Repeated
The same disk cleanup, the same service restart, the same ticket sorted the same way. Skilled people spend their day on the predictable.
Solution · By Challenge
Manual, Repetitive Operations
If your team fixes the same alert, routes the same ticket and chases the same approval every week, a platform should be doing that work. Autointelli runs approved fixes when alerts fire, routes and escalates tickets by rule, and keeps approvals moving, with guardrails and a record of everything it did.
- Detectand trigger
- Approveand guard
- Runand record
Disk usage above 90% on APP-07The same alert, handled two ways · sample data
00:00Alert fires
The problem
Toil piles up in three places
Nobody plans to spend the week on repeat work. It just never stops arriving.
Waiting
Fixes wait for someone to notice, and requests wait in an approver's inbox. The delay is longer than the work.
Remembered
Backups, patch windows and monthly requests depend on someone remembering. When they forget, nobody finds out until it matters.
Where the manual work hides
Six gaps, and how each one is closed
The same fix, by hand, every time
The manual wayA disk fills or a service stops. Someone is paged, logs in and types the commands they typed last week.
Closed withUnified360 binds an alert's trigger or recovery to an approved script, Linux bash over SSH or Windows PowerShell over WinRM, and can run it on the very server that alerted.
Scripts nobody reviewed
The manual wayA script pasted from a chat thread and run with admin rights.
Closed withScript versions move from draft to pending approval to approved, each with a SHA-256 checksum. Approval by someone other than the author is the default, and a rule only ever runs a pinned, approved version.
Automation that makes things worse
The manual wayA flapping alert re-runs the fix every minute, or a reboot fires with nobody watching.
Closed withA cooldown between runs of the same rule, a dry-run mode, a human approval for every destructive-risk run, and one bounded follow-up if a run fails.
Tickets sorted by hand
The manual waySomeone reads every new ticket, decides who should have it, and sets the priority.
Closed withITDesk automation rules test conditions on a trigger, then assign, prioritise, tag, notify, escalate, call a webhook, run a script or message the requester on WhatsApp.
Approvals chased by email
The manual wayA request waits in an inbox until someone remembers to reply.
Closed withITDesk workflows: approval steps with named approvers, any-one or quorum, reminders when a step is overdue, branches and waits. Catalogue items carry approval stages too.
Chores that live in a calendar
The manual wayConfig backups, patch windows and monthly requests depend on somebody remembering.
Closed withSchedules that run themselves: config backups by device role, discovery scans, patch windows, recurring requests, reports by email, and auto-close of resolved tickets.
How Autointelli takes over the routine
Detect. Decide. Do. Prove.
Four steps, from the alert that starts the work to the record of what ran.
Step 1 · Detect
Let the alert start the work
- An alert rule watches your servers, network and applications.
- Bind its trigger or its recovery to a script, and pass the alerting host into the run.
- In ITDesk, a trigger such as ticket created or SLA breach starts a rule the same way.
Alert rule · firesDisk usage above 90% on APP-07
TRIGGERthen RECOVERY
Automation ruleClear temp files when this alert triggers
Run on the host that alertedCooldown 300 sFailure script: Collect diagnostics
Step 2 · Decide
Only approved changes run
- Scripts are versioned and checksummed, and approved before a rule can use them.
- The approver differs from the author by default, and PowerShell is linted before review.
- Each version carries a risk level, and a destructive run waits for a human every time.
Script · Clear temp filesVersion 3 · approved
Authordev.ops
Approved bya different person (maker-checker)
Checksumsha256 9c1f…e07a
Risk: lowRisk: destructive needs an approver for every run
Step 3 · Do
Run it where it's needed
- Linux runs over SSH with a pinned host key. Windows runs over WinRM.
- A dry-run flag lets you watch a new rule before you trust it. The script decides how to honour it.
- Output is captured, trimmed and stripped of the credentials used to connect.
Run #142 · rule · APP-07Clear temp files v3 succeeded
Removed 412 MB from C:\Windows\Temp exit code 0 · 19 s
WinRM · PowerShellCredentials redacted from output
Step 4 · Prove
Keep the record
- Every run keeps a snapshot of the target, script version, risk level and who approved it.
- Ticket automations log each action's real outcome, including failures.
- Config backups keep versioned snapshots you can compare line by line.
Run history · sample
#142 Clear temp files v3succeeded
#141 Restart the service v2 · dry rundry run
#140 Reboot the server v1rejected
#139 Clear temp files v3failed
Alert-driven fixes
Wire an alert to a script, then try to break it
Set up a rule, then fire a flapping alert at it three times. Watch the cooldown, the dry run, the destructive-run approval and the failure follow-up do their jobs.
Automation rule
When this alert triggers, run this script.
Run history
What the automation console would record.
Illustrative. Scripts, risk levels, cooldowns and targets are configured in Unified360. Sample data.
Ticket automation
Change the ticket and see which rules fire
Six sample ITDesk rules. Pick a trigger and a ticket, and watch every enabled rule get checked in priority order.
The ticket
Change anything and the rules re-run.
Rules, in priority order
Every match runs all of its actions, and each outcome is logged.
The ticket afterwards
Illustrative. Rules, conditions and actions are configured in ITDesk. Sample data.
The week that runs itself
Chores that stop depending on memory
Filter by platform to see what each one takes off your calendar.
| Job | Mon | Tue | Wed | Thu | Fri | Sat | Sun |
|---|---|---|---|---|---|---|---|
| Unified360Alert-driven fixesRuns when an alert fires or recovers. No calendar needed. | |||||||
| Unified360Config backupsCaptured on an interval set by device role. Old snapshots are pruned daily. | |||||||
| IntelliAssetDiscovery scansRecurring scans on a timetable you set. | |||||||
| IntelliAssetPatch windowApproved patches roll out inside a maintenance window. | |||||||
| ITDeskRecurring requestsDaily, weekly, monthly or cron requests raised for you. | |||||||
| ITDeskReports by emailScheduled reports land in the inbox. | |||||||
| ITDeskAuto-close resolved ticketsAfter a set number of days, when you turn it on. |
Scheduled Runs when an event happens
Illustrative schedule. Intervals, windows and days are set per job. Sample data.
What changes
When the routine runs itself
Routine fixes start on their own
The alert that used to page someone starts an approved script, and the person gets the result.
Approvals that keep moving
Named approvers, quorums and overdue reminders replace the inbox that nobody watches.
Changes that are reviewed
Approved versions, a different approver by default, and a gate on destructive runs.
A record of what ran
Every run and every automation action leaves an entry, including the ones that failed.
Built with
The platforms behind this solution
Unified360
Operations and observability
Alert-driven scripts with approvals and guardrails, plus scheduled config backups with change comparison.
ITDesk
IT service management
Automation rules, approval workflows, recurring requests and auto-close on the service desk.
IntelliAsset
Asset and patch management
Scheduled discovery, patch policies with maintenance windows, and scheduled reports.
Questions
Frequently asked
What can run automatically when an alert fires?
In Unified360, an alert rule's trigger or recovery can start an approved script on a target: bash over SSH on Linux, or PowerShell over WinRM on Windows. The target can be fixed, or the server that raised the alert.
Can a script run without anyone reviewing it?
No. A rule only runs a pinned, approved script version. Versions go from draft to pending approval to approved, each with a SHA-256 checksum. By default the approver must be a different person from the author, and that can be relaxed for small teams.
What stops an automation from running wild?
A cooldown between runs of the same rule, which defaults to 300 seconds, so a flapping alert can't hammer a server. A dry-run mode, where a flag is passed to your script and the script decides how to honour it. A human approval for every run of a destructive-risk script. And one bounded follow-up script if a run fails, which cannot chain any further.
What can ITDesk automate on tickets?
Rules pick a trigger, such as ticket created, updated, resolved or closed, or an SLA warning or breach, add conditions, and run actions. Actions include assigning a group or agent, setting priority or status, tagging, notifying, escalating, setting a field, calling a webhook, running a script and messaging the requester on WhatsApp. Every enabled rule is checked in priority order, and each action's real outcome is logged, including failures.
How do approvals work?
ITDesk workflows have approval steps with named approvers, either any one of them or a quorum, and reminders when a step has waited too long. Steps can branch, wait, assign tasks and change status. Service catalogue items also carry approval stages.
Can requests repeat on a schedule?
Yes. A catalogue request can be set to repeat daily, weekly, monthly or on a custom cron schedule, so the routine request is raised for you.
Does it back up network device configurations?
Unified360's Config Backup captures device configurations on an interval set by device role, through the monitoring agent. It uses SSH drivers for many vendors and APIs for Fortigate and Palo Alto, keeps a new snapshot only when the config actually changes, lets you compare any two snapshots, and prunes old ones, after 10 days by default.
Is this a no-code workflow platform?
Unified360, ITDesk and IntelliAsset each have automation built in for their own domain. For broader, cross-system process orchestration — your own scripts, scheduled or on demand, with an API on every workflow — see IntelliFlow.
See it on your own routine
Find out how much of your week is repeat work
Tell us what your team does over and over and we'll show you how Autointelli would run it. No commitment required.