Product Office Hours - AI Governance: What is it and what can we do with it? Add your questions here!

1. 🤔 If you were speaking to our leadership team and managers, what would you discuss with them about the importance of AI Governance? Why should this matter? Keep in mind, they may not read this as a new technology stack, but as new cost they cannot afford. 
2. 💭 Does Jamf itself use AI inside its management software, and is the IT telemetry handled safely?
3. 💬 What if auditors ask if our company is using AI safely? How can Jamf help us here? 
4. 👩🏻‍💻 Does AI governance slow down employee productivity? Does AI governance offer granular controls, like a different profile or tier for different departments?
5. 🧑‍🏫 Who do you see managing this? Security teams or admins? How do you feel about it being so intelligent one day that AI manages AI governance?
6. 🤖 The AI landscape is changing as we speak and listen now. The AI field will look different in 1 year, and certainly by the end of the decade. How is Jamf staying on top of the latest AI trends and tools for governance? 
7. 🎧 I have enough work as it is. I don’t want to listen to your speech about AI Governance. That is someone else’s problem! I won’t look you in the eyes as we pass each other in the hallway. How would you get my attention on this matter? What would you say?

Jamf AI / Device Data – Questions for Jamf

1. Jamf Protect dependency and licensing

   • Is Jamf Protect required to purchase or use this capability?

   • If we do not have Jamf Protect, can we still collect and use the associated device/AI activity metrics and telemetry?

   • Which data-gathering and reporting capabilities are available without a Jamf Protect license?

2. Platform support

   • Is this capability limited to Apple platforms, or does it support other operating systems?

   • If multiple platforms are supported, which operating systems and device types are currently supported?

   • Are there any differences in functionality or available telemetry between Apple and non-Apple platforms?

3. AI governance and device controls

   • Can we create custom profiles, policies, or configurations to govern or regulate AI applications and AI services on managed devices?

   • Can those profiles or policies be deployed directly to devices through Jamf?

   • Can we define organizational controls around which AI applications are permitted, restricted, or monitored?

   • Are these capabilities limited to Apple devices, or will equivalent controls be available for Windows and other supported platforms?

4. Architecture and authentication

   • Is this capability implemented as a microservice within the Jamf Platform?

   • Does it require Jamf SSO or a separate authentication mechanism?

   • What Jamf services or licensing components are required for the service to operate?

5. Jamf AI Assistant integration

   • Is the collected data queryable through the Jamf AI Assistant?

   • Can administrators use natural-language queries to retrieve device-level AI activity, compliance, or risk information?

   • What data sources are currently exposed to the AI Assistant?

   • Are there plans to expand the AI Assistant's ability to query this data OR to action on it?

6. API and integration capabilities

   • Is there an API available for accessing the collected data and metrics?

   • Is the API part of the Jamf Platform API, or is it a separate product-specific API?

   • What authentication methods are supported?

   • Is the API read-only, or does it also support policy/configuration changes?

   • Can the data be integrated with external platforms such as SIEM, data lakes, reporting platforms, or ServiceNow?

   • Are webhooks or event-based integrations available for real-time data?

7. Data ownership, retention, and privacy

   • Where is the data stored and for how long?

   • Can data retention periods be configured?

   • Is customer data used to train Jamf or third-party AI models?

   • Can the organization control which telemetry or AI-related data is collected and retained?

How do we get started? Is there a guide or set of recommendations?

Thanks Chubs for sharing this, it really helps me. I will try it.

Cyber insurers are beginning to ask organizations how they govern and control the use of AI. How does Jamf’s AI Governance help customers answer those questions today, and how do you see it supporting future cyber insurance requirements?

Great questions ​@mvu! Answers below:

1. 🤔 If you were speaking to our leadership team and managers, what would you discuss with them about the importance of AI Governance? Why should this matter? Keep in mind, they may not read this as a new technology stack, but as new cost they cannot afford. 
   1. I would say that being able to really understand, set policy, and govern AI is going to be extremely important in the future. Not only that, it is here today and most people are actively trying to figure out how to do it. The truth is, many employees and students are using AI. Organizations are either trying to block everything or they are opening the doors of exploration with limited controls of what is happening at large. This is present only from foundation models but local models as well. Furthermore, most organizations are trying to better understand how they can use AI to actually create better outcomes, not just usage. Jamf is in a unique position to give you on-device capabilities to better understand visibility of what AI is being used today. This will help you set policy and execute governance for responsible, productive, and cost effective use of AI now and in the future.
2. 💭 Does Jamf itself use AI inside its management software, and is the IT telemetry handled safely?
   1. Two angles at this, but the overall answer is Absolutely. Jamf both uses AI internally as well as in the products we offer. The AI Governance solutions we discussed during this session is accessing and transmitting data in the same manner as we do within our overall Jamf Platform nothing net new is different about that. Additionally, Jamf also has the AI Assistant available for use. That seems a bit more aligned with the question above. More information on the architecture and security for that can be found here https://www.jamf.com/resources/technical-papers/ai-assistant-architecture-security/
3. 💬 What if auditors ask if our company is using AI safely? How can Jamf help us here? 
   1. This is a perfect use case for Jamf’s new AI Governance. This solution not only tells you what models and harnesses you are using, it also gives a breadth of controls to ensure safe and secure usage. For example, you can control which MCP connectors they are able (or not able) to use as well as control the shell commands the agents use.
4. 👩🏻‍💻 Does AI governance slow down employee productivity? Does AI governance offer granular controls, like a different profile or tier for different departments?
   1. AI Governance gives you the capability to use and maximize speed of AI adoption while also adhering to your policies. It uses blueprints for deployments, so you can act dynamically across your devices. This is the same intelligent scoping that Jamf is known for, so you can get as granular as you need it to be. As with everything at Jamf, it is built Apple first, Apple best, and the native integration means the overhead is minimized so your organization can be assured you are adhering to policy and your users feel like they are getting the experience they wanted from the Apple device.
5. 🧑‍🏫 Who do you see managing this? Security teams or admins? How do you feel about it being so intelligent one day that AI manages AI governance?
   1. AI Governance is an opportunity for IT and Security to help each other. This solution gives the telemetry needed to better understand how AI is being used on the device-level which can include much better insight than the network alone. With better information, teams can make better decisions about what users want to do vs what is acceptable use. This will enrich how the organization balance productivity and the value of AI with security needs of the organization. Furthermore, we have built the AI Governance capability intentionally in an interface adjacent, but not inside other tools such as Jamf Pro or Jamf Protect. That will allow multiple stakeholders to access the insights and actions without being overburdened by one system or the other.
6. 🤖 The AI landscape is changing as we speak and listen now. The AI field will look different in 1 year, and certainly by the end of the decade. How is Jamf staying on top of the latest AI trends and tools for governance? 
   1. This is moving very fast, and it probably isn’t going to get slower in the future.  Jamf is at the forefront of this evolution. We maintain close partnerships with major vendors in this space to understand where they are headed and consult on how to better serve your needs with features and functions that are most valuable to what we hear and see from Jamf Nation. These same partnerships give us capabilities to better align with the changing security landscape as well. Our Threat Labs team puts us in a unique position to not only understand how this impacts Apple devices and their use, but also proactively build and protect against trends in real-time. And of course we are using AI to help us build these solutions, so that means the delivery of updates and features come faster than ever.
7. 🎧 I have enough work as it is. I don’t want to listen to your speech about AI Governance. That is someone else’s problem! I won’t look you in the eyes as we pass each other in the hallway. How would you get my attention on this matter? What would you say?
   1. First, let's all take a deep breath. Then know we are here for you to make this easier and not harder. :) 
   2. AI is also here… now. And it is changing the landscape. We don't know exactly what the future will hold, but we do know the game has changed. And just like every other technology that has come our way in the past 50 years, we can use this as an opportunity to define how we can better help ourselves, our users, and our organizations get the most of that capability. That decision is up to us. We can get on that ride and have some fun, or we can stand on the sidelines and watch. So let's go.

Hi ​@ktrojano! Getting started is easy. Check out some of the great materials we have created to help shape this up below:

Jamf Short: https://www.youtube.com/watch?v=WGW7uTGtbX0
Enablement Guide: https://learn.jamf.com/r/en-US/ai-governance-configuration-guide/AI_Governance
AI Webinar: https://www.jamf.com/solutions/ai-governance/

And of course, community is key. We have a new AI for Admins group you can join and, of course, keep coming back to these series. 
 

Thank you, ​@sam! That’s awesome. And thank you Jamf for holding this Product Office Hours session.

Will you have some things at JNUC in the Expo Center? Was hoping for more sessions around AI.

Many thanks again!

​@mvu, Yes! Expect quite a bit of activity around this at JNUC. Looking forward to seeing you there.

​@Chubs, your questions are thorough per usual! Good on ya. 😅 I did my best, but there were even a few I will need to phone a friend on. 

This is a great use case for the new AI Governance capability. We built this tool to give you visibility, control, and evidence. We talk a lot about the first two, but evidence is key here. We currently have a templated “Executive AI Posture Report” as well as a full audit trail with every policy decision, deployment, and enforcement action. That can be extremely useful for these situations. As this area develops, we want to make sure we are providing everything you need, so please provide feedback on anything more that you may need in this area. 

Hey ​@sam!

When will we be able to create custom AI Governance policies? The predefined policies are a good starting point, but every organisation has different requirements. Will Jamf allow admins to define their own rules and policies in the future?

Good news. You should be able to do this now. Even the predefined policies are really just a starting point and you can pick and choose what you want from there in the AI Governance policy controls.

• Log in to the Jamf Account Portal using your Jamf ID.
• Navigate to AI Governance in the sidebar.
• Select the target environment from the pop-up menu if you manage multiple tenants.
• Click AI Policies and choose to create a new policy for your chosen tool (such as Claude Code, Claude Desktop, or OpenAI Codex).
• Configure your manual settings such as restricting specific AI models, enforcing OAuth authentication, or pinning update channels. (You can also upload a custom JSON configuration file if you prefer manual code-based definitions).
• Save and publish the policy