iOS App Intents vs Shortcuts: When an Update Silently Kills the Automation

An iOS update can kill an automation without a funeral. The Shortcut still sits in the folder. The personal automation still shows a toggle. The App Intent the Shortcut called has been renamed, permissioned, or quietly retired by a vendor who shipped a new Siri story. You discover it when the morning routine does not fire, or worse, when it fires and does nothing. For the automations that are still worth keeping after an update, see Shortcuts that actually save time in 2026. Silent is the house style.

App Intents are how modern iOS apps expose verbs to Shortcuts, Siri, and Spotlight. Shortcuts actions are the older catalog plus those intents wearing a block. Users experience one canvas. The canvas is not one API. Updates land on different layers. That is why “Shortcuts broke” is often “an intent changed and Shortcuts still has a block with yesterday’s name.”

Two layers, one yellow block

A first-party Shortcuts action — set Focus, get calendar events, send a message — lives in Apple’s house. It can still break. It breaks with company. An App Intent lives in the developer’s house. They can ship a better intent and orphan yours. They can add a parameter that is now required. They can move the verb to an App Shortcut you must re-add. Your Shortcut does not always show a red error. It shows a successful run with an empty output. Empty is how you miss a log.

I name Shortcuts after the verb and the app version I built them against, in a comment at the top. Comments are not version control. They are a clue for future me. After an OS update I search for the app’s name in Shortcuts and I run each one by hand. That is the funeral I schedule so the automation does not die alone.

What silently dies

Confirmation dialogs return. Apple decides a personal automation is now sensitive. You were not looking at the phone. The automation is a ghost with a conscience.

Calendar and Reminders intents grow new privacy prompts. The first run after update is a tap you will not be present for. I run the first run on purpose at the kitchen counter.

Third-party “add row,” “log workout,” “set thermostat” intents vanish from the block picker or stay as zombies. I have a zombie that still looks valid. It no-ops. I replaced it with a webhook I own. Webhooks also break. They break loudly if I check Healthchecks. Intents break politely.

Focus filters and modes rename. A Shortcut that sets “Work” sets nothing if Work became “Focus 2” in your head and “Travail” in the system after a language experiment. I do not experiment with language packs the same week as an OS drop.

App Intents are still the right long-term bet

When they work, they are better than UI automation you do not have, and better than a share sheet you have to remember. Developers who invest in intents are saying the verb is a product. I prefer those apps even if I pay the update tax. The tax is smaller than Tasker’s coordinate hell, which I use on Android and respect and fear.

Shortcuts-only actions that wrap a private URL scheme are the old world. They die too. I do not romanticize them. I just know their death is sometimes more obvious — the URL fails. An intent death can look like success.

The morning-after test

After any iOS or iPadOS update, before I leave the house: run the five automations I would notice by noon if they died. Car-ish, morning log, a file save, a webhook ping, a Focus set. Watch the screen. If a prompt appears, that automation is no longer unattended. Decide to live with the prompt or move the job off the phone.

I do the same after a major app update for the two apps my Shortcuts hug. App updates are OS updates in a trench coat.

I keep a paper list of the five. The list is not in Shortcuts. Shortcuts might be the thing that is lying. I have also kept the list in a pinned note on the Mac. Either works. The medium must not depend on the automation that might be dead. Paper is rude and correct.

When to stop using an intent

If the vendor has changed the intent twice in a year, I stop hugging it. I go HTTP if they have an API. I go manual if they do not. A Shortcut that I rebuild every spring is not an automation. It is a seasonal craft.

If the job is safety or money, it does not live in a third-party intent. I have said this about Pushcut and I will say it here. Intents are for convenience. Convenience can die silently. Safety needs a pager and a path that is not a yellow block.

Scriptable and webhooks as shock absorbers

I still use Shortcuts as the trigger. I move logic to Scriptable or to a server when the logic must survive a block rename. The Shortcut becomes “run this script” or “POST this.” The script I can read in git. The intent I cannot. After an update I fix one wrapper instead of twelve blocks.

That is not purity. It is fewer funerals. I still keep a couple of one-tap Shortcuts that are only first-party blocks — log a glass of water, toggle Low Power. Those I will rebuild in five minutes if Apple sneezes. I will not put a vendor’s “create invoice” intent in the same basket. Invoice is a wrapper or it is a browser bookmark I press like an adult. The five-minute rebuild budget is the classification system. Over five minutes belongs off the canvas or in a script I can diff.

Siri as a second casualty

App Intents power Siri phrases you trained your mouth on. The phrase can remain while the intent behind it no-ops. I re-test the phrase in the kitchen test. Talking to the phone after an update feels stupid and it catches the stupid failure. I do it anyway.

Family Sharing, work phones, and the second iPhone

A Shortcut you built on your phone does not always survive a restore onto a new iPhone with the same Apple ID. Some intents re-bind. Some ask for permission again and then sit. I treat a phone migration like an OS update: the five, by hand, on the kitchen counter, before I retire the old slab. iCloud sync of Shortcuts is a rumor with a folder. I have seen duplicates. I have seen an old block win. I delete the duplicate I do not run. I do not keep two “Morning” Shortcuts and hope.

Work-managed iPhones will forbid the interesting intents. MDM is a silent killer that is not silent if you read the profile. I do not fight it with a personal Shortcut I will have to explain to IT. Personal automations live on a personal phone. That sentence has saved me a policy meeting.

Kids’ phones and Shared iPads are worse. An intent that needs an Apple ID you do not want them to have will fail closed. Good. Do not build a household on a child’s Shortcut. Build it on a wall switch or a Home hub that does not require their Face ID at 7 a.m.

WatchOS updates break the same intents on a smaller screen. The Watch might be the only runner of a complication-driven Shortcut. I include the Watch in the morning-after test when the automation is a wrist tap. Leaving the Watch on the charger overnight and testing at 7 is how I find the death before the commute.

Beta iOS is how you volunteer for extra funerals. I do not run betas on the phone that runs the five. I run betas on a spare if I have one. If I do not have a spare, I am not a beta person. I am a person who wants the car-ish Shortcut to work on Monday.

The choice I would make again

I would still build on App Intents for apps that treat verbs as a product. I would still wrap them. I would still run the five the morning after. I would not stack twenty vendor intents and call it a system. I would not update iOS at 11 p.m. on a Sunday before a Monday that depends on a Shortcut. That Monday is how silent death gets a story. Give the death a morning and a list. Then the update is just an update, and the automation is something you still own — or something you finally had the decency to delete. Deleting is underrated. A dead block you keep is a lie you will trust next month. I delete the zombie even when it makes the folder look small. Small and true beats a museum of yellow that does nothing when the season changes.