Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

AI Security, AI Cloud Ops & Infrastructure | ClearTech Loop

AI Security, AI Cloud Ops & Infrastructure | ClearTech Loop artwork

Your AI Is Making Decisions Your Company Can't Defend with Maribel Lopez

Oct 21, 202615 minAudio
ShareXLinkedInFacebookWhatsAppEmail

Send us Fan Mail Your company has an AI governance committee. Great. But when an AI agent makes a decision that harms your business, who actually has the authority to shut it down? And here's a bigger question: Would you even know it made the wrong decision? In this episode of ClearTech Loop, host Jo Peterson sits down with Maribel Lopez, Founder and Principal Analyst at Lopez Research, to examine the uncomfortable reality behind enterprise AI adoption. Organizations are moving quickly to deploy AI, but many haven't established clear accountability, reliable data practices, or meaningful ways to measure whether their investments are actually improving the business. Maribel challenges the assumption that forming a governance committee solves the problem. She explains why AI can make technically logical decisions that are completely wrong for the business, and why governance becomes even more critical as AI agents move from recommending actions to taking them. In This Episode Who's actually in charge? Why AI governance committees can fail without someone empowered to intervene. When AI goes rogue: Why human oversight, permissioning, and rollback capabilities matter. The productivity trap: Why making someone's job 30% faster doesn't necessarily create business value. The business context problem: How AI can use accurate data and still make a disastrous decision. Data you can't trust: Why disconnected systems and conflicting information undermine AI outcomes. Decisions you can't explain: What happens when organizations must justify an AI-driven decision. The executive responsibility: Why AI governance extends beyond IT and security. The question isn't whether AI can make decisions. It's whether your organization can stand behind them. About Our Guest Maribel Lopez is the Founder and Principal Analyst at Lopez Research and an enterprise AI analyst who advises organizations on emerging technologies, AI adoption, and business transformation. Website: https://lopezresearch.com/ About Our Host Jo Peterson is the CIO of Clarify360 and Chief Analyst at ClearTech Research. Through ClearTech Loop, she explores the technology trends, risks, and decisions shaping enterprise leadership. Worth Reading + Watching AI Has a Business Context Problem | Maribel Lopez Why accurate data alone isn't enough for AI-driven business decisions. https://lopezresearch.com/2026/05/06/ai-has-a-business-context-problem/ The Hard Part of Agentic AI Isn't Building the Agent. It's the Space Between Your Tools | Maribel Lopez The challenges of connecting AI systems while maintaining context, security, and governance. https://lopezresearch.com/2026/08/02/the-hard-part-of-agentic-ai-isnt-building-the-agent-its-the-space-between-your-tools/ AI Agents Are Already Working for You. Do You Know Who They Are? | ClearTech Loop with Michael Krigsman A related conversation about visibility, ownership, and securing an AI workforce. https://cleartechresearch.com/ai-agent-security-michael-krigsman/ Stay in the Loop For more conversations challenging the assumptions behind enterprise AI, cybersecurity, and emerging technology, follow ClearTech Loop. Subscribe on YouTube: https://www.youtube.com/channel/UCR7xPyjXPgQJKMBvnjoGBQw Subscribe to Jo's LinkedIn newsletter: https://www.linkedin.com/newsletters/7346174860760416256/ ▶ Watch on YouTube: https://www.youtube.com/@ClearTechResearch/videos 📰 Subscribe to the Newsletter: https://www.linkedin.com/newsletters/7346174860760416256/

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

Jo PetersonHey everyone! Thank you so much for joining this week's session of Clear Tech Loop. I'm Jo Peterson. I am the CIO of Clarify 360 and the Chief Analyst at Clear Tech Research, and I've got a wonderful guest today. Talk to us about AI governance. So it's Maribel Lopez. Hi, Maribel.

Maribel LopezHey, how you doing, Jo?

Jo PetersonGood. Thank you for making time. In case you're not following Maribel, you should be if you're interested in AI. Maribel is the founder of Lopez Research, and she is an enterprise AI analyst. So let's just get going. For non-technical executives, AI governance is an organizational design and risk management issue rather than an IT problem. Leadership must focus on accountability, operational guardrails, and measurable business outcomes. So, from a strategic perspective, there's a few AI governance questions that every C-suite executive should ask, and I wanted to get your thoughts on them.

Maribel LopezOkay.

Jo PetersonAll right. So the first one here we go. AI governance often fails when accountability is spread too broadly. How important is it for the C-suite to understand who exactly has the authority to stop or override an AI model, and does that team member know it's their job?

Maribel LopezYeah, this is actually an interesting question. So you started the dialog with the question about AI governance, and one of the things that we found when we talk to organizations, or when you look at a bunch of the published research, it turns out that many organizations didn't have governance practices to begin with. So there wasn't necessarily a governance group. There were people that did aspects of it at legal and in HR, etc. but there wasn't a formalized governance group. There may or may not be formalized governance strategies. So for many organizations and and CEOs, anybody in the C-suite, they're now faced with the introduction of AI, creating the need for them to have governance policies, procedures, things that they don't already have a baked-in skill set for so I think that's challenge one.

Maribel LopezChallenge two is that many of them approach this like, well, we need a committee. So I've talked to many organizations that have you know AI risk committees, AI governance committees, and there's nothing wrong with that. But at the end of the day, because you do want input from different groups within the organization of how a policy policy might look, but somebody needs to be in charge for certain aspects of this. And I think the third challenge, which you were trying to get at, is that nobody actually knows if it's their job. So if we dial it back a little bit, and we think about other technology trends that we've looked at, if you think about what happened with security, if there's a security breach, most people don't know who to call. They they will arguably call IT and say, "What do I do? Right? If we built when we were building mobile, you know, we got all the way to the end of building a mobile app before anybody looked at it in security and decided if this would be okay or if it would leak data or any of that other stuff. So, if we back up the bus for a minute and say that as we're looking at building our application services and making our technology decisions, there has to be some discussion around what data that uses, how that's being protected, what kind of controls are we buying within our systems? So I think that that is that actually shares some of the responsibility to each of the groups that are making those decisions. But at the end of the day, there has to be somebody that can say, "This we turn off until we figure out what's going on. So it's not an easy answer to your question because I don't think it's you know you run everything under the planet and the company through one individual who looks at it all and says yes or no that's not a scalable model I think you need to be thoughtful as different groups within the organization of understanding that there are data security and governance risks, and asking the right questions about well, how do I know? And then there's a minimum viable discussion of you also don't want to make it onerous because that stops innovation, right? So, what's the thing that we really know that we have to protect? And then we're going to work through the other edge uses cases we go on, and we'll probably get to talking about agentic AI and other things as as we go forward.

Jo PetersonYeah, and you make a really good point because your crown jools for AI may be different than your crown joules for other things.

Maribel LopezYes,

Jo Petersonright. It just might be. So, how are non-technical C-suite execs that you chat with approaching that issue of minimum viable guardrails for human oversight and data privacy?

Maribel LopezYeah, right now they're mostly. I would say the first thing is most of them aren't, but assuming you were right. If you were, you're actually spending a lot of time talking to your technology teams and starting with the principle of, can we build things that do no harm? You know, that's that's sort of your minimum viable product. Is that we are not putting our corporate data just randomly into large language models. If we're going to roll out agents, we might start in a more slow roll where we look at our SaaS applications that already have a certain type of identity and security and permissioning built into them that can be extended to the agents. Right. So I think we we look for smaller ways while we try to build a bigger picture of where are the risks and what could go wrong. And it's actually a very dynamic environment. So if you look at it every month or two, something's changing. You know, we just had not one, not two, but three model providers say that their models broke out and you know infiltrated companies while they were doing testing.

Maribel LopezSo, and that's just in the past couple of weeks. So you can see how past week, even maybe it's moving pretty fast. You can see how this is this is one of those things where it's not a set it and forget it. So I think the first thing that you that non technical C suite executives do is they make sure that they're having a cadence talking to different people in the organization. Are you testing things? Is it working? Do we have rollback policies? Like, is there somebody that can hit a button and turn off that agent if it goes rogue? You know, we need to make sure we have those kinds of guardrails embedded into the new systems that we're putting in, because while big decisions are going to require human review, we're building things that move so fast and at such scale, we're not going to be able to have a human review every agenda transaction that's going on. So some of this has to be really thoughtfully planned at the outset.

Jo PetersonIt feels that way, right? Last question: Governance should be measurable, not assumed. So, how are non-technical execs measuring if AI is actually improving business outcomes?

Maribel LopezYeah, this is interesting, and that and that's slightly different from the governance per se. I think when we think of measuring governance, you know, we have to be able to go back and say, you know, here's how we protected the data, here's how we allowed the permissioning of the data, here are the tools that we have put into place that test the model outputs that you know the agentic outputs that are happening to see if they're working the way they should, and we can either modify those or roll them back, right? So I think there's that whole technical aspect of it. When we look at AI in general, though, there's a question of is it measurable, and I think we we don't want to just get into the trap of, oh, you made someone's job 30% faster; they could do 30% more because that that's not really the thing. When I talk to organizations about how you prove value and where where the risk versus reward is, particularly when you're thinking about data and governance, you have to look at whatever your top three initiatives are in the company.

Maribel LopezMaybe it's to improve customer satisfaction. Maybe it's to reduce billing issues or improve your supply chain, and make sure that the AI solutions that you're rolling out are going to have some impact on those first. They'll either make that process a smoother process so that it operates more efficiently, they will either make that process better, so that you get better answers. There's a lot of discussion of business context right now, which I think business context and governance are really very tightly aligned because the assumption is you're just going to roll out AI and it's going to look through all the data throughout your company and find interesting correlations and make interesting judgments and insights, and then in fact actually act upon those. So once you get to the action step, then you're really in governance land. Like, was this the right thing to do or not? Was I thoughtful about how I created this? You know, frankly, the practicality of will I get sued if something goes wrong. So I think there's a lot of discussion of all the problems we have with data management before it still exists. They're just compounded with AI, and then we layer on the fact that for the first time in a long time, we're starting to get these governance regulations like the EU AI Act that have teeth around them if something goes wrong. So you could really be on the financial hook for things. So I guess when I I think about the AI improving business outcomes, I say you have to have the use case. You have to discuss with somebody at the outset how you're going to measure that. Is it going to be that you shipped 30% more code?

Maribel LopezIs it going to be that you ran X percent more things through the process. Is it going to be that you contained Y amount of calls in the contact center that didn't need a human? Right. There's all these different ways you could measure things. Some of them are direct monetary. Some of them are process improvement. Some of them could be our ability to take in marketing ideas and create minimum viable products to test to the market. That's not all the way to revenue, and it's not all the way to cost, but it's a definite business outcome that you could look at. So those are some of the things we've been working with organizations on.

Jo PetersonYeah, that's really good, and I like that you married context and governance because they do feel close right now, and the thing that we're not saying out loud, out loud in that same sort of vein is trust. Right, yeah. It's there, and we're all sort of thinking about it, and we're all sort of thinking about how does that apply to me and my company and my brand image and all those things, right? But it's not-it's a bubbling idea that I don't think we'll see it get teased out over the next couple months.

Jo PetersonWhat do you think?

Maribel LopezI think we started down the path, and I think that the connect hasn't quite hit in organizations' minds. So, for example, we we've all heard about the garbage in, garbage out. That actually makes perfect sense with AI. If you put in bad data, you're going to get bad data out. But if you think about it, the largest challenge that organizations have right now when they move into AI projects is in fact their data. Do they have the right data? Is the AI accessing the right data? Now you can also have your AI too disconnected from your data in the sense that it is making interesting observations and possibly taking actions, but they're not tied well enough to your business, for example, it could know that I am waiting for parts, and I am this category of client today, and it will decide that well, maybe you get the parts later because Maribel's not the top tier client today. Well, it could be that there's business logic if it had access to the right information, that says, you know, my pipeline is for me to be a huge client by the middle to end of next year, and if you don't know that as the AI, you'll make a very different decision than if you knew that, and that's where that business context really starts to matter, right? And that's where you have to be confident that the data sources that you're using are the truth. I mean, I've been in meetings with retailers where they would argue over sales numbers: was were the sales X?

Maribel LopezWere the sales Y? You would think that that would be a very standard, basic thing that you would know when you go into a meeting, but it turns out that you can have different sources of data. You could be looking at Excel spreadsheets that are outdated. By the way, the world still runs on Excel, and not all those Excel spreadsheets are being updated by Claude CoWork right now. So we're in a we're in an environment where I think trust in the data. You know, we know AIs hallucinate in general if they're not grounded properly, but then even if you ground them in your data, if they're not grounded in the right data, that's a secondary issue. And then there's a whole third issue. If we really get to machines starting to make autonomous decisions for people, we also have to be able, from a governance and ethics standpoint, be able to back that up and understand how that decision was made. So, for example, if a mortgage was or wasn't approved, I need to be able to explain that to the customer at the end of the day.

Maribel LopezSo, I think we have multiple dimensions in the trust dialog that we're still working through. But it's a very you know, there's a lot of opportunity, so I'm very excited about our ability to impact change moving forward.

Jo PetersonYeah, yeah, and I'd love to have you back and talk about this idea as it develops further. But for now, thank you for taking time to visit an audience. I hope you found this valuable because I did, and we'll talk at you later.

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. AI Agents Are the New Insider Threat. Are You Ready? With Tia Hopkins

    Nov 18, 202614 minAudio

    Send us Fan Mail What if the next insider threat to your organization isn't a person at all? AI agents are moving beyond answering questions. They're accessing systems, executing tasks, requesting credentials, and...

    Credit the people on this episode

  2. NinjaOne Katana AI: Can AI Fix IT Problems Safely? | Jeff Hunter

    Nov 12, 202621 minAudio

    Send us Fan Mail Can AI safely troubleshoot and resolve IT issues without putting your entire endpoint environment at risk? NinjaOne Field CTO Jeff Hunter returns to ClearTech Loop to join host Jo Peterson for a...

    Credit the people on this episode

  3. Shadow AI Is Already Inside Your Enterprise | Anthony Russell

    Nov 3, 202616 minAudio

    Send us Fan Mail Your engineers may be using AI agents your security team doesn't even know exist. Those agents could be accessing sensitive data, handling company code, or operating with credentials nobody is...

    Credit the people on this episode

  4. Your AI May Be Smarter Than You. Can You Still Control It?

    Oct 28, 202615 minAudio

    Send us Fan Mail What happens when AI becomes capable of making decisions its creators can't fully anticipate or control? AI agents are gaining autonomy, accessing enterprise systems, managing credentials, and making...

    Credit the people on this episode

  5. AI Agents Are Getting Hired Without Background Checks. Here's the Risk| Rasha Horn

    Oct 13, 20269 minAudio

    Send us Fan Mail Your company wouldn't hire an employee without onboarding, supervision, or defined access. So why are AI agents being given responsibilities without those same safeguards? In this episode of ClearTech...

    Credit the people on this episode