Generally Available — and actively evolving. Applivery Intelligence is generally available and supported. It is also an AI product under continuous research and improvement: answers are generated, capabilities change between releases, and mistakes are possible. Review what it produces before acting on it.
Access and seats
How do I get access?
Your Applivery organization Owner buys a seat and assigns it to you. Then open Intelligence from the ✨ button in the Applivery dashboard top bar. There is nothing to install or configure. See Getting started.
Do I need to enter an API key or connect my Applivery account?
No. Intelligence runs embedded in the Applivery dashboard and reuses your active session. If you can see the dashboard, you are already authenticated.
Who can buy seats?
Only the Applivery organization Owner. The role is read live from the Applivery directory, so it always matches who is Owner right now.
Why is Insights, Workflows or Cowork missing from my navigation?
Those three require a Max or Ultra seat. Chat is available on every tier.
I have several Workspaces. Do I need a seat in each?
Yes — a seat is granted in one Workspace. Conversations, dashboards, workflows and monitors are scoped to a Workspace too, so switching Workspace switches all of them.
What happens to my work if my seat is unassigned?
Nothing is deleted. Dashboards, workflows and monitors belong to the Workspace, and keep running.
What it can and cannot do
Where does it get its data?
From your live Applivery Workspace, through the same API the dashboard uses. There is no separate copy of your fleet and no cached snapshot it answers from.
Can it change things, or only read?
Both. It can lock, rename, restart, wipe, move, disenroll, install and uninstall; create and update Policies; author and assign scripts; manage Segments and Audiences. Every one of those asks you first.
Can it do things I cannot do?
No. It acts as you, with your Applivery role. If the dashboard would refuse you, so will Intelligence.
Will it answer questions that are not about my fleet?
No. Requests unrelated to device management are declined. It is scoped deliberately.
Can it answer questions about the Applivery product itself?
Yes. The Knowledge Agent answers from the public Applivery documentation and REST API reference, with links to the pages it used.
Why did it say it created a CSV, a report or a Policy, and nothing appeared?
It should not, and there are explicit rules against exactly that. If you see it, treat the claim as false — check the Applivery audit log or the artifact panel for what actually happened, and report it. CSV, specifically, is not a separate action: any table artifact carries its own download button.
Why did it give me a different answer to the same question?
Answers are generated, and the fleet moves underneath them. For anything you need to be reproducible, pin it as an Insights widget — a widget re-runs a fixed query rather than re-deciding how to answer.
Approvals and safety
What needs my approval before it runs?
Out of the box, everything — reads included. Every capability starts on Needs approval, so a new seat is asked about each call until you relax the ones you trust. You do that in Settings > Integrations > Applivery MDM > Tool Permissions.
Where do I change that?
The gear icon in the top bar, then Integrations > Applivery MDM > Tool Permissions. Each capability can be set to Always allow, Needs approval or Blocked. A sensible starting point is allowing the reads you use constantly — list Devices, get a Device, get Device apps — and leaving everything that changes something on Needs approval.
Is that setting mine or the Workspace's?
Yours. It applies to your own conversations only. Relaxing a capability does not stop a colleague being asked about it, and blocking one does not take it away from them. There is no Workspace-wide tool policy in Intelligence — the shared backstop is your Applivery role, which no setting here can widen.
Which actions always ask, no matter what?
Wiping a Device, locking a Device, and anything that deletes. These cannot be set to Always allow by anybody, including Applivery. They show a Locked badge in the list.
Can I see what has actually been called?
Yes — the History tab beside Tool Permissions lists the calls from your own chat sessions with their input, response, latency and time, over the last 7 days, 30 days or 3 months.
What does an approval prompt show me?
The exact capability being called, on which object, with which parameters — before anything happens.
What if I ignore an approval prompt?
It expires and the action does not run. Nothing proceeds by default.
Do Workflows have the same protection?
Workflows run under the identity of the person or schedule that started them, with the same Applivery permissions. For bulk operations, put an approval node in front of the destructive step — the run pauses and waits for a person, and if nobody answers within the timeout it takes the rejected branch.
Can Cowork change my fleet on its own?
No. Cowork proposes; a person approves. When you approve, the remediation is a fixed template shipped by Applivery, restricted to the objects in that finding, executed under your credentials. It can be reverted, and the revert only undoes what the execution actually changed.
Is there an emergency stop?
For Cowork: disable the monitors on Cowork > Monitors. A disabled check does not run and reports nothing, and disabling all of them stops Cowork in that Workspace. For chat: set the capability to Blocked in your own Tool Permissions, which stops it for you, or narrow the person's Applivery role, which stops it everywhere including the Applivery dashboard itself. Applivery additionally operates platform-level switches that can suspend Cowork across the service — ask your Applivery contact if you need that.
Will the actions show up in the Applivery audit log?
Yes, under the name of the person who ran or approved them.
Limits and usage
How is usage counted?
In credits, against your seat's allowance for the billing period. Reading Devices, reasoning about them and rendering results all consume some.
Is there an hourly or daily cap?
No. The only limit is the one for the whole billing period.
What happens when I run out?
Your plan allowance goes first, then the Workspace's purchased credits if your seat is allowed to draw on them. If neither is available, chat pauses until the reset date. Nothing is lost.
Does my allowance roll over?
No. It refills in full on the reset date.
Do scheduled workflows and Cowork monitors eat my personal allowance?
No, deliberately. Background work spends from a separate shared pool funded by Max and Ultra seats, so a job running overnight cannot leave you with no chat in the morning.
Why does my Workspace have no automation pool?
Because no seat in it is Max or Ultra. Free and Pro seats fund nothing.
Is there a limit on how many dashboards, workflows or monitors I can create?
No numeric limit. Those objects belong to the Workspace, not to a person, so capping them per seat would produce nonsense.
How long is history kept?
Workflow node-level history 30 days; alert history and dashboard refresh history 60 days; usage records 90 days. Workflow run status is kept indefinitely. Conversations, dashboards, artifacts, workflows and findings are kept until you delete them.
Guardrails
What is filtered before the assistant sees my message?
Attempts to override its instructions or extract its configuration, requests outside device management, and messages over the configured length limit. The scope and the limits are set per Workspace by Applivery, not from your dashboard — ask your Applivery contact if the defaults do not suit you.
Can somebody get it to leak another customer's data?
Every request is resolved to one Workspace, and credentials are resolved per Workspace on each call. There is no shared account and no cross-Workspace query path.
What stops it inventing a fleet when it cannot reach one?
If the Workspace exposes no Applivery capabilities — a misconfigured Workspace, or a role with no permissions — chat refuses to answer at all rather than answering from general knowledge. You get an explicit banner instead of a plausible fiction.
What stops a Cowork monitor from spamming the inbox?
Per-run caps on findings, tool calls and iterations; a daily cap on runs per Workspace; a daily cap on executed actions; and its own token budget separate from chat's.
Does an empty Cowork inbox mean everything is fine?
Only when it says so. Cowork distinguishes "ran and found nothing" from "could not run", and only shows the all-clear on evidence that the fleet was examined.
Data and privacy
Is my fleet data used to train models?
No. Your data is used to answer your request.
What is stored?
Your conversations and the artifacts they produced, your dashboards and their last refreshed data, your workflows and their run history, your Cowork monitors and findings, and operational records of usage for billing and support.
Where does it run?
On Applivery's infrastructure, over HTTPS end to end. The Applivery credentials it uses are held encrypted and resolved per Workspace at call time.
Can I delete a conversation?
Yes, from the conversation menu. Dashboards, workflows and findings are deletable too.
Who in my organization can see what?
Conversations are yours alone — they are listed against your own account and colleagues in the same Workspace do not see them. Your Tool Permissions are yours too. Dashboards, workflows, Cowork findings and the Workspace's integration credentials are Workspace-level and shared with everyone who holds a seat.
Troubleshooting
"Connect Applivery" appears instead of the chat.
The embedded session has not been handed over. Reload the Applivery dashboard and reopen Intelligence.
"Session expired".
Your Applivery session ended. Sign in to the dashboard again.
It answers, but never reads real data.
Usually a Workspace without Endpoint Manager, or an Applivery role without Device permissions. Confirm both in the Applivery dashboard.
A widget shows "Overdue — refresh may not be running".
The dashboard's schedule has not fired within its expected window. Try Refresh all; if it keeps happening, check that the Workspace still has an automation pool — that is, Max or Ultra seats.
An alert never fires.
Check the notify mode — on change stays silent while the condition remains continuously true. Also check the cooldown, and confirm the widget is actually refreshing.
A monitor produces nothing.
Open the run status. skipped and credential error are distinct from "found nothing", and the inbox tells you which one it was.
A long answer stops mid-way.
The turn ran out of output budget. Ask for a narrower slice — one platform, one Segment, one page — rather than the whole fleet at once.