Episode 423 – Non-Human Identities in Microsoft Entra with Eric Woodruff and Chris Brumm
39 minAudio
Welcome to Episode 423 of the Microsoft Cloud IT Pro Podcast. In this episode, Ben is live from Workplace Ninjas, joined by Eric Woodruff, Chief Identity Architect at Semperis and Microsoft MVP in Security focused on identity, and Chris Brumm, Cyber Security Architect at glueckkanja and Microsoft MVP in Security with over 16 years of experience in cybersecurity. Together they dig into the often-overlooked world of non-human identities in Microsoft Entra ID. They cover what service principals are, why they tend to fly under the radar compared to user accounts, and how attackers actively exploit that gap. The conversation spans credential management best practices, the risks of improper owner assignments, the challenges of multi-tenant app configurations, and why managed identities should be your go-to wherever possible. They also discuss the growing challenge of AI agent identities and what IT pros need to start thinking about now before that surface area explodes. Show Notes Eric Woodruff on LinkedIn Eric Woodruff on X (@ericanidentity) Eric on Identity Chris Brum on LinkedIn Chris Brumm on X (@cbrhh) Chris Brumm’s Blog Application and service principal objects in Microsoft Entra ID Workload Identities Securing service principals in Microsoft Entra ID Securing managed identities in Microsoft Entra ID Conditional Access for Workload Identities Microsoft Entra Audit Logs Microsoft Sentinel Detection Templates Eric Woodruff Eric Woodruff is the Chief Identity Architect at Semperis and a Microsoft MVP in Security with a focus on identity. He specializes in all things Microsoft Entra and Active Directory, with a passion for helping organizations understand and secure both human and non-human identities. You can find Eric on social media as @ericanidentity. Chris Brumm Chris Brumm is a Cyber Security Architect at glueckkanja based in Germany, with over 16 years of experience across virtually every corner of cybersecurity. He is a Microsoft MVP in Security with a primary focus on identity security. His team operates SOC services and he brings a detection and response perspective to identity risk, helping organizations build lifecycle processes and monitoring strategies for non-human identities in Microsoft Entra. About the sponsors TrustedTech is a leading Microsoft Cloud Solution Provider (CSP) specializing in Microsoft Cloud services, Microsoft perpetual licensing, and Microsoft Support Services for medium and enterprise-sized businesses. Our robust team of in-house, U.S-based Microsoft architects and engineers are certified in all 6/6 Microsoft Solutions Partner Designations in the Microsoft Cloud Partner Program. M365 Licensing Consultation M365 Tenant Assessment Copilot Readiness Assessment At Intelligink, our focus is singular: the Microsoft cloud. Our Microsoft 365 and Azure experts help you work securely and efficiently by unlocking the full value of what you’re already paying for, so you can focus on running your business.
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.
Welcome to episode 423 of the Microsoft Cloud IT Pro podcast recorded live from Workplace Ninjas US in December 2025. This is a show about Microsoft three sixty five and Azure from the perspective of IT pros and end users, where we discuss a topic or recent news and how it relates to you. In this episode, Ben sits down with fellow security MVPs, Eric Woodruff and Chris Brum, to discuss the often overlooked world of nonhuman identities in Microsoft EntraID.
We cover service principles, why they tend to fly under the radar, and how attackers actively exploit that gap. They also spend time talking about credential management best practices, challenges around nonhuman identities, and how you should be thinking about them, especially with the growing challenges of how they relate to AI. Let's dive in. I'm sitting here with another episode of Microsoft Cloud IT Pro podcast at Workplace Ninjas, the first one in 2025.
We all get to be honorary attendees, so Scott, who I normally do the podcast with, if you guys if people listen, is not here once more. But instead, I have Eric Woodruff and Chris Brum. And, we are going to talk about non human identities in Entra. So, this is an interesting topic because, yeah, I'm excited to dive into it. But first, do you guys wanna tell us a little bit about who you are, where you are, company that you work for? Yeah. Sure. Yeah. So I'm Eric
Woodruff. Currently, I'm the chief identity architect at Semperis. I I do a whole mixed bag of things, which I I won't sorta go down the rabbit hole on there. But, yeah, MVP in security focused on identity, so all things, Entra, some active directory thrown in the mix. Yeah. And it's just it's, you know, great to be here. You know, I will say you can find me on the socials as Eric on Identity. It's maybe not super hackery sounding, but it's hopefully easy to remember. Okay.
Yeah. There you go. Easy to remember is good. Hi. I'm Chris Broom. I'm from Germany, and I'm a cybersecurity architect with Glucanje. And, yeah, I'm doing cybersecurity for, I don't know, sixteen years or so. I think I did everything that you could do there, but my main topic is identity security. And like Eric, I'm also an MVP in that area. So my company is doing, yeah, managed SOC services mainly. Okay. And,
yeah, we are protecting our customers, and I think identity is one of the main main topics there. Alright. So no red teaming? Are my devices safe around you? Yeah. Yeah. Yeah. I'm I'm in the blue team. Absolutely. So Yeah. That's kind of funny. Like security conferences are like, don't ever connect to the WiFi because you never know who you can connect. Yeah. You just have to scan this QR code. Yeah. So that's been funny. This conference, the amount of QR codes and everybody's like, this is a security conference. Yeah. Yeah. And we just found here USB sticks. Have you plugged it in yet? No.
I don't do this in general. Oh, good times. So moving on to non human identity. It's like when I kinda reached out and said, who wants to talk on the podcast? Who wants to do some interviews? You guys were like, well, we did a session on non human identities in Entra. And this is an interesting one because a lot of times when people think Entra, they think, oh that's how I log in. I have my username, my password, I'm doing MFA whatever Entra. But there's a whole another component to it around
non human identities and Entra that maybe sometimes gets forgotten and I think that was some of the premise of your presentation. Kind of what we wanna talk about today. Yeah. Our basic idea was to to just collect all the stuff that should become a knowledge by an intro admin and just to make sure that everyone knows it. So about permissions and credential handling and all the stuff that yeah. What could go possibly go wrong? Yeah. What are our topic? Yeah. Yeah. And I'll just say, like, I think,
really mean the the term is new, but nonhuman identities have been in Entra for as long as it's been Azure AD. Right? I mean, like, we've been talking about service principles and stuff. They've always been there. But, yes, it's like a a hot topic in the industry now with nonhuman or workload identities, whatever you wanna call them. But yeah. And for as long as these have they have existed, there's abuses on them that attackers love. So trying to talk about those things. Yeah. So if somebody's newer to this because sometimes I think we
even make the assumption, oh, everybody to your point, everybody should know about non human identities. When you start talking about non human identities, we're talking about SPNs. What are those? For somebody that may be newer to enter is like, wait a minute, I need to think about non human identities in Entra. Where do you kinda start with those when you talk to people about non human identities in Entra? I mean, I would say Chris is like, oh, no, not this topic. I mean, I guess it depends if they know AD. Right? If they know AD, I
would try to relate it to, you know, GMSA's service accounts. Right? And if they don't know anything in general, I guess not that they don't know stuff. Right? I guess I'd be saying that this is like for you know, you think of applications that might go do stuff. You have SaaS apps that are doing things behind the scenes with, you know, data and like SharePoint or OneDrive or, you know, Exchange. Right? There's still identities to all those things. Right? Those services under the covers and that's essentially what, like, service principles are that are operating, like, without people. That's the the
non human part. Yeah. And a lot of those have access to data. And I think that was a big part of your talk is or part of your talk, and I'll let you guys kind of direct it is managing those SPNs and to your point, they have access to data much like a user would, but yet I don't think they maybe get us necessarily as much attention as users and what they have access to. Yeah. I think there are two big topics that we have to talk about. And the first thing is what permissions can f apps have.
So and apps can have huge permissions. So and we should if an app has the same permissions like a global administrator, we should handle it like a global administrator. And the other part is, okay, who has permissions on apps? And there are some really bad practices out there. So as an example, yeah, a lot of companies are assigning owners at at the app registration and the service principles, and they are assigning that owner permission to normal user accounts. So and an owner is such, yeah, is a person who can
do everything with that app. So including adding new secrets and log in as that app and use the permission of the app. So there's a big potential for stuff that could go wrong. Yeah. So as you're working at identity then, you're working with this SPNs and the owner rights and doing all of that. How do you start guiding people through some of that management of those SPNs? Yeah. I mean, I'll I'll say it's tricky at times because but there's always places where
customers love Microsoft docs and customers want to do nothing against Microsoft. Right? They wanna do things as, like, this is what Microsoft wrote. That's what we wanna do. And occasionally, right, the rest of the industry disagrees. And and, like, owners is a thing. In particular, like, when I worked at Microsoft, I would tell people, go add, you know, Joe, you know, owns Salesforce in the business. Right? Go make Joe an owner. So that way you sort of know, right, we gotta go bother Joe about Salesforce. Not realizing that again, if an attacker
targets Joe who's not treated as a privileged user. Right? That now attacker, right, compromises Joe, phishes him, whatever. And now attacker can move into Salesforce. And and to your point about data. Right? You see in the news that data is the thing attackers want. Yeah. But identity So it's a little bit. Yeah. Yeah. And but the I mean, I think to what we were saying about the permissions that they have on data. Right? Like, obviously, like, it's very broad. Right? Like, if an application
has permissions to exchange, right, many times, it's like it has permissions to everyone's mailbox and can do whatever it wants with all that. So, yeah. Yeah. I think about that. Like, I've I've done migrations before. Right? Whenever you start migrating email, they're like, oh, go create the migration tool, the exchange web I think it's exchange web services Yep. In the SPN. Go create an SPN for it. Go create access so it can migrate all the email and it is it's that concept of wait a minute if this is migrating all
the email it's touching all the email which means it's going to have access to all of the email and am I treating this in that way? Yep. And in this case, it's not only about write permission. So people don't usually, when when I talk to people, say relax when they heard say, it's only read permission. So, yeah, but read permission to your whole file system or your whole mailbox is, yeah, also. But Yeah. Pricing. Data data exfiltration is a is a huge problem right now. So yeah. And I think we you maybe have or we see better
practices regarding authentication method policies and conditional access rule set, but it doesn't affect work nonhuman identities usually. So there are other policies. There are application management policies, and there is conditional access for workloads by yeah. But I rarely see this. Yeah. Yeah. And I mean I think it's almost like right like workloads, service principles, they're treated like a sort of second class citizen. And and I get like as a sys admin right you might feel like you already have too much to do. So
it's easy to say you have to do more. But on the other hand, attackers love going after service principles because they they tend to know that assist admins don't have a strong understanding of what they do or what they have access to. And and again, if you see a tax on, you know, Microsoft three sixty five going after, you know, OneDrive, SharePoint, Exchange, many times it is using those, you know, enterprise apps as your service principles that they'll they'll leverage. So
This episode is brought to you by Trusted Tech. If you're managing Microsoft three sixty five, you already know licensing isn't simple. E three versus e five, the brand new e seven licenses, Copilot security bundles, and with price changes coming in July 2026, the stakes are getting higher. Trusted tech helps IT leaders make sense of their Microsoft three sixty five environment with a free licensing consultation. Their engineers analyze what you're licensed for and what you're actually using and where you can optimize, whether
that means consolidating security tools, preparing for copilot, or eliminating wasted spend. Plus, they offer proactive and reactive support, which has been awarded yet another solution partner designation from Microsoft for support services. If you want a clear, data backed plan that has helped over seven five hundred subscribing customers save up to 20% on Microsoft three sixty five before your next renewal, visit trustedtech.team/mscloudprom365
and schedule your free consultation today. We talk about security risk for unused user accounts. Right? Like, you have an unused user account that's active. Part of the reason that's a security risk is it's not monitored or watched as closely because the user's not using it. I feel like service principles can kind of fall in that boat as, yeah, they're being used, but they're probably not being watched Yeah. Quite as closely as a normal user account Yeah. So to speak.
And attackers probably tend they don't wanna be seen. Yeah. Yeah. They're gonna go after things that aren't as visible. Yep. Yep. Yeah. I mean, I would say, you know, using service principles for persistent. Right? Persistence, like, assigning credentials to them, but in the way that they can dwell. Right? Like, they might get access to something and then wait and see if they're detected. Right? And then after some time, then they go on to do their their nasty Yes. So yeah. And as I said, my team is running SOC services. So and monitoring credential addition to a service principle in a
bigger environment, that's so loud. We see a lot of modifications. We'll see a lot of credential adding and so on. You really have to correlate this with the permissions of the app to get a reasonable signal to noise ratio. So Right. So do you have guidelines, Chris, in, like, how you guys do that from a SOC perspective that if end users are doing this or not end users, but other sysadmins, companies are doing it internally is how do
you go about even starting to monitor things you maybe watch for? You have to do both. You have to do prevention and detection. And I always start with, okay, do you have any life cycle processes? So companies usually have managed to have life cycle processes for their users. Admins is getting more rare and for non human identities. So, yeah, it's lost in vain. There are some good practices. So there are
a lot of good practices around how to implement OpenID Connect. So this is for the developers. You have to be the go to person for your developers regarding identity. And there are some good practices how to configure the applications. And you should leverage that. Yeah, that workload ID features. Is that it's a separate license, but, yeah, we for powerful apps, I think it's reasonable to protect them. And if you want to monitor
your apps, you have to rely on on the intro audit logs and you can adjust them in Sentinel. A good starting point is Microsoft has has a repository full of detections. But yet, to be honest, this is very noisy because it's just detecting, yeah, someone added a reply URL. Someone added a credential. So without context, it's very, very hard. So and this is what yeah. This is what what SOC providers are doing. So someone added a credential. Okay. We have a sign in
risk. Okay. But if we have a sign in risk and then this person is adding, credential to an app. This is interesting. So it's regarding context. And another context is, okay, are we talking about an app without big permissions or are we talking about an app with very high permissions? And bringing all that stuff together. But, say, yes, there's it's hard to do a one on one on this. Yeah. There's a lot of data there. And it was interesting you mentioned developers and because
I do IT pro stuff, I like to pick on developers and probably pushing that because developers probably have a tendency too to say, oh, I'm writing an app. It needs a service principle. Just give it these permissions versus really trying to narrow down and really hone what does it really need versus what makes a developer's job easier. Yep. Yeah. I mean, it's it's easy to sort of dump on developers and, you know, no offense to them. And I get that they have a job. Right? Like, if Yep. If you're a software company and you're not making software, then you'd have no reason to be there. But, yeah, there it it's I think the problem is
finding in the business, like, the balance. Right? And it's not the same across the board. Again, companies just wanna be sort of told what to do, but it's it's more of an art, I'd say, than a than a science because, you know, like, developers, right, you need to make sure that if if you create too much friction they'll go spin up their own tenants their own subscriptions right and they'll go do all their developer stuff possibly with like real data and now you have shadow IT things going on. Right. So yeah, you definitely have to like configure things in a way where you can let developers do their stuff, but I think I'll just
say also security people tend not to like to talk to anyone. Right? Yeah. Yeah. We really don't. But but I mean it is one of the the points we try to drive home in our talk is like you need to work with your developers, right? Like you you can't expect everyone to be an expert in like OpenID Connect. I mean barely security people in the industry. But you need to work with them, right, to make sure that when they're implementing things that they're doing it right and not, I'd say, almost like shaming them for not, like, understanding every OAuth flow out there.
Yeah. Yeah. And and I think another important thing is, so, yeah, people make mistakes. We don't know stuff. Yeah. We know today maybe stuff that we didn't know two years ago. So and it's a good idea to make regular reviews. And there's a lot of open source community tools out there. So stuff like, entra ops. So greetings to Thomas Nalheim. And, yeah, just use it and review your environment and find that stuff. Yeah. Yeah. And another thing that kinda goes with developers, and I know another thing you touched on in
your talk was not just these SPNs and the permissions they have, but there's secrets and certificates and how these how you actually authenticate to these applications Yeah. And managing some of that. Do you want to talk a little bit about what you guys touched on best practices, things that people should be aware of when it comes to credential management for non human identities. Yeah. I mean, that that is a ball in itself. Right? Like, I mean, if you're talking about credentials, I mean, you have secrets which are the easy
thing, but it essentially is just a password. Right? It is a password for an application. It's a better name. Yeah. Right. Yeah. We should just call it app passwords or something. Although we used to have those in Entra. Well, I think in graphs, either on app registrations or service principles somewhere, like, it It's a lot of secrets in the UI, but you then you have certificates, right, which are better, but certificates also, right, people can still do silly things, leaving them laying around. Yep. Like that. Save them to
their documents so they sync it up to OneDrive. Yeah. So I mean, Lori, the the North Star is is like federated credentials, which, again, is something that's difficult to understand for some people. But in if you're, like, solely like a Microsoft shop, right, like, it's manage identities, which are easier to deploy. Right? Because then you don't have credential management. Nobody's having to hold secrets or Right. Use them. And another advantage of the managed identities is that you have an integrated life cycle. If you delete that function, the,
the managed identity is also gone, including its permissions. So Yep. Yep. And I mean, I I will say, right, there's security research out there that will show there's sometimes cracks in sort of hybrid things like with ARC relative to manage identities. But I think if you sort of zoom out, right, and look at all the issues people have with secrets, right, it's still better to use manage identities whenever, you know, possible than, you know, certificates or or secrets. So yeah. And and I think we have to talk about this secret password. So if we if the
time comes, we have to get rid of it because I know it's hard to do it, and it's even harder than with end users. But we see attacks where attackers are phishing developers and or attacking developers to get their secrets. So there is an attack out there that's called, ShyHulu, and it's leveraging that NPM packages. So they're they're targeting developers
for NPM packages, then they are distributing malicious NPM packages to the max of the developers. And then there are tools like TruffleHawk that are scanning for secrets in GitHub repositories. And then you see an a login of an of an application with a correct password out of nothing, with a correct secret and so on. And it's, yeah, almost impossible to monitor. So and and even on the endpoint perspective,
so yeah. Developers are not the best protected part of most environments and lock in lock ingestion into your theme from the GitHub and so on. This is also not very common. So yeah. And at the end, we have also as I said, we have to do some preventive stuff and we have to do some detection and prepare for response. But as the best prevention is to get rid of these secrets or passwords. When I think of nonhuman identities, I probably tend to hear the most is, to your point, someone leaving a secret
in their code or they go in and check some code that has that secret into a public repo and it's like, oh, but I pulled it down five minutes later after I realized it. You don't know who saw that in those five minutes. Yeah. I feel like that's where I hear the most compromises when it comes to SPNs is due to those secrets and just I don't know that I wanna say carelessness, but just accidents happen. Right? Probably everyone whether they want to admit it or not. Right? Is guilty of like sending a password or something over Teams
and be like, I'll delete it right after. Right? But I mean, there there's absolutely, you know, risk in there. I mean, and I I've been guilty of the these things, you know, myself. Right? Yeah. I mean, the secret part I think is also tough. Just just made me think of earlier when you were asking Chris about detection. It's one of the other things with secret management as we we now are also starting to talk about, like, well, let's lower the lifetime. Right? Because it it'd be like, well, I'll go ahead the certificate as a secret, and I'll just have it not expire until I'm retiring. Right? I'll make it somebody else's problem because we all hate rotating search. But now we're we're going
down this path of saying like short lifetimes whether it's certificates or secrets which is awesome but now it creates I mean, in my eyes, right, even more noise if we're saying every every week, right, that that credential should be rotated. Yeah. And if we have different policies for lifetime, for secrets, and for certificates, maybe we are nudging the developers to rethink if a certificate would be an alternative. Yeah. Yeah. So when you work with customers, do you it I'm sure it varies customer to customer, but do you guys kinda
have like a recommended, you should maybe we're not gonna go till you retire, but it's okay to go a year, two years, or I know passwords like the NIST is now saying, oh, just rotate it when there's a risk detected. Yep. Yep. I mean, I don't know, like, I think certificates like the Microsoft recommendation now what I think is six months for the lifetime on them. I mean, I I would say in ways, right, you could be more aggressive than that. I mean, honestly, like, I'm not a developer so I don't wanna pretend that
coding is easy. But in other ways, right, like, I think sometimes people don't put the effort into the fact that, like, if you own if you have control over the currency graph, it's just like a few graph API calls to ask for a new one. Right? And and I mean, if you see a big identity systems where they have certain workload stuff like, I think Spiffy Inspire do things where it's like, you know, they're they're like rotating the secret like daily. Like that is used behind the covers for some of this stuff. I'd like to pick up the point, change password only if you have a reason. So this is
something for user accounts, not for admin accounts. And we should delete workload nonhuman identities more than admin accounts than as user accounts. And I think that don't share don't rotate passwords is there is a yeah. I think we all think or come to the conclusion, a password is not enough. And we have additional stuff like MFA, device compliance, and all of those things. And, yeah, we don't have it in place in most
environments for nonhuman identities. So the only things that is between an attacker and the and the resources is just a password at that point. Yeah. And I think, yeah, we should try to get rid of passwords. We should decrease the lifetime. And and, yes, we should introduce stuff like conditional access for workloads. So Yep. Yeah. I mean, I was just saying not not to increase my my Microsoft stock value but like for like a lot of people don't do workload identity protection. Right? Because it's like
a pay for thing. But I would just say and I try to tell people like if you can identify what is considered like the high privileged risky workloads. Right? I mean, it's, like, $3 a month or something. So say you have 10 or 15 service principles, like, start there. Right? It's only, you know, maybe a $100 a month to protect those. Right? But you could be potentially protecting the keys to your kingdom with, you know, conditional access for workloads and the identity protection features. So That's an interesting one too. Is Microsoft starting to do There's a couple aspects
there. They're starting to do a lot more with conditional access for non human identities. Around SPNs, Ignite, couple weeks ago they said you can start doing conditional access now for agents which are just Yeah. Yeah. Look, agents are exploding in creating just a massive new influx of additional nonhuman identities. Do you work with clients around the conditional access and around some of the how do we start thinking about agents? Kinda two questions in there. It it it starts. So I think we are in the beginning of that. So there's just a big
interest out there. So and our first message is the building blocks for agent IDs are human and non human identities. So they are not reinventing the wheel. They're reusing the building blocks that we already have. And if you have a good practice for protecting your human and not non human identities, you have a very good starting point for all that agents. But, I mean, I would just say with agent IDs and and workloads specific to conditional access, I think it's so when it's like being open minded. Right? Like,
in for a long time, we say zero trust does not, you know, network. And people are like, well, I'm just gonna make my conditional access policy for users say, well, if you're coming from corporate network, like, don't MFA. And we're like, that's an awful idea. Right. But if you look at workload, right, like, workload because it's different and agent ID included here. Right? If you have the ability to say, if it's coming from, you know, this, you know, data center, the this egress. Right? Network location is one of the few things that you can use. Right? So it's like, in
these cases, it's better than nothing. Right. You can't really MFA a SPN. That would be Right. Right. And I mean Certificates are MFA. Yeah. I mean, but I see I've seen people ask, like, weird things like, can you assign FIDO two keys to you know Yeah. And and yeah. But it like sometimes I think I'll see people say well we're not supposed to do network for users so why would we do it for workloads and and almost like right maybe we've caused the own problem as security people by saying don't do that so much.
But I think it's like looking at it right well, it's it's better than nothing. Right. It's one of the few things we have. So Yeah. And I'd like to add something. I think so we will see what happens with that agent IDs, but I think we will have a lot more non human identities. So there will be much more agents and applications, and we already have in many environments more applications and user. So just from the quantity, we have a bigger problem there. And I
think the permissions of agents will be a little broader because you don't exactly know what is really needed. So and I think over time, the permissions of the agents will extend. So we have more of them. They have more permissions. So, yeah. We have to take care of it. Yeah. Because I think especially with agents, to your point people are like an agent's not any good if it doesn't have access to data. Right. People are like I want it to have the access to the most amount of data so my agent's good or this data or that data.
And we've already seen, right? Copilot came out and everybody's like, Oh, all these security issues. And it's, no, it doesn't create new security issues. It just It identifies. It highlights quicker Yes. The security issues you already had. Yep. Yep. And we're gonna continue to see that and I think to your point it's just going to create a whole new security issues are gonna become more easily quickly visible. Yeah. But yeah, I mean that's like I feel like I'd see so many people you you had the similar problem with with Delve. Right? But like Yep. You know, point at SharePoint and then all of a sudden
people like, how come I can see all this corporate data I'm not supposed to because it's indexed and you can just go write query. But that's not, like, AI's fault. It's that you have, like, everyone has access to everything permissions going on. Right. If you would've used SharePoint search or Microsoft three sixty five search, you probably could've found them. It would've just taken a little longer. But no. I mean, it it's definitely, like, not even just a Microsoft, but industry thing that I feel like people are noodling on. Right? Because we have this issue where, yes,
you wanted to have permissions, but then you say, well, you only wanted to have permissions when it needs them. But also if if you're hammering a user constantly saying, like, do you want the AI to still have permissions? Like, you have fatigue and people just start hitting yes all the time to things. Right? I mean, that's like a lot of consent grant issues where because nobody reads the thing. Right? You're just like, I just want it to work. Yes. Yeah. Yeah. So it's like trying to figure out how how can you let AI do its thing, but also not let it just, like, run
away with access that is, you know, out of control. So and I don't know if there is, like, an answer out there yet for a lot of this. So Yeah. I think we can be happy that Microsoft has just launched that that AirBad stuff, and we will have conditional access and all of that stuff so that we we can, yeah, build good practices and guardrails and all of that stuff. Yeah. Do you feel overwhelmed by trying to manage
your Office three sixty five environment? Are you facing unexpected issues that disrupt your company's productivity? Intelligink is here to help. Much like you take your car to the mechanic that has specialized knowledge on how to best keep your car running, Intelligent helps you with your Microsoft cloud environment because that's their expertise. Intelligent keeps up with the latest updates in the Microsoft cloud to help keep your business running smoothly and ahead of the curve. Whether you are a small organization with just a few users up to an organization of several thousand employees.
They want to partner with you to implement and administer your Microsoft cloud technology. Visit them at inteligync.com/podcast. That's intelligink.com/podcast for more information or to schedule a thirty minute call to get started with them today. Remember, Intelligink focuses on the Microsoft cloud so you can focus on your business. Anything else that
you wanna talk about, you wanna make users aware of when it comes to managing those nonhuman identities and and try. I think we covered some of that credential management, the permissions. I think we touched on ownership a little bit, some of the agents. Anything else? I think when I look at our presentation from yesterday, there are two things that are, yeah, problematic. And we already discussed the first thing that was the owner assignment, get rid of it. Yep. So use administrative roles for this. So like a cloud application
admin, scope it to the dev. And the other part was, yeah, multi tenant apps are really challenge. So and we see we we saw a funny incident with, yeah, multi tenant apps by accident. So I think the Bing Bang story was one of the more known or someone it's just a radio button. Who is able to sign in? And Yeah. I mean, I I I tell you like with that. Right? So Bing Bang was when Wiz did research and found they could basically they they got into, like, some Bing back end and they could edit, like, search results.
Got it. But yeah. Right. It it's like a radio button, but a lot of times you'll see devs again maybe sort of jumping on dev. See, right? They'll they'll pull up basically all the authorization sort of checks on enter. Right? And, like, people conflate authentication and authorization. So they won't have any logic in their app and they're just like, well, if they've authenticated, then I guess they should be here. Right? And now it's a multi tenant app that shouldn't be. And so you end up with anyone. Right? If they stumble across a thing, can go sign into it. And interestingly
enough, you'll see a lot of times devs won't invest effort on, like, an admin sort of interface. Right? So maybe they have all this authorization stuff on a front end where all their customers are logging in and then there's some other admin portal thing that if you can get into it, right, you just have free reign to do whatever you want. And again, that's like basically what Wizz found in a lot of their the research. I mean, they were going after Microsoft stuff specifically. So Got it. I think the common the main problem with multi tenant apps is that we have a lot of controls
in the app that where is the service principle, not in the app where the app registration is. And the attacker can have its own service principle and has control over assignment and role mappings and all of that stuff. And, yeah, Eric just yesterday showed those in no hours attacks, so you can find it also on, it's on your on your blocks. Right? Yeah. And and we know of, is on the December's blog. I I would spend an hour trying to explain it because I ramble about I mean, I I think I think a
simple thing with with multi tenant apps is like, if you are not a software developer, like, as a business, you probably shouldn't have the app there. And and that is something again that you could use, you know, free tooling, paid tooling, right, to go find those sorts of things. So Yeah. So and if you are troubleshooting and switching configurations of your app I've seen that a lot. Don't touch this. It won't solve your problem, but bring you new problems. Yeah. Yeah. And and as I was saying, again, it's good to do reviews. When I I'm doing re
entra reviews, so we are selling this as a as a package to customers. So and one of the main things I'm looking for, okay, do you have I'm always upfront asking, do you have any multitenant apps? No. We don't have. So and then I just turn turn on the tools. Okay. We found 20 of them. And usually, you just can turn them off. And when there are some that are not by accident, you have to endure that people running the application are doing additional stuff. Yeah. So Yeah. And I think
that highlights again, like we talked about earlier, the need for security and developers to talk to each other. Yeah. Whether developers don't even, some developers may not even know, I don't even know what the difference is between single tenant and multi tenants. And really having that conversation and I've seen this across all of Microsoft March is one thing I feel like the cloud has done is it's caused IT people that don't typically like to talk to each other Yep. To have to talk Yep. Because all this stuff is so tightly integrated and has some
of those broader reaching effects. Yep. No. I mean, absolutely right. Like, you start to see conversations about soft skills and communication, all that. Right? Like, coming into the industry because, like, sometimes the best security person isn't because you have the biggest brain. It's because you can work across the organization, right, to help implement strong security practices amongst your developers, amongst your end users. Right? This isn't like the old days where you could be that bastard operator from hell, you know, and just tell people to do
this thing. Right? Because then you get shadow IT and and all sorts of stuff. So, yep. Awesome. Well, thanks guys. Yeah. I I appreciate the conversation, talking a little bit more about non human identities. We'll include a bunch of links in the show notes. I'll get links from you guys to social media where you want, if people wanna connect with you there, websites. If you have links to tools, I'll get some of those that we can put in the show notes as well. Any last things you wanna share, highlights, telling people to come to this conference next year or 2027,
I guess. It was really a nice event, so Yeah. No. I mean, this has been great being at the inaugural Workplace Ninjas and hopefully I mean, we're a little ahead. I'm I'm hoping I can claim now that I will become the intra idol from the inaugural workplace Oh, do I need to go I have not voted yet. Do I need to go vote for You need to go vote for me, Ben. You have to you have to vote for Eric. Yeah. But, no, I mean, as far as myself, I'll be actually over in Europe at MC2MC and then
Hip Europe a couple days after that and then Experts Live Denmark. We're co presenting Chris and I at MC two MC so Okay. Are you gonna go most important question if you're going to Denmark. Are you gonna go to the Lego headquarters while you're in Denmark? Oh, I do. Are you a Lego guy? Yeah. Yeah. It's not really close to Kommang. Is it? Oh, I I have no idea. But, yeah, I I am a Lego guy, so You can go get those secret Lego sets. There's what? The two LEGO sets that you can only get at LEGO headquarters in Denmark.
Oh, I did. Oh. I'll have to how far is it? I think for for Americans it's not that far. I will go rent the biggest SUV in all of Denmark. Go buy Allego. Right? We're in Texas. We're like Texas is huge over there. Everything's closer than driving across Texas. So, yeah, I'm really looking forward. So this year was also was already a cool event in in Copenhagen, but and and next year
will be a really huge, I think. So Yeah. And but I also in enjoyed this web with New Year's US. It was also a very nice event. And, yeah, we have a full second day. So Yeah. Yeah. Alright. Well, I will let you guys go enjoy your second day. Thanks again. Enjoy the rest of the conference. Awesome. Talk to you guys later. Thank you. Bye. If you enjoyed the podcast, go leave us a five star rating in iTunes. It helps to get the word out so more IT pros can learn about Office three sixty five and Azure. If you have any questions you want us
to address on the show or feedback about the show, feel free to reach out via our website, Twitter, or Facebook. Thanks again for listening, and have a great day.
Welcome to Episode 441 of the Microsoft Cloud IT Pro Podcast. In this episode, Ben and Scott take a first pass at Microsoft Ignite 2026, starting with a question a lot of IT pros are asking: is the pace of change...
Welcome to Episode 440 of the Microsoft Cloud IT Pro Podcast. Recorded live at the North American Cloud and Collaboration Summit in Branson, Missouri, this MS Cloud on the Road conversation has Ben sitting down with Jay...
Welcome to Episode 439 of the Microsoft Cloud IT Pro Podcast. In this episode, Ben and Scott dig into two Azure agents with confusingly different names: the Azure Copilot Troubleshooting Agent and the Azure SRE Agent....
Welcome to Episode 438 of the Microsoft Cloud IT Pro Podcast. Back in Episode 429, Ben and Scott introduced LLM wikis. In this episode, they come back a few months later to compare what those wikis turned into after...
Welcome to Episode 437 of the Microsoft Cloud IT Pro Podcast. In this episode, Nick Burling, Chief Product Officer at Nasuni, joins Ben and Scott to talk about getting unstructured file data ready for AI. Nick digs into...