Turn insight into action, automatically with Omnissa DEX

Digital employee experience monitoring gives IT visibility into how devices, apps, and networks perform for employees. But seeing a problem and fixing it are still two different jobs, and the gap between them is where IT time disappears. At Omnissa, DEX AI is how we're closing that gap: a set of AI capabilities across the Experience Management product line that catches problems early, explains what's wrong, and increasingly fixes them before anyone opens a ticket.

At the center sits the DEX Agent. When something breaks, spotting it is usually quick. The hours go into what comes after: working out what's actually wrong and deciding what to do, while the employee who reported it waits. That's the part the DEX Agent takes on.

It investigates across data sources to find the real cause, picks a Playbook your team has already vetted (or builds a new one when nothing fits), runs the fix, and then confirms it worked instead of stopping once it ran. It's built on tools you already run, with Workspace ONE Intelligence handling detection and Freestyle Orchestrator handling execution. Investigation runs on its own, but nothing gets written or run on a device without an admin's approval.

Want the clearest way to see this in action? Follow one admin through three situations, with the same DEX Agent behind all three.

One agent, three situations

Situation 1: Seen it before

Insights v3 now runs org-specific models and catches step changes and trends on top of anomalies. It flags a spike in crashes before employees start filing tickets. Experience scores can start the same investigation when a group of devices slips below its peer baseline.

Here is the new Insights screen:

 Guided RCA v2 pulls from multiple data sources and uses AI to trace the problem back to a real root cause, not just a correlation. It checks which device models, OS builds, and app versions crash more than the fleet would predict, what devices were doing just before they crashed, and what else broke alongside them. An AI agent chooses which checks to run as the picture builds, then tests the likeliest cause against comparable devices that never got the change.

The DEX Agent takes that diagnosis, checks it against the patterns it's automated before, and recognizes this one because it already has a fix. It recommends the Playbook, the admin approves, and the agent runs it, pausing at any step that needs a human. When it's finished, it explains what happened and why in plain language.

Situation 2: Building the fix

Same detection, same diagnosis, but this time no fix exists yet. The Playbook generation service builds one on the spot, either a Playbook or a remediation script when that's the more direct route, drawing on Assist Session Summaries from past sessions where a help desk agent worked through this by hand.

The admin reviews it, approves it, and it runs like any other Playbook. It also stays in the library, so the next device that hits this issue gets fixed without anyone stepping in.

Situation 3: Starting from scratch

An admin asks the DEX Agent directly: "Why is this device slow?" Or a spike in ITSM tickets for the same issue kicks the question off instead. Either way, no Insight flagged this and nothing matches a known pattern, so the agent investigates live. It pulls session data, checks the device's experience score, reviews relevant Assist Session Summaries, and checks the DEX Knowledge Base and the open internet to see whether anyone else has hit this and documented a workaround. It shows its work along the way instead of returning a verdict with no explanation. Once it lands on a root cause, the Playbook generation service builds a Playbook or script, the admin reviews it, and the fix goes into the library for whoever runs into this next.

All three work the same way underneath. Insights v3 and Experience Scores detect, Guided RCA v2 diagnoses, and the DEX Agent decides whether to run a known fix, build a new one, or investigate from scratch, with Assist Session Summaries and outside knowledge feeding it along the way.

Situation 3: Starting from scratch

An admin asks the DEX Agent directly: "Why is this device slow?" Or a spike in ITSM tickets for the same issue kicks the question off instead. Either way, no Insight flagged this and nothing matches a known pattern, so the agent investigates live. It pulls session data, checks the device's experience score, reviews relevant Assist Session Summaries, and checks the DEX Knowledge Base and the open internet to see whether anyone else has hit this and documented a workaround. It shows its work along the way instead of returning a verdict with no explanation. Once it lands on a root cause, the Playbook generation service builds a Playbook or script, the admin reviews it, and the fix goes into the library for whoever runs into this next.

All three work the same way underneath. Insights v3 and Experience Scores detect, Guided RCA v2 diagnoses, and the DEX Agent decides whether to run a known fix, build a new one, or investigate from scratch, with Assist Session Summaries and outside knowledge feeding it along the way.

Where the device lives doesn't change any of it. The fix might run as a UEM script on a Windows or Mac desktop, a Horizon API call on a virtual desktop, or a call into Intelligence for a mobile device, but the agent and the flow stay the same.

How much the agent handles is up to you. Insights v3 and Guided RCA v2 work well on their own, so an admin who'd rather investigate and fix things by hand still gets the detection and the root cause. The agent picks up where you want it to.

Knowing where to look

Omni finds the story

Omni already answers questions and pulls data across the Omnissa consoles, and it now works on the DEX dashboards too. They hold a huge amount of data by design, spread across groups, filters, time ranges, and charts. Omni summarizes all of it down to what's worth focusing on, so an admin isn't left working out which numbers matter.

Device Timeline explains itself

It already puts every data point about a device next to a breadcrumb trail of what the user was doing, and the summary tells you what the two add up to: real memory pressure on the device, or a user with a heavy browser open and nothing wrong at all. Check it out:

 Experience Scores mean something again

Instead of one fleet-wide bar, a model scores each device against its peer group, which admins define themselves. A laptop that looks fine next to the whole fleet can be clearly struggling next to devices just like it, so a low score means something is genuinely wrong rather than a device slipping under a fixed threshold. The calibration runs continuously against your own fleet, with nothing to configure and no services engagement to buy.

The score explains itself too, showing what's pulling it down against the peer baseline, what low scorers have in common, and which devices have been struggling for weeks rather than just having one bad day. Real warranty age from HP, Lenovo, Dell, and Apple Business Manager feeds in too, which makes refresh decisions easier to defend.

Beyond the console

None of this has to stay inside Omnissa. With Intelligence MCP Server support, other agents like ServiceNow, MoveWorks, or Salesforce Agentforce can call directly into DEX to pull device data or kick off a root cause investigation. Omnissa has already shown this working with Agentforce: a device crash opens a Salesforce case, Agentforce pulls the device details and root cause straight from DEX, and the case closes itself once the fix runs.

Not just for admins

Assist Session Summaries turn a support session into something a person can read in seconds instead of piecing together from raw logs. The same summary serves two audiences: a helpdesk admin uses it to hand off or escalate with confidence, and an employee can follow it as self-service instructions if the same issue comes back – no ticket required. Check it out:

 DEX AI also shows up in the apps employees use every day. In Boxer, AI email summarization is already helping people get through a full inbox faster, and in Content, AI summaries do the same for long documents.

What it adds up to

Here's the payoff: fewer tickets, problems caught before employees notice them, and faster resolution when something does break. For frontline teams whose devices tie directly to revenue, that means uptime that keeps climbing toward 100%.

See it for yourself

Boxer and Content AI features are in your hands today. Be on the lookout for the BETA coming soon, here. Contact your Omnissa representative or request a conversation with our team to see the roadmap or get early access.