, so not having the proper controls in place.
Right. I think it was Uber's CTO sat down like January and was like we've run out of all our money now and we have eleven months to go.
And I've worked with organizations that have implemented some very nice AI solutions. The users love them, they're very excited about having them, and then you find out it costs sixty thousand dollars a month to run. And they're like, well, that's not just that's I
Right.
can't justify spending that much money. So again, in the in line with your cloud analogy is any good organization that uses cloud is monitoring their cloud usage, they're right-sizing the servers in the cloud to minimize cost. You need to do the same thing with your AI workflows. Do I really need to use a large language model? Can I use a different large language model that's not going to take as much compute power and there's ways to manage the costs that come out of the workflows. I think also another learning from last year is some of the most efficient or great use cases aren't very exciting to talk about. So you know, improving your inventory controls, your procurement processes, things of that nature that aren't aren't as exciting as what you hear in the media. But those tend to deliver very significant business value with less effort. So just as I would never recommend to somebody let's move everything to the cloud and get it done very quickly. We've had that same sort of idea with AI of, you know, where can we apply AI and let's let's do a whole bunch of things all at once of again finding where you're going to get the most bang for your efforts up front and start there. And have the right controls in place both to monitor it while the implementation is going on and then later. I mean, we tend to throw governance around AI governance around as an answer, but that's certainly needed. And I think related to that is having a way of understanding what AI workflows are in your organization. So anyone, ourselves included, can download Claude and start building AI agents and use it in the enterprise and that may be a good thing or it may not be a good thing. So while we don't want to stifle innovation, we still need ways to control and certainly understand what has been deployed.
So what are some like with cloud how there was you know monitoring solutions and controls that were necessary to be put in place as organizations adopted that. With AI, what are some of those solutions, tools, controls or frameworks? What are some of the prerequisites organizations should either, you know, do before they adopt AI or as they adopt AI, start putting it in place?
Yeah, so I think part of it is not allowing people to use publicly available tools. So the technology leaders need to deploy tools within their four walls in a secure manner, that does have a cost associated with it. But I don't want you sending my enterprise data to the public version of ChatGPT. There are some very nice newer companies that have deployed solutions that automatically monitor what AI agents are running, what data they are using and such. So I would highly recommend having some sort of solution like that so you know what's out there and what's working, who's using it, so you can work directly with the individuals if they need help or they're trying to do something that maybe is not in your best interest because of what data they're accessing or how they're accessing it. And certainly having tools in place to monitor the costs. You know, how much compute are you using, how you know, tokens and other things that have a direct hard dollar cost. They're it's difficult as in I don't think you can mandate a lot of rules and have the organization actually follow them. People are trying to do the right thing. They're trying to drive the efficiencies. So I look at more setting a baseline, having a way of knowing what who's trying to do what, and giving you a set of tools that should meet the majority of your needs. So you don't have to or feel obligated to go out to a free version or using your corporate card to go buy a pro subscription to something.
Right. Okay. So as we wrap up, looking back on your career, what are some of the biggest lessons that you've learned as you've implemented or seen technology transformations implemented?
Well, I think that the first one and it still applies today is the promise of the technology and what you can deliver in a reasonable amount of effort and time are two quite different things. So you know, starting with a smaller scope that's controlled so you can be successful and that builds a lot of momentum. I think we have graduated from, we're going to you know transform the whole world all at once in one very large project and those don't tend to go very well and they move very slowly. I think giving people like we're just talking about with AI, giving people a safe place, sandbox training system that they can innovate with, that they can use it. You always have a core group, a smaller group of people who are very anxious to get into the technology and say experiment, innovate with it. So give them that opportunity and allow them to fail and fail often because you get a lot of learning out of failure. And ensuring that I'm doing something with a clear business objective in mind. I'm not just implementing AI because, well, everybody else is doing it, so I should do it too, or feel I should do it.
Okay. And if you could give one piece of advice to executives preparing for a major technology transformation, what would it be?
I would emphasize the change management again, that you cannot forget the people and the answer to probably every question in your mind starts with people and engage the right people throughout the organization and in the initiative.
All right, sounds good. Well thank you everybody for watching. If you enjoy this conversation, be sure to like, subscribe, and leave a comment with thoughts or suggestions for future guests and topics. And thank you again to David Williamson for joining us today.
You're welcome. Thank you for having me.
Conversations With CISOs, Security Leaders & Technology Executives
Hear practical conversations with the executives responsible for protecting complex organizations. Each episode explores leadership, risk management, infrastructure, incident response, governance, and the decisions security leaders make every day.