The complaint that Siri on Mac 'turns everything into a web search' points to something specific: Siri on macOS has almost no hooks into the apps people actually work in. On iOS, Shortcuts gave developers a way to expose app actions to Siri. On macOS, that integration is shallow, inconsistently implemented, and requires every developer to opt in — which most don't. So when a user asks Siri to do something reasonable, Siri has no action to fire and falls back to a search.

Apple's incentive to fix this is weak because Shortcuts on macOS is genuinely hard to use for non-developers, and Apple is focused on building out Apple Intelligence features that require newer hardware. The users who want more from Siri on Mac right now — the ones saying it feels 'forced,' 'limited,' and 'not nearly as powerful or customizable as expected' — are left with a gap between what Shortcuts can technically do and what they can actually configure without writing AppleScript.

What's missing is a visual builder where a non-technical Mac user describes in plain language what they want Siri to do — 'when I say open my standup template, create a new document from this file and put today's date in the title' — and the builder translates that into a working Shortcut with Siri trigger, without the user ever seeing the Shortcuts interface. The builder handles the boilerplate, the naming conventions for Siri invocation, and the testing. The user just describes their workflow and approves the result.

This recurs every time someone changes their workflow, gets a new job, or updates macOS and finds their old Shortcuts broken. The need isn't one-time setup — it's ongoing maintenance of a personal automation layer that nobody currently makes easy.

What to build

Build a Mac app where users describe automations in plain English, the app generates and installs working macOS Shortcuts with valid Siri triggers and tested app actions, stores a personal library of these automations, and automatically flags broken ones after macOS updates so users can repair them with one click.

Where to start

Start with a library of 30-40 pre-built, tested automations for the ten most common Mac productivity apps (Mail, Calendar, Notes, Finder, Safari), let users install and trigger them with one click, and use the customization requests from that cohort to validate which plain-language builder features to build first.

The hard part

The hardest problem is that macOS Shortcuts actions vary wildly by app — some apps expose rich actions, most expose almost none — so the plain-language-to-Shortcut translation will fail or produce useless output for a large percentage of what users try to automate, and managing that failure gracefully without frustrating users early in the experience is genuinely hard.

How it makes money

Free to install up to 5 automations from the pre-built library, then a flat annual subscription for unlimited custom automations, the plain-language builder, and update monitoring after macOS upgrades.

See the evidence. The complaints behind this idea, the products they came from, and similar ideas in Operating System.

More ideas in Operating System