Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
AI at a crossroads: The state of the industry on trust, leadership, and execution
248 Plays21 days ago
In this episode, Phil Fersht is joined by members of the HFS Global Advisory Board to explore why AI has reached a critical inflection point for enterprises. The panel features Malcolm Frank, Steven Hill, Mark Hodges, Cliff Justice, Mary Lacity, and Jesus Mantas, bringing together decades of leadership experience across consulting, academia, and global enterprises.
Together, they discuss the growing gap between AI experimentation and real execution, and why trust, leadership conviction, and organizational design are now the biggest barriers to scaling AI. While individuals are rapidly adopting AI in their daily work, enterprises remain slowed by process debt, siloed structures, and fear-driven decision-making.
The discussion also examines how AI is reshaping services, workforce models, and business outcomes, and why success will depend less on technology and more on culture, accountability, and the willingness to redesign how work gets done.
To watch the full webinar, visit here: https://www.hfsresearch.com/webinar/ai-at-a-crossroads-the-state-of-the-industry-on-trust-leadership-and-execution/
What you’ll hear
• Why enterprises are struggling to move from AI experimentation to execution
• The AI velocity gap between individuals and organizations
• How trust, fear, and leadership behavior are slowing adoption
• Why process debt is a bigger barrier than data or technology
• The shift from effort-based services to outcome-driven models
• How AI is reshaping workforce structures and middle management
Key takeaways
• AI challenges are more about people, culture, and leadership than technology
• Trust in AI remains a major barrier, especially at senior leadership levels
• Process ownership and accountability are critical to unlocking value
• Enterprises must redesign workflows, not just layer AI onto existing systems
• Services firms must move toward outcome-based models to remain relevant
• Organizations that act with speed and conviction will outpace competitors
Chapters
00:00 Introduction and panel overview
00:20 Meet the HFS Global Advisory Board
02:20 The AI velocity gap in enterprises
04:17 Why adoption is slowing inside organizations
05:02 Trust vs capability in AI adoption
06:41 Lack of strategy and enterprise readiness
07:59 Building trust and securing AI systems
09:07 Process debt as the biggest barrier
10:40 Leadership, culture, and change management
12:46 Overcoming fear and driving adoption
14:42 The shift to services as software
16:48 Industry disruption and business model change
18:28 The future of services and global impact
21:09 Key takeaways from the panel
24:50 Closing remarks
Recommended
Transcript
Introduction to AI's Current State and Trust Issues
00:00:00
Speaker
my name is phil first and today i'm very excited to be joined by our global advisory board we're going to be talking a bit about ai at a crossroads so what's the state of the industry what does it look like what are the big issues that we're seeing right now in terms of trust and execution malcolm if you can kick this off i've been in this industry 35 years now and been fortunate enough to be with a lot of fast growers that have ridden different waves from any client server to offshoring cloud Next up is Stephen
Panelist Introductions and Expertise
00:00:27
Speaker
Hill. I'm a former vice chair of KPMG currently working with firms like Oak Trust Group who are at the intersection of AI, cybersecurity, and data. Someone else who I've known for over a couple of decades, Mark Hodges. I knew you were going to point me as soon as you dated that.
00:00:44
Speaker
Mark Hodges with Acresist. We work with founder owned tech companies, usually 5 million to 50 million in tech and tech services. Somebody who I met not long after Mark, Cliff, Cliff Justice. 30 years in the industry, joined KPMG and then acquired Equitera and integrated Equitera into KPMG and so operated that business for a while and then took over for Steve as the head of innovation at KPMG. I retired at the end of last year so now I'm pursuing independent investment startups and disruptive opportunities in the professional services space. Next up is a dear voice in industry, Mary Lester. How are you doing, Mary?
00:01:23
Speaker
I'm awesome. I'm coming to you from our campus at the University of Arkansas in Fayetteville. And I'm a teacher and a professor, and I've been doing research, studying all you guys for over 30 years. And my research and teaching really focuses on how to big, old, traditional incumbent enterprises get business value from AI. And last but very much not least, Jesus Mantis. I graduated from...
00:01:50
Speaker
IBM, I was leading the business consulting unit. I was there for 23 years. What I'm doing recently I'm in a few boards. So I'm in the board of a company called Biogen, a company called International Flavors and Fragrances, a company called Neurofab, a company called DVLP Medicines. And the most recent one is in the board of Patrick J. McGovern Foundation, which is one of the largest funders of ethical AI applied to societal problems.
AI Velocity Gap and Enterprise Integration Challenges
00:02:17
Speaker
Let's get in then to the first question du jour. And my favorite topic is the AI velocity gap where fear is your biggest competitor.
00:02:25
Speaker
So as individuals, we're sprinting ahead because we can tolerate imperfection. But as enterprises, it's siloed systems, tribal knowledge, compliance paralysis. you know How do we close this gap? Because I think enterprises are going to lose their top talent if they can't really embed AI as well as we're doing in our personal lives. I think we're beating up on on our big old incumbents a little too much. They do face realities that we as individuals don't, right? They have to mitigate risks and they do have to comply with all ah laws and regulations. And probably the bad news is they didn't start when ChatGPT came out. They started a long time ago. And so you also talk a lot about data debt and technical debt, and they've largely dealt with that. So they're in a better position than to launch off and get business value from from AI initiatives. We're really finding that in the future, every leader is really going to be managing both a human and a digital workforce.
00:03:20
Speaker
We learned from one of our retailers, was like they don't micromanage their human employees, but if one of them goes rogue, bam, they have security measures. Immediately you isolate, you deny the access. And they're saying they're building the same types of things for their AI agents. What's surprising when selling to large enterprises is how siloed and fiefdoms they are and how they equate the size of their budget and the number of people to their success.
00:03:45
Speaker
So I think that the compensation scheme hasn't caught up to the business outcomes that HFS does a great job ah of emphasizing. I think the enterprises that have done well with RPA and intelligent automation, how they introduce that and how they govern it, they're going to be a lot more successful at taking advantage of AI.
00:04:05
Speaker
like I'll also ask the views of Steve Hill. You've seen a whole plethora of tech impact industry. Do you see this gap as very real, Steve? And what can we do to close it? I think the gap is real. And I think the gro the gap is very human. There's sort of three things that I think this gap is created by. One of them is this the noise in the marketplace. There's so much wash around AI that it's confusing. And I think organizations, when they're not confident that they know what they're getting into, that's a problem. Secondly, in a lot of firms, it takes 50 people to say yes and one person to say no. I think a lot of leadership teams suffer from that consequence. The third thing I would say is change management. I mean, I think people are afraid of change fundamentally. So I think those three things are really creating a speed issue with this
Human Factors and Mistrust in AI Adoption
00:04:53
Speaker
adoption. I think the next question is around this agentic tipping point, which is all about, I think, less advising and more
00:05:01
Speaker
executing, are we really facing more of a trust problem than a capability problem with AI? I've got some data here that just shows how 78% of enterprises, these are large enterprises, are really not trusting the AI. They're still keeping a lot of human interjection and supervision into the urgent tech. Yes, Seuss, could you share a little bit about how you're seeing the evolution here? This goes all the way up to the boardroom. I literally had this week, we were working on our proxy for one of the companies that I'm in the board. We had a question and I actually basically had a research agent go through ISS, whole president, a lot of paper and come up with a recommendation what we should do. The chairman of the board came back, he's like, well, I'm glad you trust AI because I don't.
00:05:43
Speaker
ah The explanation that this person gave is like, well, I i don't trust what I don't understand. and I said, well, do you get on a plane? Because I'm pretty sure you don't understand all the aerodynamics that go into that plane flying. I think there is a couple of other things that get in the way when you get in the execution side. I think there is a bias for inaction, meaning if you do something and it's wrong, you could get fired. But if you do nothing, you don't get fired. So you're basically motivating inaction. as opposed to taking a risk. And the one thing that I have i'm hopeful about, I've seen actually one company build a observability layer of AI, but they can tell, well, yeah, you're you're using AI to rewrite emails or summarize articles, and somebody else is using AI to, like, Vibe code, somebody else is actually using AI to make decisions. I think then started surfacing that factual use creates a motivation for people then to come to the other side, which is to actually use AI for
Strategizing AI Implementation in Enterprises
00:06:41
Speaker
execution? Getting people to share their personal experiences is the best practice within a company. People often don't. They feel like they've got an edge. Another thing we picked up on is just this, there's a real lack of a strategy. And you can see here across Global 2000, these are leaders in organizations. Only 14% actually have a clear strategy. Do you think this is a serious issue. It's still early for large enterprises to rewrite their whole business model and their whole operating model around something that's as transformative, especially as agents that actually carry out tasks. There's going to be things that you'll trust the agents to do that if an agent does make mistakes, it's just not as critical. So trust will
00:07:21
Speaker
be seen as a different imperative in some areas than others. Take autonomous cars or autonomous planes. You'll probably have a pilot if you have planes carrying two or 300 people for quite a long time because that's just a different calculus than a small drone taxi with one with one person. In many cases, enterprises, this is still very much a people business. And when decisions are made, somebody needs to be held accountable.
00:07:46
Speaker
And it's very difficult in these murky strategic choices to hold an AI accountable for those decisions. So let's let's get into then some of the things that are causing, I think, adoption issues. And let's start at cyber and giving agent system level access without machine speed security. It's reckless, right? So how do we architect trust? You've got to build trust in, you have to build security in by design. Unlike previous generations of technology, AI accelerates attacks.
00:08:16
Speaker
to machine speed. There are new risks that AI brings to bear. There's new attack surfaces and you know data and model drift and agent behaviors are all things that can happen to AI that you didn't have to worry about before. But where I think...
00:08:31
Speaker
It starts to get easier, which is you start at the use case and you figure out what kind of machines, what kinds of ai you need, and you go from there. Mary, I know you're doing lot of research into this area.
00:08:43
Speaker
How are you seeing enterprises getting exposed when it comes to like agentic investments One company I was just dealing with last month said that their board kept yelling at their CEOs saying, what's your, you know, where's your AI? Where's your AI? Where's your AI? And arguing for an AI first strategy doesn't really make sense to me. It's like, what are you trying to accomplish as a business? And then what are your toolkits of AI? Just maybe one of them. Let's get a bit more into these like debts. We try to understand what is holding enterprises back from achieving their goals. The biggest issue is process debt before everything else, even data and people. We think like a $10 trillion dollars debt pile that companies are suffering from. So how do we start to eradicate these debts? The main question to ask, is there
00:09:27
Speaker
clear and accountable process owners that have the competence and the mission to reinvent those outcomes by whatever mean. And by whatever mean it could be, they can outsource them, they can improve the technology, they can do whatever. In my experience, the technical debt is a consequence. There really isn't an established governance around what their process and process outcomes are. And because of that, that actually results on technical debt. I'm working with a couple of companies right now. They're going through their ERP upgrade cycles. And that constant is still true, is there is no real benefit to upgrade technology if there is no process ownership. So what I recommend the companies start with what Steve Jobs used to call the single accountable individual. Just one person that says, you're going to redesign account receivables and it's you. And that's it.
00:10:19
Speaker
And you make all the decisions. You can bring everybody.
Leadership and Workforce Integration in AI Transformation
00:10:22
Speaker
You can do as many design sessions as you want, but you're going to reinvent by all means necessary in that context. If not, you you end up with another three-year ERP upgrade cycle. Yeah, I've seen those cycles even longer. Let's take this to the debt that we didn't discuss. Steve, you've got a big view here that it's really the leadership that holds us all back. It's about conviction. Change management was one of the biggest challenges underappreciated needs of real transformation around ai A lot of people are afraid of AI. They're going to think that it's to take my job.
00:10:53
Speaker
I think leadership has a big responsibility to focus on how to build the right messaging and how to build a journey around AI that is responsible and trusted in the organization. Getting the right use cases put to the table first that aren't as disruptive, showing people that AI can help them in ways that is extraordinary. People are leveraging AI every day today in their personal life. So on their phones, they're using AI. They're going to chat or to claw to leverage that momentum to help the organization understand that that kind of impact can happen at the enterprise level too. it I love what Steven just said. I'm going to use the term culture too. have to say, i think Walmart is an example we can all look to on how you have good leadership in this context. For the longest time, their tagline has been that they are people-led, tech-powered, and they think AI is going to transform nearly every job. And they've offered AI training to 1.6 million of their workers. And they've already deployed four super AI agents, one that is customer facing. So in that context, when you say this is going to be in service to humanity, AI is in service to humanity, not the other way around, you really help take the fear away. And the other thing they've also done is that for many of their leaders now have KPIs for eugenic AI use that will be part of their annual performance evaluations. The other thing is their investment that they had made in good data
00:12:20
Speaker
um I think the first egenic AI case study we did at Walmart was published in the Harvard Business Review in November of 2022, where they were using AI agent chatbots to negotiate contracts with their tail end suppliers. I mean, that that's a long time ago in the world of AI. I'll just put something out there. i think we all need to reframe this issue and we have to stop protecting jobs and start expanding value. And that's how you're gonna win. We hear it every day, the machines are coming for our jobs. I'm not gonna insult anybody by saying that fear isn't real. Of course it is. But every time we faced a productivity revolution, the people who thrived weren't the ones who clung to the old ways of work. They had the courage to redesign work and what it actually means. And in particular, rethinking middle management and replacing that with AI. The data is clear here, over half employees are just plain scared. but How do we get past this fear of AI to more confidence and more excitement? The firms I'm seeing that are being most successful are going after the low-hanging fruit with AI and then communicating that back
AI Adoption Tactics and Service Model Evolution
00:13:23
Speaker
to their talent. So I think of an outside-in and bottom-up framework.
00:13:27
Speaker
So what is outside-in? Every Fortune 500 spends hundreds of millions on outside vendors. Those are the areas that they should be automating very, very aggressively and doing that with their employees because that's money you're spending on outside resource. It's not a threat to your employees. It's like in gaming. When you go from like looking at the map to the first player view, you have to look at this from a first player view of each employee. In my prior firm is we actually came up with a framework that had three simple premises. We call it make it clear, make it easier, make it worth it.
00:14:01
Speaker
So every single quarter, we would make clear what the desired behavior change was. The second one is you make it easier. Human behavior is easier to nudge when you give the human a choice architecture that is A, B. So if if A is don't use it and B is use it, make B easier to do than a And then the third one, which is make it worth it, because in in many companies it's still like, hey, you need to have like 100 people reporting to you before you get promoted. Well, do you think I'm going to be motivated to actually automate anything or will I be better not automating and having a lot more people reporting to me so I can get promoted and then is somebody else's problem? Let's get onto the next topic, which is the convergence of services and software. And you can see we had this tech vision. We actually reduced it to 2028 and 2030, but you can see the shift away from effort-based services to platform tech-based services. This is where enterprise clients. want to go. They're aspiring to get to services, software, outcome driven models within sort of three, four years. And you can even see different ways that the SAS and ERP traditional world is converging with service providers and natives to to drive this shift. You know, let's hear from, I mean, maybe Malcolm, do you want to kick this off? It's starting to play out very quickly, Bill. And we've got a parallel. We've lived through this before. We just don't recognize it as such that when you wanted to do anything 15 years ago, you had to get the Oracle database, you had to get the Sun server, you had to stand the whole thing up and then you could get started. But that was a quarter million bucks. It's going to take you a couple of months before you produced anything. What do do now? You just go up to the cloud and boom, you know, you're off and running. The same is starting to happen in services. And it's sort of perplexing to me, Phil. It's weird that when I look at all of the GSIs, most of them are treating AI as the next tech market opportunity, pretty selfishly. The way we saw it with SAP or Salesforce or AWS, they're going after it as a revenue opportunity as opposed to fundamentally changing their means of production. And clients are figuring this out. The client just hears, you're just protecting your billing model.
00:16:11
Speaker
You're not, you know, solving my problem in a completely new way. The new game is services of software. So sell the outcome, sell the result. But the GSI community is going through a classic Clayton Christensen innovator's dilemma. So if you're the client partner, and let's just say that American Express is your account, and last year you billed them $100 million, dollars well,
00:16:31
Speaker
If you went full service as software, you may bill for the same amount of stuff, 40 million. Are you going to drive that down? or are you going to go through price elasticity and start to provide much more value for the same amount of money? That is a really, really hard challenge. Most of those firms aren't going to be able to manage that change in their economic model, in their delivery model, and in the way they motivate their staff. If I can jump in there, Malcolm, that's so true. All the infrastructure built by these large firms are the very things that are not needed in this new world. There's never been a better time to start a native AI startup and go after these industries because managing that type of economic turmoil When you are a large company, it's ah it's a very challenging and very costly transition. The Industrial Revolution unfolded over 50 years. This is going to be something like five. And the impact of this is going to be 10 times larger than the Industrial Revolution. The disruption is coming and it's it's coming very quickly.
00:17:36
Speaker
and And clients who don't care about your business and protecting your business. They care about driving more value into their business. And at the same time, their business is being disrupted as well. One of the canaries in the the AI coal mine is the customer experience function where it used to be labor based call center BPOs. And it's it's fascinating. We have a client who provides advisory services to the to the buyers of customer experience and call center and the difficulties they have in coaching and training legacy customer experience providers to actually provide AI solutions without compromising their margins is incredibly difficult. So I think that's a good business function to look at to see the speed that the provider and the customers are trying to peddle keep up with that. We had a comment from Vinod Colesleva, a famous entrepreneur, where he predicted that ID services, the way it is, is just going to be dead in 2030. So let's think a bit more about, can this industry survive? Where we're going? Is India in a perilous situation because it's built up millions of jobs India?
00:18:45
Speaker
GCCs and service providers to support the way it was? It's a great question. So, ah Phil, Wall Street may misvalue a specific company for a short period of time, but it almost never misvalues an entire category. Wall Street looks at AI hitting the GSI model and it can't figure out the terminal value. You can't determine what are the five year cash flows of these firms. And as such, they're just getting out. So that's the bad news for those firms. The good news is The client needs are only going to expand. And all of these clients are going to become much more tech-centric as a result of AI. So I think the services model is going to expand. But those firms that can't switch quickly enough, they just have these slow and incredibly painful to watch. They're not even deaths.
00:19:32
Speaker
They just become irrelevant. I don't think this is a function of India. I think it's just a function of what leaders have the courage to recognize what is the new delivery model for services and move to it very, very quickly and aggressively. I agree with you, Malcolm. I think what is a irony is that as much as customers get frustrated with the lack of innovation and speed and adoption from their GSIs or service providers. The irony is that those service providers are some of the most agile animals for self-preservation, you know, decades beyond their expiration date. And many of them just keep innovating with the changes we've talked about, offshoring cloud, client-serving, they're still around because they know they know how to do that. We'll read about the ones that die because they make great case studies. You think of like Blackberry versus the iPhone, but most of them will
Key Takeaways and Future Outlook
00:20:25
Speaker
adapt. If your value proposition is still on selling cheaper hourly labor, the clock is ticking for you.
00:20:34
Speaker
If your value proposition is outcomes to your client and you're able to manage and navigate that transition from low cost labor to outcome orientation, then you're going to be successful. There are several ah high profile Indian companies, I think that are doing a great job at that, but That's not what's reflecting in these numbers. the The numbers are there are many companies out there that aren't quite as well known that make up a significant part of the labor pool that are not navigating that change very well. And I think that's what's what's going to make a dent in the overall macro numbers that we're talking about here So I'm going to get everybody to just sum up one thing that they really took away from today's conversation that they'll implement in their own post-retirement, whatever they're doing, jobs tomorrow. So maybe Malcolm, what was the one big takeaway you've got from today? I'm looking at Jesus. This is an IBM thing. Where were Lou Gerstner's book on ah teaching the elephant to dance? And we need to teach elephants to dance because...
00:21:33
Speaker
AI is moving at 100, individuals are adopting it at 10, and companies, big companies are adapting at a pace of one. That ratio is going to break a lot of companies. So how can we get that clock speed of these organizations to move faster? And I think don't boil the ocean, but find those use cases, find those where you can enable those very aggressively with AI. The firm will learn, it builds confidence, and then you can just go step by step and start to build from there. So that's my takeaway. Mary Lester, one big takeaway. Yeah, I love to collect practices that I can share with others. And my favorite one came from Jesus. Make it clear, make it easy, make it worthwhile. I love that. Cliff, what did what did you take away from Ola? Yeah, the defensive companies are always going to be on their back foot and will probably get overtaken. The companies that go on the offense are are going to be the companies that win.
00:22:26
Speaker
And these are going to be the companies that innovate, that change their business model, that drive change. It's going to be uncomfortable. The larger the company, it's the more uncomfortable it's going to be for the workforce and for the the markets. But I don't think there's ever been an opportunity greater than the one we're in right now to make these changes and shake up the market. Thanks, Claire. Mark? Some larger corporations are making some acquisitions that are basically going to be the guinea pigs for how they adopt AI for the acquiring company. So they're they're giving that new subsidiary permission
00:23:02
Speaker
to change things, break things, fail in in ways that I am not used to seeing from larger organizations. So I think that's an interesting model that you you buy to experiment and break as opposed to just absorb. Steve Howell, one big takeaway from today. I think one of the key things to making this happen in a positive way is culture. I mean, leadership has to guide this. And i think leading from fear is the wrong first step. You've got to lead with confidence and you have to lead with, ah I think, with an ah offensive strategy. Looking at this as a cultural transformation, as well as an organizational or process transformation opportunity is super important. doug
00:23:41
Speaker
Who better to finish this off than Yusus Mantis? One big takeaway for today. My takeaway is confirmation that it is not a technology challenge. It's a people, behavior, and organization design. I think the technology will work.
00:23:57
Speaker
I think is how we rewire the organizations, who does what, and then how we rewire the interest. I mean, think about that. The org charts that we use to operate most of our companies were designed like 200 years ago. it It definitely, it can possibly be the best way to organize the work. So I think my takeaway is confirmation from my fellow panelists here that I think the bigger work is if we fix that, I think everything else will get fixed. I think a lot of this stems back to middle management bureaucracy. It's held us back for a long time, but more than ever. And I've seen big agentic deals really struggle because the C-suites decide what they want to do and then they go down to their teams below them. And it's at that level that they really struggle with the change. And I think you've hit the nail on the head. These organizations aren't designed 20%.
00:24:47
Speaker
30. They're designed for 2015 or 2010. And that that needs to be a big shift. Well, this has been a terrific hour and a bit. And I just wanted to thank everybody here for their time. Really thank you all as well for being part of our global advisory board. And this is what we're meant to do is get together, share our collective news and learn from each other and share it with the industry. So I look forward to seeing many of you in new York City for our summit in May 13, 14. And thank you very much for your time today.