Balancing Agility and Control in Rapidly Transforming Organizations

Elena Legendre:Thanks Danielle. Hey everyone, my name is Elena Legendre. I'm in the transformation area in our advisory services. My background is a little different from most people who do this type of work. So I'm trained in organizational psychology rather than IT or finance, but that really just means that I'm asking a little bit of a different question. So rather than will this system work, which is David's question, mine is can the people, can our employees actually absorb what leadership just committed them to? And most of my work since I've been at the firm for the past five years has been in places where breaking things isn't an option. So healthcare data, child welfare, federal programs, and it still has to move really fast. So that's basically what this entire conversation is about today. So I'm glad we're having it and I'll kick it over to David.

David Fahr:Hi, my name is David Fahr. I am a director on the strategy and transformation team. I've been with the firm six months and really starting a second career. I've been an IT leader for about 25 years in multiple industries, including food and beverage chemicals. As Elena said, I get involved quite a bit with transformations, implementations, ERP projects, IT strategy and AI strategy. So looking forward to the conversation today.

Jen Clark:Amazing. We're going to go to our first polling question now. So we would like to hear from all of you. Where is your organization on its AI journey today? As you just heard from our panelists, we're really going to have a conversation and discussion around meeting this unique challenge probably once in a lifetime. Give everybody just a couple of more seconds. All right. So a mix of perspectives here. A lot of folks starting to learn and explore. A few of you already running early pilots and a lot of you scaling, which is really interesting. And I think we'll get into a lot of that discussion today. A decent portion of you also kind of operating AI and production at scale. I think that speaks to everybody got started with this at different times or at different parts of their journey, probably even within different departments within your organizations themselves.

So today we have a packed agenda. We're going to set up this unique problem that I think we're all feeling today. We're going to talk through the four perspectives as my panelist just outlined, and then we're going to give you some actionable takeaways. It's like what does this actually look like in practice to do well and be successful as well as help you with a playbook? So to set up this problem at the beginning is this tension of agility versus control, there's two opposing pressures that don't always go well together. Leadership teams are really feeling this pressure to move fast with AI, to deploy it rapidly so they don't fall behind. We hear this all the time in our conversations with our clients, that teams feel like they're not moving fast enough, they may be losing their competitive edge. But on the other hand, we often work in highly regulated environments that require structure and thoughtful protection.

We really need to establish and maintain audit readiness and protect sensitive data and IP. So whether you've just rolled out a tool like a Clutter or Copilot or have that kind of operational AI at scale, the opposing pressures really stay around. And the instinct is to really treat these two as a trade off. What we're seeing with our clients is that it's possible actually to have both. It's possible to redesign your operating model even in slow, thoughtful iterations so that you can move quickly and safely with this technology. And I'm sure every one of you in this room feels this probably daily. So I'm actually going to turn it back to my panel. I would love to hear some perspectives on, as you just outlined, each of you have led transformations where this tension has showed up. What is genuinely different about managing with AI?

And I'm going to go to Elena first.

Elena Legendre:Sure. I'll answer that from where I've been sitting most recently. So for the past couple of years, I've been inside a healthcare data program of about 175 people. They are replacing legacy systems that hold a lot of sensitive information, as you can imagine, lots of claim information. So it's one of those environments where they need to move fast, but they have to be sensitive not to break anything. So the tension that we're talking about on this slide is very, very familiar to me. I feel like I've spent most of my career in it and I think organizations that handle it really well never really resolve it. They just kind of stop treating it as a choice. And what I think is really genuinely new in this AI space is that on every program or project that I've run or been a part of with change management, we kind of get to control when people meet that change.

So I spent two years at a university consolidating brand new finance and HR systems. We got to decide which group got trained and when, replacing 17 systems in a child welfare program. We sequenced who moved and in what order, a call center. We got to actually provide the scripts that they used. So AI is that first change where people are meeting it before I did. So as you probably all know, your people are already using it today on very real work. So the first move is not a communication plan, which is what a lot of people think of when they hear change management, but it's finding out what already happened and meeting your people where they are today.

Jen Clark:Awesome. That's great. Danielle, how about you?

Danielle Keller:So navigating transformation when AI as the driver is different because you don't have the option to not move quickly. So your team's going to be using AI whether you are ready or not, and that introduces a tremendous amount of risk to your organization. So you have to find a way to adapt quickly, and this includes adjusting your governance model to mitigate that risk. So you need the appropriate policies, processes, and oversight in place to ensure AI is used responsibly, safely, and legally within your organization. So I think that's really why it feels different now.

Jen Clark:Awesome. Agreed. And David, from the IT perspective, what are your thoughts?

David Fahr:Yeah, I'll tell you, Jen, from my 25 years of leading IT initiatives, the one item that is kind of common thread through all of them is really the corporate culture and organizational culture just drives so much of the success or failure of initiatives. And a high performing culture empowers, it trusts, it allows the teams to experiment, make decisions, even fail forward, and puts guardrails in place to manage the governance around that. Honestly, AI just accelerates everything. So if it's a solid culture, it's going to continue to be solid. If it's dysfunctional, if you have solid organizations, it's just going to just increase the size of that challenge going forward.

Jen Clark:Absolutely. I think your point and building on what Danielle and Elena said around really accelerating that spirit of innovation and that innovative culture, AI really shows that you either have that kind of connective tissue or it hasn't been built yet. And so building on what everyone just spoke about is AI isn't just a faster version of what has come before. It really demands a kind of different operating model across every layer of your organization and likely every department. The kind of four places where we see AI changing the equation the most is really one is around this spirit of experimentation and change is we see this in our engagements daily. Now all of a sudden overnight is overnight since 2023 I think, but certainly in the last year, a dashboard that took six months to build is now prototyped within an hour by a person in the business and then they're struggling how to figure out how to deploy it across multiple teams.

In another case, we worked with a client on a monthly reconciliation that normally took 40 hours and after building and deploying a cloud skill, the task now takes two minutes. And so typically organizations would have time while the technology was being built or integrated to think about these controls and these changes and how to thoughtfully roll these things out. That's just not the case anymore. The tool, the utility, the problem solve is nearly instant and everywhere across your organization all at once. The second thing that we see quite often is really around this data quality perspective, as your data quality really gets exposed and the data investment that was maybe pushed to a later date now becomes really important. AI surfaces every data weakness nearly instantly, and we've seen with our clients that maybe are having a data warehouse struggle now that they can't get two insights from their data, even though it is connected to one of those platforms like a cloud or a copilot.

We've also seen teams struggle with basic things like access rights when the folder structure and the data permissions were never formalized or rolled out across the organization. And so the problems that teams were kind of working around or adjusting are now kind of front and center with AI. The third thing is really around monitoring is all of a sudden monitoring and management of these tools is now, it's for lack of a better term, continuous. It's not just periodic anymore. New features from these AI platforms and core systems drop sometimes daily with or without knowledge from a central team. And so that pace of innovation makes that set and forget kind of periodic management basically nearly impossible. And then last but certainly not least, seeing regulation moving kind of sector by sector. It's inconsistent. Most of the time we are all expecting it to be significantly lagging.

So what you're doing today might be okay, but next quarter or next year it might not be.

In this case, an ounce of prevention is really worth its weight, but nobody has a crystal ball. So it's very hard to know what you have to do in the future. And the thing that is really interesting is there are probably 10 others, especially of you that are a little further along in your journey that you could list, but these aren't separate problems. They kind of hit all at once and compound on each other at all the same time. So would love to kind of turn back to the panel for a second and talk about what is different now that these all things are happening simultaneously and also faster. The speed is really, in some cases, a little bit concerning. Where have you kind of felt that pressure the most in previous transformations? And I'll go to David first for this one.

David Fahr:Yeah, thanks Jen. I think for me, what's kind of near and dear to my heart is the data topic. So based on my experience, whether it's ERP, CRMs, other digital transformations, typically master data and historical transactional data can blow up the timeline or scope of a project. So AI hyperscales that and we all kind of know the phrase garbage in, garbage out. And what's really new here is that the speed and scale is that AI just does a better job of distributing that garbage. So I believe we've all kind of experienced hallucinations. And so when AI hallucinates on bad data, it can go really, really wrong. A lot of organizations think they have an AI challenge when underlying it, they really have a data challenge. So it's really hard to go super fast when your data's just not in a good spot. So that's my perspective.

Jen Clark:That's great. Danielle, how about you?

Danielle Keller:Yeah, thanks Jen. To me, the speed at which AI is being adopted by users and organizations is mind blowing. I know you kind of spoke to that. And on top of that, the pace at which AI is changing is exponential. If you think through new model versions, new vendor terms and conditions, new use cases from your team, the list is endless. And so for me, from a risk perspective, it's finding a way to put controls in place that can keep up with the pace of AI and help manage that risk associated with how fast AI is moving. That's really what feels so different in this case compared to previous transformation. It's just transformations. It's just the sheer speed and adoption and change that's happening all as we speak.

Jen Clark:Yes, definitely. And completely agree. Elena, how about you on the people side?

Elena Legendre:Yeah, absolutely. So I would like to talk about kind of where all of this actually lands because it's not landing evenly with our people. So I think about how I spent about 10 months teaching supervisors, like 1500 federal supervisors and managers live while the policy underneath that curriculum kept changing as it was being taught. And as I'm going through this curriculum, none of their questions were really necessarily about the content. They were more concerned about how do we lead people through something that they don't even fully understand themselves? And I think that's the same position that managers are in now. So they're put in the middle every single time. We have our executive team that's looking at strategy and then our front line is looking at tasks, getting the work done. And then you have that supervisor or manager kind of caught in between the one who has to answer, what am I telling my team on Monday about something that changed on Friday?

And I found that this group is typically and consistently the least supported group in any program that I've worked on. And we kind of keep designing it that way, not on purpose. It's just what happens. And then to build on that a little bit too, to go to the question about what's new, I think this regulation and data quality piece is important because we're starting to ask managers again to do some hard work here and review work that maybe they can't actually check the same way they always have. So reviews may have always been something that someone handed you, you know roughly what to look for, right? You have a checklist to go through in your head where the soft spots usually are, where people kind of go wrong, what to look for first. Now someone's handing you a deliverable, a document, something that requires your review that took them 20 minutes that used to maybe take them a day or two.

You don't know how it was made. You're not really sure what to look at. And it's not a competence problem. Our managers are good people. They really do want to learn and embrace what's new, but they have to learn what to check now. Nobody knows yet though. We're all learning. So in 10 months of helping those supervisors, it was an interesting question to help them answer because it's not something that you really fix with training. It's something where you have to kind of redesign and training isn't necessarily going to fix it. And I don't see a ton of organizations doing that work yet, but that's what we're here to help you with.

Jen Clark:That's great perspective. I will even add on that, that in the scenario that you talked about, our reviewers and our managers and our subject matter experts, that's when a human may be co-producing something with AI and agentic autonomous output complicates that picture even further. So to kind of keep going down around, not to be a total downer around failures and what goes wrong, but really we wanted to make sure that we called out a couple of things that we see consistently as places where organizations trip up a little bit, especially during this rapid overnight sort of like almost all at once transformation. And the first is really around ownership, and this is multilayered. AI, as we've just talked about, AI lives inside every function's core work. It's in every leadership role. It is now in every kind of end user role as well. When accountability is really unclear at the top, it's also unclear everywhere else.

Everybody is just kind of looking around going, "How does this work? Who owns this?" We see this in our engagements all the time. Leadership teams often think that assigning kind of a single person for this responsibility solves the issue, but AI really requires a complex matrix of accountability. Who is responsible for new tools and use cases? Who is responsible for a bad output? Who is responsible for value and cost? And as we just talked about, agents make that even harder, especially when you're talking about autonomous workflows. We work across multiple industries. This is often the hardest solve because again, it often requires some type of redesign. I'm redesigning part of my role, I'm realigning part of my responsibilities, or in some cases I'm blurring the lines between two departments that were very clear before. The second is really around that fragmented governance, so not having a shared framework and language around controls.

So without multiple teams executing through a unified policy, it compounds your risk. We really truly do see governance as your guardrails and the foundations that allow these teams to experiment and move really fast. Governance isn't just something that you're going to be audited against. It's also just how the rest of the organization participates and keeps everyone safe. And then finally, again, the last but not least, people are such a huge part of this. I often like to say that this is not a technology transformation. It's usually a people and process transformation. Adoption really follows usefulness and trust, not leadership mandates. It's not optimizing your tokens, not trying to get maximum usage. Without deliberate design, your transformation is going to stall. And change shows up in very unexpected ways for every different type of organization, as David kind of mentioned earlier, of every organization has its own culture.

We hear all types of different stories about how clients are coaching employees to really kind of lean into their integrity and not overstate their abilities because it's very easy to seem like you're an expert when you have something like a Cloud or a Copilot or a ChatGPT. Two, we also hear things like there's a real pressure to have every single output be polished, which can then lead to overall alliance on the tool. So change is really complex and unique and requires time and attention. So again, I'm going to kick it back over to the panel to keep the discussion going. Which of these things do you think organizations underestimate the most and which one almost no one might see coming and it kind of catches them by surprise? And I'm going to go to Danielle first on this one.

Danielle Keller:So first, I think all of these are great choices, but being the cyber risk person on the panel and coming here to talk through governance, I feel like I've got to go with fragmented governance here. Fragmented governance is quiet. And what I mean by that is that most teams feel like they are doing the responsible thing and they're trying to follow the rules and do things responsibly. But in the fast-paced world of AI, people have to make decisions quickly. And if you don't have a strong enough or agile enough governance structure to be able to handle it, a lot of risk gets introduced into your organization and you may not realize it until it's too late. And so I think that is something that organizations aren't really underestimating.

Jen Clark:That's great. And completely agree. Elena, how about you?

Elena Legendre:So I'm going to disappoint everybody and not pick my own box here because I think the one that's most underestimated is ownership here. And I think that change management is very important and it's underestimated for a lot of different reasons, partially because of the way it's been marketed for the past 20 years is it's all communications and training. So it turns up in a budget, maybe looking like overhead or something that can get cut. But I really think that ownership is such an integral part of change management and of it being successful. And I really think that what nobody sees coming isn't even on this slide. So to give you an example, I love telling stories. A few years ago, I was brought onto a state child welfare program to help restructure their change management plan. Keyword there being restructure, right? There already was a plan.

It was written, it was approved. It was in the project library that everybody likes to keep organized and never really look at and it wasn't working. And that's a failure that everybody can probably relate to, but you may not even realize, right? Well, I have this plan, I have this policy, it's written, there's a document, it's got to be working. And with AI, it can get a little bit worse because policy that really nobody's read and it's changing all the time and the actual behavior of using AI is happening somewhere that you can't even see it, can't really monitor it. So the problem is not so much that transformation is stalling here. It kind of looks fine. Your plan exists, your licenses are being used, nobody's complaining, but not any one of those things are really telling you whether anything has changed. So I think starting to measure in different ways is also really important here.

Jen Clark:Totally agree. That's a great point. And like you said, it's not explicitly called out, but knowing what success looks like and being able to measure is really important as well. David, last but not least, anything to add?

David Fahr:Yeah, I agree with Elena 100%. I think ownership would be my vote here. I think it's a consistent thread through all the projects I've ever done. Projects can still go sideways with great ownership for a variety of other reasons, but they cannot succeed with poor ownership. So having good ownership, good engagement at the executive level alignment from an executive perspective is clearly is so key. AI throws just a big wrench in this topic because the ownership topic is even more cloudier. One of the biggest mistakes I see is organizations asking who owns the AI platform. And it's always, well, IT owns AI, IT owns AI. And this is about business outcomes and AI is so, its tentacles are in all parts of the business. And IT can help enable, govern, support the capability, right? But really the alignment with the business function and good ownership with the business team is really how you drive the value in the good leadership going forward.

Jen Clark:Totally agree. Thanks David for that. All right. We're going to go to another poll before we continue our conversation. So this poll is really around which transformation risk is the most underestimated in your organization? We'd really like to hear from you. Again, I think our panelists have done an amazing job kind of setting up the real tensions and problems that organizations face and feel as AI has really kind of taken a hold in most departments and like we said, kind of frontline work. But everyone feels these pressures differently going back to our point around we're organizational culture. But once everybody has a chance to answer the poll, we're going to actually turn our focus to some solutions and how to set this up right, knowing that this is a constant and ongoing job that all of us are going to face for multiple years. Awesome.

So it looks like the leading answer is all three, and that's not surprising from what we hear from our clients all day, every day. So as I said, so now that we've set up that tension and the problem that most people feel all the time, really going to turn to some thoughtful solutions. And again, you can hear it in all of our answers, it's like we believe governance grounds everything. So we're going to start with Danielle first.

Danielle Keller:Sure. Thanks, Jen. And I will say the second answer in that poll was fragmented governance. So if we're keeping score here, I think I've just beat Elena and David, but we're not, we're not. Okay. So we're here to talk through designing AI or agile governance models. So the word governance makes people brace for red tape often, which is fair because most governance models were built for a slower environment, think annual reviews, quarterly steering committees, things that are done on a very scheduled basis. The pace of change used to match the pace of governance. What's shifted now is that boards are asking about AI use in the same meeting that they're asking about financial controls. Regulators are starting to expect answers in areas that most organizations haven't even though of or even drafted a policy for yet. And then I guess additionally, internally, business units are standing up their own pilots faster than you're able to even inventory them.

So the pace of change has increased significantly and your governance model needs to adapt to that. So I'm going to cover six areas that can help you make that shift to a more agile governance structure. So the first is going to start with clear decision rights. So this sounds obvious, but it's one of the most common gaps we see. Who actually owns the decision to approve a new use case or shut one down when it's not performing the way it should? If you can't answer that in a sentence or two, you don't have clearly defined decision rights. You have a gray area and that's really where things can stall out or go wrong. You don't want to find yourself in a situation where let's say you have legal, IT in a business unit who all believe that they have the final say on the same decision.

So defining that early is critical and that can be done as simple as a one-page racing matrix to help with that.

The next is going to be tiered risk escalation. So not everything needs to go to the top, and I think people know that. So if everything did, nothing would get the attention that it actually needs. So escalation speed should match the level of risk associated with it. Thinking of this from an AI perspective, a chatbot that's answering FAQ questions and then a model touching regulated data should never be sitting in the same decision queue for one person to look at. So if you think through tiered risk escalation, most organizations that are able to implement this right land somewhere amongst three tiers. So you have your low risk items, maybe those can move through a documented self-assessment. Moderate risk items get designated a reviewer within a set number of days. And then your high risk items, think anything that's touching regulated data, informing financial decisions or potentially affecting customers can get escalated to a committee.

Anything more than three tiers, I mean, I guess you could go with four tiers, but I think anything more than that, people are going to be spending too much of their energy figuring out which tier they're in.

The third would be steering at the right altitude. So this is really just about having the right people involved at the right time. I think it kind of speaks for itself, but you need a cross-functional group meeting at a cadence that matches how fast the work is actually moving. So for most organizations, rather than meeting quarterly, it might make sense to actually meet monthly or even maybe biweekly for fast paced and critical initiatives. And then the other side of this is not just the cadence of how you're meeting, but it's who's in the room. You want to make sure you have people who can actually unblock something that day if needed, not this broad committee that doesn't have any decision makers in the room that you then have to go find and have another meeting after the meeting.

Fourth option here would be, or not option, but way to get that agile governance model is going to be pre-approved answers. And I think this is a favorite one because it's an easy win and everybody likes an easy win. So if you have the same questions that keep coming up, can we use this vendor? Can we use this type of data this way? Those are candidates for pre-approved answers. Let your common scenarios move without another meeting and save those more complex scenarios for the meetings. This might look like building a decision tree. So three or four questions, and if the team member lands in a green zone, they're cleared to move on without escalating it to a higher committee.

All right. And our fifth one here would be embedded controls and owner. So controls that live in a policy binder don't proactively stop anything. Controls that live inside a workflow, so an automated access review, a data classification check that's built directly into an intake form, those actually work because nobody has to remember to go do them separately because I think we're all busy and so you just forget the controls, but if it's embedded within the process, it's a much smoother transition. The other piece of this is that each of these controls needs a name attached to it, not a department. So IT is responsible for this isn't true ownership. I can't tell you the number of times that I get that response where if you ask who owns this or who do I need to go to and it's a department or a very large.

So how can you implement these embedded controls and owners? I alluded to this kind of earlier, but think of an intake form. When someone requests a new AI use case, use the form to force an initial data classification answer before the request can even be submitted. That also helps with the risk tiering based on the self-selected data classification. And then based on that answer, the intake form can automatically route to the correct owner for review.

And then last would be real-time reporting. So leadership needs to see what's happening now, not in a retrospective three months later. A dashboard showing current status and open exceptions is much more actionable than a quarterly deck because with how quickly everything's moving now, by the time you get that deck, the thing that you're worried about or that you should have been worried about has either already happened or it's resolved itself, whether that's good or bad. Maybe it's good if it's resolved positively. Quarterly reporting still matters for trends and historical purposes, but it shouldn't be the only lens that leadership has.

So I guess to wrap this up, because I want to be cognizant of our time, these six things are intended to compliment each other and reinforce each other. Clear decision rights make tiered escalation possible because you know who things actually need to get escalated to. Pre-approved answers only work if the controls behind them are actually embedded in the process. So if you're wondering where to start, I would start by defining ownership and clear decision rights. It's the fastest to implement and everything else depends on it. The tiered risk escalation, the dashboard, those might take a little bit longer and that's fine, but get the ownership and decision rights nailed down and the rest can build upon that. And with that said, I'll pass it back to Jen.

Jen Clark:Awesome. Thanks, Danielle. Great overview of making sure that you have a strong framework of governance that can keep up at pace. So now we know that no structure works without people coming along with it. And with that, I'll turn it over to Alina.

Elena Legendre:Thanks, Jen. Yeah, I think one thing we can agree on is that change happens a lot. The change is always different. People stay relatively the same. So I'm going to talk through a couple of numbers here that are on the screen. If you're familiar at all with change management or Prosci, you probably have seen this 70% figure before, about 70% of transformations failing to meet their objectives. But I think that the numbers on this slide that are a little bit more interesting and worth talking about are the 13% and 88% here. So you can see with excellent change management, you can flip the odds here. So the 70% will tell you the odds, but then the 13% and 88% here really tell you it's a choice. It doesn't have to be your fate. If you invest early and often in change management and transformation services, which means investing in your people, you can really flip the script on this.

So I'm certified in the methodology behind these numbers, but the more useful version of this is really what this looks like from inside of a program. And I can tell you from experience that within the first two or three weeks, I can tell whether or not that initiative is going to land and it's hardly ever the technology. It's whether ownership is so obvious that when I ask who's responsible for this, where does the buck stop? Nobody has to stop and think about the question. So just what Danielle was talking about. So for instance, on that healthcare data program that I was involved in, we had a role there and it was a release train engineer. So you have 175 people on that program. And the whole reason that role exists is so that there's accountability for flow across all of those teams and people and it's not ambiguous.

So every single one of those 175 people knew where a block dependency went. There wasn't any kind of guessing. And when a role like that is vacant or it exists on a chart somewhere and no one's really doing it, everything downstream gets slower, but then people have a hard time really telling me why. So the single best predictor I've seen across every program, again, isn't the methodology, it's not even the budget or the communication plan, it's whether one senior person is personally visible in the work. And I mean that in a really specific way. So they're showing up to working sessions, they're not just in the steering committees, they're using the thing or the tool themselves, and they're saying out loud in front of their people why this matters and what happens if it doesn't land. So I think that's the highest leverage move available to everyone on this call today, and it costs just a few hours a month.

And it's one of these things that you can't delegate, right? The moment you hand that over to a manager, it stops being the thing that works. So on our next slide here, I want to talk through a couple of disruptions where we can talk through how this shows up. So being from a legacy firm in South Louisiana, we are no strangers to natural disasters and how we respond to a lot of disruptions, but we've also gone through others. COVID, work from home, we've done it all. I'm just going to pick one of these as an example here. So when we were doing the rental relief during COVID, that was federal money for a county program. People were calling because they're about to lose their housing, right? So talk about tension, lots of emotions. The eligibility guidance kept moving. So what documentation they needed, who qualified, how far back you could go.

We had about a 15 person call center that had to give really accurate answers to every person who was calling in crisis in the moment, and that correct answer kept changing underneath them. So what that really taught us is that you can't train the content. You can to an extent, but much like Danielle was saying earlier, that content is wrong before it's even reaching the person who has to deliver it. So we really had to stop focusing on training so much of the content and train on some other things, getting our people to be proactive. Where do I need to go to look and how do I know whether what I'm looking at is the most current and verified information? So they needed to go to one authoritative source that was updated once and then kind of cascaded everywhere. We retrained weekly, not quarterly. We had to find a good cadence.

So we call those weekly ongoing trainings. And that really helped our people stay up to date on the model as it was today. So think about AI too, that's changing daily. How are we training people? That's not the most effective way to prepare and equip our people. So we really focus on training them on how to verify and where our current guidance lives rather than delivering training constantly. And on our next slide here, I wanted to share with you some things that I hear often. I'm not going to read all of these to you, but I'll tell you kind of where they came from. So either they've been said to me or in a room I'm in or to colleagues, and each one kind of gives you a clue as to where something might be missing in their organization to really help that transformation stick.

So for example, the one here that says, "Honestly, no one really owns this and it's starting to slip." When somebody says that to me or some variation of it, I know that I don't have a change management problem yet. I have a sponsorship problem, that ownership piece that we keep talking about, and that'll eventually become a change management problem if it's not handled early. And here's a little psychology tip for you for fun, that word honestly at the front. When people say that, especially to start their sentences, that means they're probably finally admitting something out loud that they've known for months, and it's usually the most useful sentence in the room because they're finally admitting it. It's probably been bothering them for a lot longer than they'd like to admit. Another one on here that's very, very common, we just don't have enough people to actually do all of this.

That one is usually true and I don't really try to argue with it. And what I've watched happen is that when you push past what an organization or company can actually absorb, you're not going to get slower progress, you're going to get failure across their whole portfolio, including things that may have been going really, really well. So the answer is not encouragement like, "Come on guys, we can do this, pat on the back, we've got this." Somebody with authority really has to take something off of that list. And as you may know from experience, if nobody will, the organization's going to decide for you. Your people will start making their own choices even informally and you don't get to choose which one of those drops. So it's really important to be on top of that as leaders in the room today. That's just scratching the surface on change management.

I could talk about it all day, but I'll hand it back over to Jen for our next polling question.

Jen Clark:Awesome. Thanks so much, Elena. As Elena mentioned, we'll go to another polling question before rounding out our panelist's perspective. So what is the biggest barrier to your transformation today? We'd love to hear where you're feeling the pressure. I think building on what Elena was just talking about, AI takes time and investment. I think what you're hearing from all of us is that despite some of the market hype, it's not a silver bullet in a lot of cases. It does require kind of thoughtful ownership, leadership and decision making. There's a lot of talk in the market right now about cost and value realization of AI. Without some of these basics, it's really hard to find success in those numbers and to get your return on investment. So we really believe that adopting these kind of foundational aspects really lead to that success and show up in your success measures later.

We'll give everybody just another couple of seconds to finish answering the polling question. Awesome. Wow. Okay. So really a lot of folks leaning into kind of that adoption, training and change, which is probably not a surprise again to our group. Awesome. So again, last but not least, almost every road often leads to IT, whether or not that's a responsibility or a blame. So we'd love to hear from David about how AI is changing the role of IT in organizations and how to do it well.

David Fahr:Thanks, Jen. Yeah. I can't imagine an organization that is a department that is more impacted by AI than your traditional IT department. And so look, at the end of the day, IT, CIOs, directors have to evolve and it starts with speed. You can't be the gatekeeper. You can't be the no police. You really have to become an accelerator at the end of the day. Now, that's enabling safe experimentation and maintaining appropriate controls. And if you're not an agile organization, this is going to force you to become an agile IT organization. If you're doing traditional waterfall for portfolio management and project management, that's kind of going out the door at this point. The other thing I would say is if your IT organization is not already embedded in the business, then you are already behind. You need to know your business. Do not sit in an ivory tower and try to lead IT.

You need to get in the field, you need to visit customers, you need to go to manufacturing plants and warehouses and really get into the shoes of all the users and employees in the organization, as well as your customers and how the organization makes money. The most effective IT organizations are ones that are, they're going to be integrated with the business. They're true partnerships, right? They have a seat at the table at executive steering committees and boards and executive team rooms, and that's really where you want to be. And I have to talk about data. I talked about data before. Really, you need to start treating your data as a strategic asset. It needs to be organized. It needs to have people supporting it. It needs to be centralized as much as possible to keep it clean because now AI just exasperates the challenges and the benefits with really good or bad data, depending on how it goes.

And you really need governance that doesn't slow you down. What you don't want IT to be is a wall. You want to set guardrails like, "Hey, I'm good with you being on this side or this side, but when you get out there, you're going to get a call from me." That's really the way IT needs to begin managing things going forward, encourage experimentation, have sandboxes, have it all secure, but really support experimentation within the business. And then the last thing I would say is just IT needs to evolve to a completely new operating model. AI is not a project. So if everybody is treating AI like a project, you're kind of already behind. I think most people did initially to say, "Hey, we got this project. Let's spin up AI." At this point, it's becoming ingrained in corporate culture. So really your IT organization really needs.

They can't be keeping the lights on. They have to be enabling the organization in leveraging AI to be more efficient, drive growth of the organization. Look, right now it's tough to be a CIO. It's one of the toughest positions. It's been that way, but it's really tough. But at the end of the day, IT organizations are going to have to adapt and evolve. That's just the way it's going. Back to you, Jen.

Jen Clark:Thanks, David. I completely and totally agree. I think our CIOs are in a unique and pressured environment, and really adapting to these new models can really put them in a position of power and influence. So kind of wrapping this all together in a nice little bow is thinking about. We've covered a lot of perspectives here, but what do we think about kind of top actions that leaders should take right now as sort of your quick wins, as Danielle mentioned earlier. It really comes down, this panel really thinks about these top five things, right? Of course, you're going to hear us talk about ownership. It's so important that you have that really complex accountability across your organization. You've got to build the right foundation, both from a governance and a data and IT operating models. This foundational work can begin in parallel. It doesn't have to come first, but really, really important to spend some deep thinking time there.

You've got to agilize, for a lack of a better term, your controls. The pace and speed does not allow for this periodic governance model. You have to shift to this real time environment. You've got to establish those strong frameworks way before it's needed, as Danielle listed. And then finally is just, again, last but not least, change management is ongoing and constant. It needs to be involved kind of first thing or even earlier than that, if possible, and needs to stay throughout the entire life cycle, not just bolted on at the end. So final kind of panel lightning round. If organizations could only prioritize one of these, where would you start? And I'm going to go back to David.

David Fahr:Yeah, Jen, I think I'm going to stick with ownership. That's going to be my theme. I've talked about culture and ownership. If you don't have good executive ownership, authority to make decisions, allocate resources around the troops when things go wrong, then governance, change management, adoption is all just going to struggle. That's my vote.

Jen Clark:Totally agree. All right, Elena, how about you?

Elena Legendre:I'll go with number five this time, my bucket, but I think doing this early isn't just about getting people comfortable, but it's one of the main things that's going to generate all the data that you need to make all four of these other decisions to know whether or not that adoption is real, where it's failing, and whether you can safely go faster. And if we're starting it really late, then we're kind of just doing it on vibes, as they say.

Jen Clark:And Danielle, last but not least, how about you?

Danielle Keller:Yeah, I'm going to stick with my AI governance number four, but I do think that also includes number one, establishing ownership. So I kind of get two for one there, but building an agile governance before you're forced to, because ownership, controls, and change management all need somewhere to plug into, and governance is what that is. That's what tells you who decides what and what is acceptable even once things start moving quickly. So I'm going to go with number four/number one.

Jen Clark:Awesome. All right. How we can help. So this team, obviously we think about AI success and transformation as a four-legged stool. It is really important to have all aspects of this to transform your organization thoughtfully. You have to have IT along with governance, along with change, and a deep understanding of the technology in order to make real progress. We deliver in a variety of engagements, both from just light touch advisement to medium touch, getting your programs launched or accelerated to the next milestone, especially as you're thinking about scaling. And then heavy touch is really that end-to-end delivery where we can be embedded with your teams. So one last polling question. We'd love to hear from the audience of which area would benefit your organization the most right now, kind of bringing you into the conversation and thinking about how this change, what would accelerate you, especially as all of you are in different kind of maturity stages, likely again, sort of need different things at different places, but interested to hear directly from you all.

Give just another second.

All right. Daniel, not to say that there's a winner, but I think Danielle moves around with AI governance design and controls, but with change being a close second. So to round us out and end the conversation as pulling it all together, you will hear us say this over and over again, the organizations that will succeed in this AI transformation will have clear ownership, this agile governance model and integrated change from the very beginning. And the goal isn't to do this slowly. The goal is to rebuild your organization in a way that can carry the pace of innovation because it's certainly not slowing down. So with that, just with a couple of minutes left, we're going to take a couple of questions. I see a couple of questions in the chat. The first one is really around regarding ownership, speaking to the risk of allowing employees to identify and use individual free AI tools versus the organization deciding what larger tool.

I will answer very quickly and then would love to hear from the panel. I would say that almost always we recommend some type of enterprise protection. It's just so hard to control, especially sensitive data types and IP, what is getting outside of your bounds. And so really making sure that you've got clear protection and clear kind of acceptable use is essential in not letting the toothpaste out of the tube, so to speak. But Danielle or David, anything to add there?

Danielle Keller:I would completely agree. I mean, it goes back to the governance, but you need to protect your data and putting that governance out there to say what your team can and can't do. I think this is something that you would certainly want addressed in there. So I completely agree, Jen.

David Fahr:Yeah, just quick comment, get away from using personal AI tools in an organization. It's wrought with risk and failure. We've seen it. Get on enterprise tools. We can help you do that if that's what you need.

Jen Clark:Absolutely. Oh, one last one and then we'll close this out. Addressing employee concerns that AI could replace their job, kind of basically making it take longer that employee will grab on and adopt. I think from my perspective, but also would love here Elena's, I think what I always call measured communication is employees are going to very easily see through platitudes and having clarity around those success measures of why you're placing AI. Again, not that top down mandate of where we think it's going to be successful and then that senior leadership aspect that we're talking about. But Elena, anything else to add there?

Elena Legendre:Yeah, I completely agree, Jen. I think People tend to, of course, fill silence with the worst case. So they might hide what they're doing or hide their concerns. Rumor mill starts churning. I think it's important for, again, that person in that senior sponsorship spot to be really vocal about what the goal is and talk about what the guardrails are around usage, what they're expecting out of people, and then kind of incentivizing the usage as well to really get them excited about it. We've had agent-a-thons here and it's been really fun to see how creative people can get and it just energizes it in a different way.

Jen Clark:Totally agree. Well, the very big, huge thank you to my panelists, Danielle, Elena and David for their perspective and leadership. As we said, the goal isn't to move slower, it's to build the model to allow you to move fast safely. So I'm going to pass it back to Bella.

^{Transcribed by Rev.com AI}