[Day 3] Dedicated to Dedicated: Non-negotiables for EMM/MDM in Rugged Android Deployments
Disclaimer: The following article captures the opinion of Matt Dermody, Senior Director of Enterprise Mobility at Manhattan Associates. The stances contained within are a reflection of Manhattan's specific focus on line-of-business Android devices, built on years of being "Dedicated to Dedicated." Background Manhattan Associates is a B2B software company specifically focused on best-in-class, line-of-business enterprise deployments of enterprise software such as Warehouse Management (WMS), Transportation Management (TMS), and Point-of-Sale (POS). These software deployments command high expectations of uptime and availability, and that ultimately encompasses the complete solution, including the mobile computers that the software runs on. Manhattan is dedicated to ensuring that our end customers have the best possible experience, and that involves ensuring that the dedicated devices running our software are also properly maintained and supported. Google defines a "dedicated device" as a company-owned device that is fully managed and locked down for a specific work purpose, often using a single app or a small set of apps. These devices are restricted from personal use and are used for business functions like point-of-sale systems, inventory scanners, or digital signage. Or in other words, all of the device types that Manhattan sells, deploys, and supports alongside our software solutions. In that sense, I guess it can be said that we are… Dedicated to Dedicated. Managing rugged, line-of-business Android devices is not the same as managing BYOD phones and laptops. These are mission-critical endpoints running specialized apps in warehouses, stores, yards, and DCs—downtime costs money and local IT is increasingly rare. Your EMM/MDM must control versions, files, firmware, and field support with precision. Anything less adds risk and operational drag. Situation Imagine having to explain to a CIO that a business-critical mobile app has automatically upgraded to a new version that breaks functionality, and there is no easy rollback available. Here is a preview of how that might look. That situation is all too common but can be prevented with the right EMM/MDM strategies. The mere thought of that possible situation keeps the Manhattan team up at night. We have spent years developing strategies to add predictability and stability into enterprise device deployments to prevent bad situations from ever happening. Philosophies & Strategies Here is a preview of some of the core philosophies surrounding Manhattan’s tailored approach to managing mission-critical device deployments. Some of these might be controversial, but these are the strategies that work for us. 1. App Distribution and Version Discipline Rigorous version control of enterprise apps—stage, canary, bulk rollout, and rollback—is a must. Rugged ops cannot afford “surprise” app updates or version creep. If you can’t downgrade quickly, you don’t control your risk surface. An EMM/MDM should offer direct installation of private APKs on fully managed devices. Auto-upgrades to the “latest only” through Managed Google Play can lead to instability and version drift. Look for a console that can deploy specific app builds to specific groups on your schedule. If your tool can’t install an APK directly onto devices, it’s the wrong tool for rugged. Period. You must target different versions/configs by environment—Stage, QA, Prod—often per site group. That includes app versions, config files, etc. 2. File Management for App and Scanner Configuration LoB apps often externalize key settings via JSON and similar external config files. For example, Zebra DataWedge uses .db files placed in a specific auto-import directory to control mission-critical scanner settings. Your EMM must place, update, and replace these files on demand and at scale—ideally without anyone touching a device. Emergency changes (host cutover, DNS rename, scanner tweak) should be a file push away, not an onsite scramble. 3. Remote Control and Log Retrieval Treat full-fidelity Remote Control as table stakes. Support must see what the user sees, drive the screen, and pull logs and files in one session. Anything “view only” or bolt-on only erodes speed to resolution. Relying on reports from the floor or grainy pictures of an error taken from another device are not sufficient tactics for troubleshooting mission-critical device deployments. When issues hit, you don’t want an insurance policy like Remote Control that can be used to quickly diagnose and test solutions; you want a tool. An EMM admin without Remote Control is effectively blind with their hands tied behind their back. 4. OEM-level Controls (Zebra/Honeywell) There are numerous configuration settings that enterprise-grade OEMs extend beyond the baseline Android Enterprise configuration APIs. These OEMs are generally years ahead of what is available in base Android from a configurability standpoint and often introduce configuration settings that may otherwise never arrive to the base OS. These granular configuration layers ultimately are what set enterprise-class devices apart from consumer-grade technology. It is therefore imperative that an EMM managing these devices has the capability to manage OEM configuration extension features directly. For Zebra, this involves execution of their MX XML, DataWedge behavior, button mapping, radios, and other rugged-specific controls—through native profiles or integrated mechanisms. OEMConfig is useful, especially for parity across EMMs, but you will hit practical limits in closed networks and with Play-dependent timing/visibility. OEMConfig is a lowest-common-denominator functionality that was designed as a bridge to enable limited AMAPI-aligned EMMs to manage OEM-level settings with the limited tools at their disposal. Your EMM should support both OEMConfig (at a bare minimum) and offer the flexibility of direct MX/file workflows so you’re not boxed in by the limitations of distributing device settings through a complex web of Google Play server infrastructure. Your EMM should offer the ability to manage settings directly on the devices it manages, without the added layers and black boxes of complexity. 5. Firmware and Security Patching Over-the-Air (OTA) upgrades are great, but only when the EMM admin is in complete control. Auto-upgrades from the OEM pushed out over the air can bring production to a halt when critical business functions break. At a bare minimum, they can bring a network to a standstill as large upgrades are forced through the ISP connection into the building or site. An EMM should therefore offer integrations with the OEM-specific OTA and/or firmware upgrade protocols to put the controls in the admins' hands. 6. Lockdown and Kiosk Modes Rugged devices should boot into the work, not into Android. Enforce kiosk/lockdown, strict app allow-lists, settings restrictions, and consistent UX across every DC and store. The EMM should offer configurability over what is displayed on the lockdown, including personalization and customization to offer links to additional items such as launching apps, toolbars, or script executions. 7. Enrollment that Fits the Reality of Rugged Use Android Enterprise Device Owner (AEDO) with a barcode-driven process (e.g., Zebra StageNow). It’s fast, repeatable, and minimizes user taps and mis-taps on the floor. Wi-Fi credentials can be encrypted in the barcode rather than shared haphazardly and manually entered by end users into the Setup Wizard. More granular control over initial network connectivity is also afforded as compared with the limited options available through DPC extras if using the designated AEDO QR method. Avoid Zero Touch Enrollment (ZTE) for rugged Wi-Fi-only devices. ZTE is not "Zero Touch" as it realistically pushes many touches (and possible errors) to end users. There is overhead and maintenance to unenroll and re-enroll devices into the portal as they go in and out of repair. Enterprise-grade devices are often covered under repair contracts due to the nature of the environments they’re used in. This means they are going in and out of repair relatively frequently, and ZTE portal management ends up causing more bottlenecks than the steps it’s otherwise designed to free up. StageNow barcode flows are fewer steps and far more reliable for DCs and stores. 8. Closed Networks and Offline Constraints Many rugged sites have limited or no access to Google services. Your EMM must support managed app configuration and device policies in ways that don’t depend on real-time Managed Play orchestration. If your only path is Play-mediated, you’ll struggle with timing, visibility, and outcomes. Look for an EMM that offers “offline” or standalone Managed Configuration support by reading and exposing the configuration schema of an uploaded Enterprise app. 9. Health Analytics, Drift Detection, and Scripting Device health analytics (compliance, connectivity, install status) are critical for early detection and fleet stability. Pair that with a scripting engine and policy-driven rules (e.g., automatic relocation, auto-heal) to keep devices in line without manual human intervention. 10. What to Deprioritize (and Why) BYOD-centric EMMs that can’t directly install private APKs, can’t push files, and don’t include Remote Control as a first-class capability will drag deployments and support. Many EMMs specifically lack the granular APK/file control, versioning/rollback discipline, and integrated Remote Control required for rugged Android in DCs and stores; workarounds add fragility and cost without closing the gaps. Bonus – Identity and SSO Newer EMMs are offering advanced capabilities around Identity Management and SSO across business apps. As enterprise-grade devices become more multi-purpose, more mobile apps are being installed, each often with its own separate login requirements. Over time, there will be increasing needs to supply SSO workflows on-device across these business apps and to offer a clean pathway to script and automate the cleanup of a prior user’s session across all apps as they log off and make way for the next user to log in. If in the EMM selection process today, look for an EMM that offers these capabilities. Even if those features are not needed today, it is almost certainly the next set of features enterprises will look for and need to adopt. The Quick Scorecard If you can’t answer “yes” to these with your selected EMM/MDM, you’re taking unnecessary risk: Can your EMM install a specific APK build directly to AEDO devices? Can you canary a new version to one site, schedule a 2 a.m. cutover, and roll back instantly if needed? Can you push a JSON config change and a DataWedge .db to 500 devices in under 10 minutes—no manual touches? Can support remotely control the screen and pull logs/files from the same session? Can you execute Zebra MX XML, enforce kiosk/lockdown, and set scanner behavior centrally across models? Can you deploy LifeGuard/.ZIP OS updates by group, with maintenance windows and rollback? Can you enroll with StageNow barcodes (AEDO) instead of relying on ZTE flows designed for non-rugged scenarios? Can you operate cleanly in sites with limited/blocked Google services, including offline managed config workflows? Bottom Line A capable rugged EMM/MDM gives you deterministic control over versions, files, firmware, and front-line support—at fleet scale and on your schedule. Prioritize direct APK delivery, file distribution, OEM-level controls, Remote Control, AEDO barcode enrollment, and firmware orchestration. Deprioritize BYOD-first tools and any workflow that forces you through black box Play timing or pushes enrollment burden to associates on the floor. I’d love to hear what the comments have to say. Am I way off base? Do you fundamentally disagree? Or were you nodding along as you read through this. Let me know below! Oh and "AI", forgot to mention the buzzword. Matt
Zero Touch phones randomly wipe themselves
Hello, We are a large corporate and mostly use Samsung phones as Android devices. Enrolment is being done via ZT portal to a default profile which is Corporate Owned Work Profile provided via Microsoft Intune. We are noticing an increased amount of cases where users set up their phones (no QR code, no text token) with default configuration added using DPC extras and within first few hours they would reset to a factory default state without any notice. This has become a real issue as it is affecting more and more people. Devices enrolled without ZT do not suffer from this issue, even though they are using the exact same enrolment profile. I saw many posts like this here and elsewhere on the internet, but no actual solution. What is the problem here and is it being actively looked by Google?
Samsung Devices: Can't call from a personal app
Hi everyone we received some reports from our users in the last couple of month that suddently the phone app on COPE devices (Samsung A-series) starts to show "Can't call from a personal app" - Your organisation only allows you to make calls from work apps. Workaround: Reboot the device. For most of the reports this workaround has to take place once and the message is gone forever. A very small amount of devices starts to show this message again after a couple of weeks. Rebooting is resolving the issue again. Any idea of how to prevent this? Even emergency calls are not possible if this error is appearing! Does anyone else have seen this behavior? Raised a case with Samsung today. Thanks! Daniel
AMAPI prepareEnvironment() failing with ApiLevelException on Android 8 despite using DEVICE_POLICY_CONTROLLER role
Hi everyone, I’m implementing a custom DPC (device owner) and integrating AMAPI locally on the device. On Android 8 / 8.1, calling: val request = PrepareEnvironmentRequest.builder() .setRoles( listOf( Role.builder() .setRoleType(Role.RoleType.DEVICE_POLICY_CONTROLLER) .build() ) ) .setAdmin(admin) .build() immediately fails with: com.google.android.managementapi.common.exceptions.ApiLevelException On Android 10 and above, I don't have this exception. According to the AMAPI documentation: If the device's SDK API level is insufficient for certain requested roles (this may be in addition to a general minimum API level requirement for the call itself).{@code Role.RoleType.DEVICE_POLICY_CONTROLLER} requires API level 23 or above. Any other roles require API level 28 or above. I am using the latest AMAPI client library: com.google.android.libraries.enterprise.amapi:amapi:1.7.0 Questions Is AMAPI (EnvironmentClient + Device Policy Controller role) still officially supported on Android 8/8.1? Any clarification on the real minimum supported API level for AMAPI prepareEnvironment() would be greatly appreciated, as the documentation suggests Android 8 should work, but the behavior indicates otherwise. Thanks!
[Day 2] Mission Intune : When Migration Becomes a Mission (Almost) Impossible
Good Morning Everyone 🕵️ Deep within the digital infrastructure, a high-stakes mission is being prepped. Five mobility experts have been deployed to solve a massive puzzle: migrating tens of thousands of smartphones to Microsoft Intune. The Goal: Ensure a fluid, secure, and uninterrupted transition for thousands of users. The Battlefront: A complex landscape filled with legacy policies, mixed configurations, and strict deadlines. It’s a race against the clock where one wrong move could start a domino effect. From scripts to security protocols—nothing is left to chance. Failure is not an option. Following Broadcom’s acquisition of VMware in 2023, the Workspace ONE product is now owned by Omnissa. Broadcom’s commercial strategy, which has influenced its spin-off companies, had become highly aggressive toward all customers. Consequently, we have decided to migrate the management of our Android and iOS tertiary fleet to Microsoft Intune.. While we are familiar with Intune, several limitations should be noted: Reporting: Intune offers basic reporting through Microsoft Endpoint Manager and Power BI integration, but lacks the advanced, customizable dashboards available in Workspace ONE. Deployment Performance: Application and configuration deployments can be slow, with status updates often delayed due to Intune’s reliance on periodic device check-ins rather than real-time communication. iOS Management: Intune provides full functionality only for devices enrolled via Apple Business Manager (ABM). Non-ABM devices have restricted supervision capabilities, limiting advanced configuration and app deployment. Error Handling: Intune does not display granular error codes in its console. Troubleshooting often requires log collection from the device or use of Microsoft Support tools, increasing diagnostic complexity. Conditional Access & Compliance: Intune integrates tightly with Azure AD for conditional access policies, which is a strength, but requires additional configuration and licensing for advanced scenarios. App Protection Policies: Strong for Microsoft 365 apps, but less flexible for third-party apps compared to Workspace ONE. Migration Strategy Overview The project aims to migrate the entire mobile fleet—a few tens of thousands Android and some iOs devices—between September 2023 and December 2024. Cybersecurity requirements mandate a shift from COBO (with personal Google accounts allowed) to COPE, reinforcing corporate control and reducing exposure to security risks. Key Challenges Technical Constraints: Devices incompatible with Android 13 require hardware replacement. For most employees, migration involves full device reset and Intune re-enrollment—a complex, time-consuming process. Security Limitations: Backup tools cannot be authorized, increasing the risk of data loss and user errors. A recurring issue is failure to remove Microsoft Authenticator configurations, creating significant support overhead. Performance Impact: The Samsung Galaxy A32, previously adequate under COBO, performs poorly under COPE, affecting user experience. Status and Strategic Decision By June 2024, progress is far below target. To mitigate operational disruption and support overload, the strategy shifts: forced migrations are discontinued. Migration now occurs only during: Hardware replacement (obsolescence, failure, or breakage) Voluntary device reset This approach prioritizes stability and resource optimization while maintaining compliance with security standards. We’ve been with Intune for almost two years, we make do with it and we are hardly surprised anymore when something doesn’t work. If you have any questions, don't hesitate to reach out via the comments below Kris[Day 1] Mobile Devices With a Sixth Sense: What Android Can Learn From Detection Dogs
Good afternoon everyone! Intro Alongside my passion for Android, which I’ve also made my profession, I spend a lot of my personal time working on scent detection training with dogs. Over the years I’ve trained my own dogs to search for items such as data carriers, phones, cannabis, and most recently one on cash. I wanted to participate in the festival because I had to skip the opportunity last year. But to contribute meaningfully, I wanted to create something that connects both worlds, Android and my other interests. This article is the result of that cross-pollination. The article is just a different perspective to discuss, a thought I had and a look in to what I think could be a good future. Android & detection / search dogs Enterprise mobility is still too often reduced to policies, profiles, and compliance checkboxes. A device shows compliant, an app is locked down, and the job seems done. But anyone who has worked with a well-trained detection dog knows that control is only half the story. The real value comes from analyzing behavior and context, and the ability to anticipate on what’s coming. Fun fact: Our nose, and a dogs nose, contain olfactory receptors, nerve cells that detect odor molecules, which is what we use to recognize a scent. An average human has around 2 to 6 million of those. A dog’s nose has around 250-300 million. They are capable of detecting so much more scents than we do. A detection dog doesn’t just smell an object. It smells the contents, the ingredients of what it’s made of and It detects deviations. It recognizes not only what is present, but also when a situation doesn’t match the pattern it expects. If something has disturbed the soil, it will recognize that. And as a handler you should be able to read to signals and act on it. If you want to go right, and the dog is showing that it recognizes a scent on the left, you should really go left and trust the signals your dog is sending you. As a dog handler I’m trusting my dog to make the right decisions, I just follow and guide the dog where needed. Lift him to higher grounds, or maybe mark areas of extra interest that I can see and I’ve been told to search. Its teamwork. Devices as Sensors Imagine a device that doesn’t only enforce policy but also understands what normal looks like in its environment. Not only checking whether something is allowed, but noticing when something is unexpected. A phone that has spent months connected only to Wi-Fi inside the warehouse but suddenly appears on 4G at two in the morning in another city, that may not be a direct policy violation, but it is something you and I would ask questions about. Any detection dog would pause, tilt its head, and quietly signal that something’s off. The ingredients to make devices smarter already exist. Smartphones capture motion, location, battery patterns, network behavior, app usage, and user interaction. Individually these are datapoints, but together they form a pattern, just like scent particles form a track for example. The interesting part is: the hardware has been ready for years. What we lack is interpretation. Fun fact: Did you know that when a dog is searching/sniffing, it can inhale and exhale up to 300 times per minute? If we would do this, we will start hyperventilating within seconds. I think Android could evolve in the same direction by learning baselines of enterprise-normal rather than relying solely on static policies. Once a baseline exists, devices can flag changes proactively, early before things escalate. An example Consider a warehouse worker scanning goods along the same aisle, during the same shift, using the same three apps every day. Android sees that, learns it, and identifies it as normal. But one Monday everything is different: roaming is active, a new route is taken, unfamiliar apps are running. Instead of asking only is this allowed?, the device could ask is this unusual?, should I report this?, is this risk or intentional deviation? As an IT admin, you could check those signals and take appropriate action. But maybe we want Android Enterprise to take their own actions up to a certain degree? This isn’t just security, it also improves stability, efficiency and less downtime. Combine all these and you might even have an employee who is actually happy with the work IT is doing. Instead of being the team who keeps blocking things, you become the IT admin that makes the devices just work when they need to. Closing note I am aware of different MDM’s providing such solutions such as WS1 and Knox Asset intelligence. But I think it could and should be so much better than that. It should be part of core Android OS, present for everyone, not just the one who can afford it but also the smaller companies with less budget. It shouldn’t be depending on a third party whether or not this works. Android Enterprise has matured. Policies are essential, but they’re not the finish line. The real opportunity lies in devices that understand normal, and detect subtle deviations before users even notice. Maybe it’s time our Android fleets developed a sense of intuition. Maybe it's time for Android fleets to develop their own sixth sense like a detection dog that quietly sits, nose raised, because it notices something no one else does yet.
Unable to register device using Google Services Framework Android ID - Page unresponsive
I am unable to register a new device utilizing the Google Services Framework Android ID. Upon entering the ID and clicking the register button, the page is unresponsive. I have tested this on multiple browsers (Chrome, Brave, Firefox) with extensions disabled, but the issue persists. The page simply does not react.
GBoard - Suggestion Strip
Hi, We want to use GBoard on kiosk devices but we aren't able to remove the suggestion strip using managed configurations. All other settings can be configured fine though. The show suggestion strip configuration is set to disabled. But with versions 15.x and 16.x of GBoard it's still visible on the devices. And when checking the setting locally on the device it's still enabled (Disabling manually works fine) Back in version 14.x this configuration worked fine. Anyone else who has experienced the same thing? We've tested this on devices from Samsung, Bluebird, ELO, and Zebra. Android version doesn't seem to have any impact, just the GBoard version. // Magnus
Unable to upload bulk CSV file to ZeroTouch
Hi Team, Is there currently an issue uploading a bulk .csv file to ZeroTouch? It's giving me an error. See below. Steps below: I downloaded the sample .csv file then updated it with my data, then uploading it again to the portal as is without changing the name or file extension as seeing above, yet its giving me an error. This was working not long ago, just wondering if there is currently an issue. Thanks
Enable ADB debugging is grayed out - This setting is managed by your administrator
This issue was documented in 2021 but with no solution. My Chromebook is managed by my company and I am the manager. But Google tries to find the managed option to unlock for this to work in the administration interface for more than 15 days without success. By the way there are thousands of options in the admin interface it could be a clever feature to number them. If you are in front of the same issue please add your comments to this post. I hope that Google support will succeed to solve the issue soon because I developed my first app for Android on my Chromebook with Android Studio and I was able to download it to my phone before these 15 days.
