Building the Five Foundations of AI Value With Mobile Mentor

[00:00:04] What happens when employees adopt AI faster than their company can train, support or protect them? Well, today I have returning guest Dennis O'Shea, CEO of Mobile Mentor and friend of the show. He's going to be joining me to discuss findings from the 2026 Endpoint Ecosystem Study and also share some lessons from his company's own AI rollout. And I always like to give everyone listening valuable takeaways.

[00:00:33] So we're going to talk about the five foundations that he developed from the mistakes that he made and why Gen Z can be both an AI adoption engine and a source of shadow AI and why data, agents and spending now absolutely demand management's attention. And on that note, please join me in welcoming back friend of the show, Dennis O'Shea, as I reintroduce you to him now.

[00:01:01] So a massive warm welcome back to the show. For anyone that has missed our previous conversations, can you tell everyone listening a little about who you are and what you do? Thanks, Neil. Great to be back on the show. Dennis O'Shea, born in Ireland, lived in the UK, Finland, Netherlands, Switzerland, New Zealand. Now I live in Nashville, Tennessee. And I run a business called Mobile Mentor that's about to go through a rebrand to just mentor. And we're a Microsoft partner.

[00:01:31] And we do a ton of work with all the Microsoft tools, the security and now the AI tools and a whole bunch of the Microsoft stuff. So we've been running this business for 22 years and working with enterprises and mid-market organizations and learning every day and having fun every day. Well, we've spoken a few times now. And the first time was nearly five years ago when nobody was talking about AI. Of course, everything has changed there.

[00:01:59] And one thing that hasn't, though, is the Endpoint Ecosystem Study, which I know is very close to your heart. And there was a new one released this year. So tell everyone listening a little about that study, who you surveyed or spoke with, and what it means to you. Sure. So the Endpoint Ecosystem is a bit of a geeky term. And what it really looks to do is understand the relationship between end users and their technology in a work setting.

[00:02:25] And when we think about the Endpoint Ecosystem, we're thinking about all the devices people use. Might be their smartphone, their tablet, their work laptop, maybe a desktop. And then all the operating systems, all the applications, their sign-in experience, and all the tools and the supporting processes. So all that cluster that comes together to either make somebody productive or frustrated. And so our research seeks to understand what's going on in that space and what's it like from the perspective of the end user.

[00:02:52] And so we surveyed 1,000 people in the U.S., 500 in the U.K., 500 in New Zealand, 500 in Australia. They're the markets we're interested in. And so we look in to see what's going on, what's the state of play. And like you mentioned, five years ago during COVID, it was an extraordinary time because Gen Z had just entered the workforce. They had a horrible onboarding because we basically gave them a laptop and said,

[00:03:17] go home and stay at home and don't come back because you're not welcome here in the office and learn how to work for the first time in your life remotely with all these people you don't know. I mean, that was a diabolical time and a diabolical onboarding experience for the new workforce. And then two years ago, we were kind of coming out of COVID, trying to persuade people to come back to the office and work together and collaborate again.

[00:03:42] And now third time around, of course, AI is the frontal lobe consideration for everybody and everything and showed up really, really strongly in the research and the data. And, you know, the highs and lows were kind of seen through the lens of AI adoption in the workplace. So it's been fascinating and three wildly different responses to the study. And we learned a huge amount each time, obviously.

[00:04:09] Yeah. And obviously now we've also got a gent of KI being thrown into the mix. There's so much big changes going on here, transformational change right across the board. But AI is also exposing some cracks in the workplace. And this is something that we don't hear much about. But I noticed it mentioned very early on in your study. So tell me a little bit more about what AI is exposing in the workplace.

[00:04:35] I guess it's exposing problems that have not been addressed over the last few years or the last decade. Yeah. I kind of see AI shining this great big spotlight on things that maybe should have been but have not been resolved. The most obvious example is the data.

[00:04:52] So companies that moved all their data to online and did some data classification and applied some sensitivity labeling and got some DLP controls, they're in a good place with AI because they can allow AI tools to reason over their data without too much concern.

[00:05:09] The organizations who didn't do that and still have data locked up on on-prem file shares and don't have classification and don't have sensitivity labels and don't have good DLP, they're scared shitless now because they know that people can upload company data to a large language model or pull data down from a large language model and save it into the company data. And there's a whole bunch of issues and risks showing up.

[00:05:33] And like here in the US, we see smart people doing things like searching for W-2, which is your personal tax form that states your income for the whole year. So, you know, naughty employees would go searching for something like a W-2 or a draft employment letter, offer letter or an employment contract or a performance improvement plan or something like that. And if the data is not correctly protected and classified and labeled, those things are going to show up.

[00:06:03] And it's really tragic when it does because then, you know, the organization usually slams on the brakes and goes, whoa, stop all this AI stuff. We're not ready. We need to wind back and we've got to do all this cleanup work with our data. Stop the clock. And that takes a year or two to do all the work with data. And, of course, that's massively disruptive to the people who had just, you know, built some momentum and were leaning in and learning AI.

[00:06:30] So I'd say that's problem number one, data readiness. And something else that really stood out in the study, there was a stat around AI adoption, how it is accelerating faster than AI enablement. And, yes, while many organizations have introduced AI tools, only 29% of employees say AI delivers regular or essential value in their work.

[00:06:54] And 48% here, and that's a massive figure, report receiving zero AI training or unsure whether training even exists for it. I mean, did this surprise you? It did. It did. And when we double click on it and go a bit deeper, it's interesting. If we look at it by industry, because we tend to go deep into healthcare, government, finance and education. Those are four industries that are mandated to protect something.

[00:07:21] Patient data, citizen data, financial data or student data. So they're really interesting to look at. And healthcare and government shocked us. Half of employees in those industries get no AI training at all. None. Zero. Zero. The finance industry is the furthest ahead. They do generic AI training for everybody across the board and role-specific training. So they're very good and very systematic in deploying AI in a deliberate way for a specific use case.

[00:07:52] And maybe with some agentic, you know, some agents working as well for those use cases. And then they provide role-specific training. So finance is reporting the greatest returns and value from AI. Not surprising. Healthcare the least. And it's interesting because healthcare is probably the industry that needs it the most. And then when we look at ages, sorry, generations, we look at five generations.

[00:08:17] Gen Z, young millennials, old millennials, Gen X, which is probably our generation, Neil. And then the baby boomer generation. And it is extraordinary, the difference across the generations. Gen Z is doing a fabulous job in adopting AI. But let's put an asterisk on that. I want to come back to Gen Z because there's some funky stuff going on there. And then the millennials are pretty good.

[00:08:44] Our generation, Gen X, pretty basic in AI use for the most part. And the baby boomer generation, a lot less. So there's a very clear correlation between the generations and AI adoption. And as he said a few moments ago, the rise of AI is amplifying many of the same technology frustrations and frictions that employees have faced for years.

[00:09:07] I mean, employees are still navigating carefully around security policies just to get their job done in their own way. There's password-related risks. They remain widespread. But the future belongs to those that reduce friction. And one positive stat out there said 71% of workers believe that their company is very or somewhat open to ideas on how to improve technology. So there is room for optimism here, right? There is room. And I'll tell you one other area that came out really strong.

[00:09:37] You might remember in the last couple of times we spoke, we had really gone deep into understanding password behavior, how people manage their passwords and all the crazy things they do to manage and save and remember their passwords. And we're seeing a positive shift now where people are having fewer and fewer passwords to manage. We're not out of the woods yet. We've still got a whole bunch of bad behaviors going on.

[00:10:02] People saving passwords in the browser and in a spreadsheet and a Word document and on their phone and a notebook and all that. But we are seeing better adoption of biometrics as the point of entry to the device, which is wonderful. And in single sign-on for all the corporate applications. And of course, multi-factor is being adopted more and more as well. So we are seeing very positive momentum towards a password-less future. And we see that as one of the key things people need to do to be secure.

[00:10:31] Get rid of passwords. You know, there were a wonderful invention in 1961. But what other technologies are we still using from the 1960s? Yeah, yeah. Very few, right? It's tragic that we're still asking employees to type passwords in 2026, when we know they're the number one reason for breaches and hacks. So, you know, we're on a huge drive to help organizations go passwordless. And we're seeing some good momentum in that direction.

[00:10:59] But some of the other problems you alluded to are, you know, some of the onboarding experiences reported by employees still surprisingly clunky. And one of the things I want to come back to about Gen Z, they're fascinating. So they're driving AI adoption faster than any other generation, right? They're also the generation that's most likely to work around and bypass our security policies and systems. So, again, we slice that by generation.

[00:11:27] The baby boomers and the Gen X are actually quite compliant and well-behaved. And they don't work around company security and systems and policies. Gen Zs and young millennials, oh my God, as soon as they hit some friction, they're going to find a way around it. And the way that shows up now is shadow AI. So they've got their personal accounts going on for other LLMs. And so if they can't find a happy path to do the task with a sanctioned AI tool inside their organization,

[00:11:54] they'll just go off and sign into their personal account in the browser and upload the document and analyze the sheet or do the thing or write the thing and then pull it back into the environment. They will do that so fast and so casually that people don't even realize what's going on. So that's the generation that is driving shadow AI. And the third thing to know about Gen Z is, bizarrely, they need the most help during onboarding.

[00:12:18] So when we surveyed people and say, how many tickets do you need to raise to get all your stuff working when you joined your organization? Gen Z needs the most. So they're not as tech savvy and organizationally savvy as other people who've been in the workforce for longer. So we do need to give them lots of support. We do need to give them clear guardrails and point them in the right direction and give them the right AI tools.

[00:12:47] And there's another big stat in there that I've got to read out here. I think it was 72% of the people surveyed in the study said personal privacy is much more important than company security. Again, did that surprise you? It's a big number, isn't it? Bizarrely, again, that stat is almost exactly the same as it was two years ago and four years ago. It only moved about one percentage point.

[00:13:10] So basically, we're 70-30 people care way more about their personal privacy than company security. And that's pretty consistent across the generations. Slight skew for older people. Older people seem to be more responsible for company security and less concerned about their privacy. Maybe because they've got less of their lives exposed on social media and they're not as cognizant of privacy issues.

[00:13:36] But the younger generation hyper-focused on their personal privacy. Probably because they've exposed their whole lives on social media and they're very aware of that. So if we were to go inside any organization, let's say in corporate America, for example, what are they struggling with AI at the moment? Well, they're not struggling with deployment. They're struggling with adoption. And they're struggling with data readiness.

[00:14:02] They're struggling with rolling out good training and getting people to use AI as they would like. They're struggling with the plumbing for agents. Getting all the plumbing right, getting all the APIs, the access to permissions, being able to manage agents through their lifecycle. And they're struggling to measure and report on ROI. And, you know, Brené Brown says vulnerability is a superpower.

[00:14:29] So I'm going to share with you for a moment all the mistakes we made when we did AI. And, you know, as the chief exec, I was like the chief evangelist for AI when these tools came out first, the LLMs. And I pushed our team pretty hard and we ran competitions and we gave away some pretty meaningful prizes, monetary prizes for people to come up with the best ideas and the best use cases. And we deployed Microsoft Copilot, I think, to about two thirds of the company.

[00:14:58] And then we got ourselves in a bit of trouble. First, we had a conversation with our board and they said, right, so you've been on this AI journey. You've been spending this money and doing all this. How's it working out? What kind of a return are you seeing? And I was completely stumped. I had no answer. And if they, you know, I couldn't even have been recommended. We're going to pull it all back and stop because it was a waste of money. And that's just going to stop tomorrow. Or it's amazing. And we're going to give AI tools to everybody.

[00:15:29] And between those two extremes, I had no way of answering the question. I had no data. So we started digging in and then we realized even more fundamentally, we had not defined our use cases. We didn't have a hypothesis to say somebody in this role doing this kind of work will be able to work better, faster, cheaper if we give them the right AI tools. We didn't have that hypothesis. We kind of gave licenses to people based on seniority, I think.

[00:15:56] Secondly, we hadn't done our groundwork with the data. And this was embarrassing because when we did our first data scan, we had 33,000 data assets with sensitive data that were overexposed and overshared. Across the organization. That was embarrassing. Third, we realized that we had done some very basic training, but we were really only using the tool, the co-pilot tool in the browser.

[00:16:24] We were really weak at using it inside Excel and Word and all that. Fourth, we didn't know where to start with agents. We didn't have any plumbing. We didn't have any APIs in place to our CRM system, our finance system, our PSA tool. And then last, as I said, we didn't have any way of measuring this and reporting to the board. So we had to get our house in order.

[00:16:48] And eventually what happened was I shared some of this with customers and they started looking at me with this wry smile and saying, you know, you're not alone here. We've got some of these issues going on as well. And I realized after a while that actually most people have most of those problems. We were not alone. And so we now call those the five foundations of AI success.

[00:17:12] So define the use cases, secure the data, train people per use case, build the agents per use case, and then measure, measure, measure. And what we advise people now is measure the ROI. And if you're not seeing an ROI in three months, pull the licenses back from that use case. Say that one didn't work, but give the licenses to the next group for the next use case and treat each of these as an experiment. And then when you get one that really works, then you pour gas on it and really accelerate it.

[00:17:41] And it's so refreshing to hear you speak so candidly about this because I think everything you said will resonate with business leaders around the world. And you can look on your LinkedIn feed and you almost feel like your organization is getting left behind. Everybody else has got all the answers. Everybody else has got it right. But a quick look behind the curtain, that's far from the truth. But obviously- Far from the truth, yeah. Yeah, and you've been on this journey and other business leaders have been on this journey too. But now, now you've learned these lessons. What are you doing well with AI?

[00:18:09] And what are you seeing others do well with AI? I'll tell you what I feel we're doing well with now is we've got ahead of the curve with co-work and the token economy. Yeah. So, you know, most people are now aware and probably still dealing with the hangover of the fact that the token economy hit us July 30th, I think it was. And, you know, for a while we all thought we'd just pay our license to co-pilot or chat or cloud and happy days.

[00:18:39] Now we realize, oh no, that's just the beginning. And now we've got to buy credits to get access to tokens. And then we've got to pay for some API calls. And then if we build something, say, in the Azure environment, we're then paying for some of the Azure services and Foundry or whatever. So, and there are actually 31 different potential spend buckets in AI. So, it's going to become really, really complex really quickly. I feel like we got ahead of that early.

[00:19:07] And just this morning, actually, I got a report from one of our BAs showing all our spend on co-work last month by individual and by cost center. So, we're building out that capability so that we can pull in all these costs and say, here's what Neil spent last month. And our report even showed per day of August, you had $16 on this day and $32 and all that. And your total for the month was, you know, a couple of hundred bucks, whatever.

[00:19:33] Now we can look at that with a new lens and say, all these workflows you set up with co-work to automate these reports you used to do manually or all these things you used to do. Are we getting more than 200 bucks of value? And so, I feel like we as an organization, we've got ahead of that. Scale your business with agentic AI with limited risk.

[00:19:56] With Denodo's AI data layer, your agents are provided with real-time company data and guardrails for company protection. Create the business you always dreamed of with Denodo. And you can do that by simply visiting denodo.com to learn more. But now, back to my guest. And before we started recording today, you and I were having a separate conversation about Nashville. Because obviously that's where you're talking to me from today.

[00:20:23] And I myself, and I indeed would imagine everyone listening would say, what do you know about Nashville? They would immediately say, music. And the music scene there is worth $10 billion a year or something like that. But healthcare over there is $97 billion. For people listening, I want to bust a few myths here. Tell them about that and how you're seeing technology in that sector in Nashville. Because you've got kind of a front row seat to what's happening here. We do. We've got the three largest healthcare companies in the U.S.

[00:20:53] right here in the neighborhood I'm in, just south of Nashville. It's called Brentwood. And the largest of those is the Healthcare Corporation of America. They're a colossal organization, 50-something years old, about the same age as Microsoft. The only company in the world that has gone public three times and gone private three times. Wow. So, a lot of millionaires around here living in big bougie houses from all those transactions. But they're an amazing organization because I think there's about 550 organizations that came out of HCA.

[00:21:23] And there's a poster we've got around here called the HCA Family Tree or the Nashville Family Tree. And at the top you have HCA, that huge company. And then the 550 companies that were spun out or split out or that they've acquired or have partial shareholdings in. And it's a phenomenal ecosystem here because what happens is the people, the IP and the capital moves around very fluidly. Very, very fluidly.

[00:21:51] There's new healthcare companies popping up all the time. There's new physician groups and new specialty groups. There are a lot of incubators that will develop IP that gets picked up by those healthcare companies. A lot of the New York funds come to Nashville multiple times a year and they're involved in all the M&A and mergers and listing and delisting and all that. It's a very, very fluid dynamic environment from a business perspective.

[00:22:17] And then, of course, from a tech perspective, there's a ton of innovation in that healthcare space, trying to do things better, faster and more cost effectively. And as you're probably aware, healthcare in the US is, it's quite a beast, let's just say. I think it's about 18 or 19% of GDP. You know, it's a colossal spend. The spend for every family is colossal. You pay for your insurance, but you're still paying for a ton of stuff out of pocket.

[00:22:44] But it's just, it's hard to, it's hard to overstate how huge healthcare here is as a business. And Nashville is this center where we've got the three biggest companies and all the others. So it kind of dwarfs healthcare. Sorry, I mean music. It's not as sexy as music. It doesn't get all the headlines and it's not, you know, your radio isn't talking about healthcare every day. But it is playing music that was produced in Nashville, I bet you.

[00:23:10] Even if it's not country, so much of global music is actually produced right here. But healthcare is colossal and drives a huge amount of the tech innovation. Absolutely lovely. And if I was to ask you to look in my virtual crystal ball and look at two, we've talked about the last five years since we first started talking, all the changes we've seen. And it's going to be impossible for you to predict the future with the pace of technological change at the moment. But how do you see this evolving and changing? Two thoughts.

[00:23:40] If I look at the vendor landscape, it's a jungle right now. And the amount of money flowing around is ridiculous. So it's very, very hard to predict what's going to happen in the vendor and the tech space. But I do think NVIDIA is going to become an absolute juggernaut, the way they're acquiring, you know, other organizations and doing all these cross deals and all that. I think the models will come and go. There will be shiny new models.

[00:24:09] There'll be an exciting, very dynamic space. Like right now, Claude is sucking up all the oxygen out of the room. And, you know, that's a very exciting platform. And Tropic in general. But that will change. I think that's going to be highly dynamic. In terms of the long-term play, I firmly believe the companies that focus on a holistic, integrated ecosystem will win out.

[00:24:34] So Microsoft will win out because they generally take an ecosystem approach. And I think what they've got is, you know, I think of it like a pyramid. And at the bottom of the pyramid, almost every organization in the world is using Microsoft applications. You know, whether it's Outlook, Teams, Word, Excel, PowerPoint, SharePoint, blah, blah, blah. And at the bottom of the pyramid, we have all our data. And I think of that as our dirty data, our unstructured data.

[00:24:59] So every email you've ever touched or deleted or drafted even, and every Teams message and every document you've ever opened and everything you've ever browsed, all that unstructured data lives inside the Microsoft environment. And then at the top of the pyramid, you have the Microsoft control plane. So tools like Entra that define the identity and all our security groups and permissions and all the attributes that flow down.

[00:25:22] And under that, you've probably got Intune to manage the devices per view to manage the data defender for the security signals. And so I kind of think that the Microsoft AI play will be the middle layer that will inherit all the controls down from the control plane.

[00:25:37] All the identity, all those attributes, all the role-based permissions, all the security groups, all the DLP definitions, all the sensitivity labels, all that gets inherited into Copilot and then Foundry and Cowork and Fabric and all that layer. And then all that can reason over all the unstructured data at the bottom of the pyramid.

[00:26:03] So I think Microsoft is going to win long-term with that view. Google will do very well too, because they've got an entire ecosystem top to bottom. And I think the models, you know, Microsoft will tuck in. They've already tucked in ChatGPT and they've tucked in Cloud and others will come so that you go into Copilot. That will just be your UI, the user interface. And then within that, then you select, I'm on Cloud or I'm on ChatGPT or I'm on both to do this piece of work.

[00:26:33] Or now you can say, I want ChatGPT to do the piece of work and I want Cloud to critique it or vice versa. Or you can go into Foundry and say, no, I want to select Grok or DeepSeek to do this piece of work. And, you know, the word is your oyster. Once you go into an environment within an ecosystem that's secure and inherits all your guardrails from the control plane. So I think Microsoft will do very well long-term because of the ecosystem. So will Google and models will come and go.

[00:27:03] And I think NVIDIA will mop up a huge amount of the infrastructure. Exciting times ahead. We're going to have to get you back on in the next 12 months and see how things are evolving. And I know it's a busy time for you. So for anyone listening, want to find out more about that 2026 Endpoint Ecosystem Study, I will include a link to that. But is there anywhere else you'd like me to point people listening? Just on LinkedIn, but can I share one more thought around what we see happening in the near term? Yes, of course.

[00:27:33] I talked about some of the vendor play. The very real, very near term issues that I see every organization having to grapple with. And these will be big, big issues in 2027. There'll be three issues, data, agents and spend. So the way I think about it is problem number one is going to be managing all our enterprise data. Classification, labeling, data loss prevention controls and preventing misuse and preventing shadow AI.

[00:28:03] Making sure people are not uploading sensitive company assets to a personal account, a public LLM with a personal account. Problem number two is going to be managing all these agents. It's going to be the next mushroom cloud in the enterprise. And I think it was Gartner predicted 100 agents per knowledge worker over the next few years. That makes my head spin. Yeah, yeah.

[00:28:28] And we're getting to a point now where, you know, we're going past individual agents to enterprise wide agents. And then soon agents will be building agents. And how we keep control of those and assign an identity to each one and the right permissions so that they can make the right API calls and access the right data, but not overreaching. And then how we keep those agents up to date and patched and secure.

[00:28:53] And then how we make sure they have the right owners so that if an employee leaves the organization, we're not left with a bunch of orphaned agents that nobody understands. Managing those through the lifecycle. That's going to be that's problem two. And problem three is going to be the spend management. How on earth we wrap our heads around all the licenses or the tokens or the API charges or the platform consumption fees and potentially do that across two or three different platforms.

[00:29:23] And how we pull that all into our organizations, make sense of it every month and be able to give a report to every individual so they can see what they're spending and be aware of it. And then a report to every call center manager. So here's a total for all the people in your team. And then to every business unit to say, here's your total bill for last month based on all the people in all your call centers. That's the 2027 challenge we need to solve for first.

[00:29:51] First data agents spend. I completely agree with you. And I think that identity management space is just going to explode over the next few months and into next year as well. At the moment, we're talking about a few individuals. But if we try to imagine, I don't know, an individual with 50 agents that he's created, he or her has created there to help them through their work and then magnify that through team, through department, through the whole enterprises. It's breathtaking, isn't it? It is breathtaking.

[00:30:21] And the organizations who will nail this, I'm seeing really strong correlation. The organizations that I see making good momentum to solve in this are the ones who solved some of the legacy enterprise problems over the last few years. Like I think we spoke about, I think we spoke about provisioning, the provisioning processes last time. And what I see is organizations that fully automated their provisioning processes during COVID.

[00:30:50] They freed up a ton of their IT time to focus on new things like security or AI or data protection. And the organizations that automated all their patching. All the five kinds of patching. They freed up a lot of time. And those who did, who implemented self-service password reset. If you get rid of the three Ps, you know, the passwords, the patching and the provisioning. That freed up a lot of IT time to now focus on the new stuff, the new frontier. Data security, AI, blah, blah, blah, all these things.

[00:31:20] The organizations who didn't solve those problems and are still taking every device out of a box and manually configuring them and doing manual patching. And their service desk is still doing manual patching research. They're in a hole now because they're so busy just doing day to day. They don't have enough time to learn the new stuff and get the organization ready for the new stuff. So I'm seeing there's going to be this bifurcation. The organizations who are in a good position are going to go faster.

[00:31:50] Those who have a lot of legacy and tech debt and on-prem infrastructure are going to be going much slower, much, much slower. And I think that is a powerful and thought-provoking moment to end on. Cannot thank you enough for coming back on today, sharing your insights from that 2026 research study, specifically around how AI is magnifying existing workplace challenges and widening the gap between innovators and laggards. And we covered so much there.

[00:32:20] So I'll include links to the study, to your LinkedIn, the company LinkedIn. I urge people listening to check those out and let's keep this conversation going. But a massive thank you as always for sharing everything. Really appreciate your time. Thank you, Neil. Delightful to be on your show again. And I'm looking forward to the next one. I think Dennis's five foundations offer a practical test for any AI program. Define the use case. Secure the data. Train the people for the work they actually do.

[00:32:49] Build the agents around a use case and then measure the result. If the return is absent after three months, maybe move that license to another experiment rather than defending a weak deployment indefinitely. And there was also a lesson in the three Ps. Companies that automated passwords, patching and provisioning. They're the ones that created room for their technology teams to work on data protection, AI and agent governance.

[00:33:18] So carrying years of manual work. So those carrying years of manual work could find that old debt is slowing down your new ambition. So I will add links to that endpoint ecosystem study and LinkedIn for both the company and guests. But thanks to Dennis for returning to the show. And over to you. Could your organization explain which AI use cases are working right now?

[00:33:45] What they cost and why employees keep using them? If not, maybe it's time to move on to the next thing. But seriously, let me know your thoughts. Tech Talks Daily I'll be back again soon with another guest. Thanks for listening. Bye for now.