Apple’s most interesting AI move right now isn’t a new chatbot, a flashier image generator, or even another Siri “upgrade” that mostly brings iPhone features closer to what Android users have already been using. It’s something quieter—and potentially more consequential—because it changes how people build with their devices.
In the first developer beta of iPadOS 26, Apple is experimenting with an AI-powered way to create automation inside Shortcuts. If you’ve used Shortcuts before, you know the basic promise: you can string together actions—open an app, pull data, send a message, adjust settings—into workflows that run on demand or automatically. But you also know the friction. Even when Shortcuts is approachable, it still asks you to think in terms of steps, triggers, and parameters. You’re essentially doing “workflow programming,” even if you never see the word code.
What Apple appears to be testing is a different interaction model: instead of asking you to specify every detail up front, the system helps you shape the outcome through guided conversation. The Verge describes the experience as looking a lot like vibe coding—an approach popularized by modern AI tools where you don’t write instructions like a traditional programmer; you steer the result by describing what you want, iterating as the system proposes steps, and refining until it behaves the way you expect.
That framing matters. Because if this works the way it seems to be designed, it could shift AI from being something you “use” to being something that quietly mediates between your intent and the device’s capabilities.
To understand why this is a big deal, it helps to look at what Apple has been doing with AI so far. Much of the company’s current AI pitch follows a familiar pattern across the industry: chat-based assistance for questions, quick generation and summarization of text, and image generation tools that can produce impressive results while also raising the usual concerns about accuracy, taste, and uncanny outputs. At WWDC, Apple leaned heavily into bringing Siri-style experiences closer to what people already see elsewhere—especially in apps built around large language models.
But there’s a difference between “AI that answers” and “AI that builds.”
Answering is useful, but it’s often a dead end. You ask a question, you get a response, and then you still have to do the work: copy text, reformat it, open the right app, set the right options, and execute the next step yourself. Summarization and rewriting help, but they don’t fundamentally change the interface between your intention and your device.
Building is where the leverage lives. If AI can help you assemble workflows—especially complex ones—then it becomes a kind of translator between human goals and machine actions. And Shortcuts is already the place where Apple has the infrastructure for that translation. The question is whether Apple can make that translation feel natural enough that more people actually use it.
The iPadOS 26 experiment suggests Apple is trying to do exactly that.
Inside Shortcuts, the AI layer appears to guide you toward an automation plan rather than simply generating a block of instructions. Instead of you writing a sequence of actions manually, the system helps you define what you want the workflow to accomplish, then turns that into the underlying steps Shortcuts needs. In other words, it’s not just “AI that writes.” It’s AI that interprets intent and maps it onto existing automation primitives.
This is where the vibe coding comparison lands. Vibe coding isn’t about replacing programming entirely; it’s about reducing the cognitive load of specifying everything precisely. You still need to communicate what you want, but the system handles the translation into executable structure. In the context of Shortcuts, that means fewer moments where you stop and ask, “Wait—how do I express this in Shortcuts terms?” and more moments where you say, “I want it to do X,” and the system figures out the rest.
If Apple pulls this off, it could make Shortcuts feel less like a power-user tool and more like a general-purpose “automation assistant” that lives where people already are: in the apps and flows they use daily.
Why this could matter more than another Siri feature
Apple’s AI strategy has often been framed around Siri: improving the assistant, making it smarter, and expanding what it can do. That’s logical—Siri is the brand’s interface to the future. But assistants face a structural limitation: they’re typically conversational, and conversations are inherently linear. Even when Siri can take actions, the user still has to initiate the right request, and the assistant still has to interpret it correctly in the moment.
Shortcuts, by contrast, is already built for repeatability. Once you create a workflow, it can run again and again. That makes it ideal for turning one-off AI suggestions into durable automation. An AI-guided Shortcuts experience could therefore convert “help me once” into “help me continuously.”
There’s also a subtle but important shift in what the user is doing. With a chatbot, you’re interacting with language. With Shortcuts, you’re interacting with behavior. When AI helps you build behavior, it becomes part of the system’s operational layer—not just its informational layer.
That’s why this direction feels different from the rest of Apple’s AI announcements. It’s not merely adding intelligence to existing interfaces. It’s changing the interface for creating intelligence-driven behavior.
The “guided automation” angle: less instruction, more outcome
One of the most compelling aspects of the Verge’s description is the emphasis on shaping outcomes. Traditional automation building often requires you to think like the system: define triggers, choose actions, set conditions, handle edge cases. Even if you’re not writing code, you’re still doing structured specification.
A vibe-coding-like approach flips that. Instead of you specifying the structure first, you specify the goal. Then the system proposes a structure. You review it, correct it, and iterate. That iteration loop is the heart of why these systems feel powerful: they reduce the penalty for being imprecise early on.
In practice, this could mean that Apple’s AI-assisted Shortcuts creation supports a workflow like:
“I want a daily summary of my calendar that I can send to my team.”
You might not know which exact actions to pick, how to format the output, or how to handle time zones and recurring events. But you can describe the outcome. The AI can then assemble the likely steps: fetch calendar events, filter them, format them into a message, and schedule the delivery. If something is wrong—maybe you want only meetings with certain tags, or you want it sent at a specific time—the AI can adjust the workflow rather than forcing you to rebuild it from scratch.
Even if the underlying mechanics are the same as today’s Shortcuts, the experience could feel radically easier because the AI handles the translation from intent to structure.
And because Shortcuts is already integrated into iPadOS, the resulting workflows could tap into real device capabilities: files, notifications, messaging, web content, device settings, and more. That integration is a major advantage over AI tools that live in a separate app or browser tab. The AI doesn’t just generate text; it configures actions that actually run on your hardware.
The iPad-specific opportunity: automation as a first-class interface
It’s also worth noting why iPadOS is the stage for this experiment. iPad is Apple’s most “workflow-friendly” device. The combination of touch input, multitasking, and the way iPad apps are designed for spatial interaction makes it a natural home for automation that feels tangible.
On iPhone, automation is often something you trigger quickly and then forget. On iPad, automation can become part of how you work: you might build a workflow for research, for organizing notes, for turning a messy set of tabs into a structured reading list, or for transforming a set of documents into a draft you can edit.
If Apple’s AI-guided Shortcuts creation is especially smooth on iPad, it could reinforce the idea that iPad is not just a bigger iPhone—it’s a different computing experience. Automation is one of the ways that difference can show up.
There’s also a broader interface trend here. Over the last few years, we’ve seen AI shift from “assistant you talk to” toward “assistant that acts.” The next step is “assistant that helps you configure how it acts.” That configuration layer is where Shortcuts fits perfectly.
Instead of asking Siri to do everything in one go, you could use AI to build a reusable workflow that Siri (or the system) can run later. That’s a more scalable model for everyday life.
The catch: accuracy, control, and the trust problem
Of course, there’s a reason vibe coding is exciting and also slightly unsettling. When AI translates intent into structure, it can misunderstand. It can choose the wrong action, omit a condition, or format output in a way that looks plausible but isn’t correct.
For automation, mistakes aren’t just annoying—they can be costly. A workflow that sends the wrong message, pulls the wrong data, or triggers at the wrong time can create real damage. So any AI-assisted Shortcuts experience has to solve two problems at once:
1) Make it easy to build workflows.
2) Make it safe to rely on them.
That means the UI needs to expose what the AI is doing in a way that users can verify. Even if the AI generates the workflow, the user should be able to inspect the steps, adjust parameters, and understand the logic. The best version of this feature won’t hide complexity; it will compress complexity into a guided process while still leaving a clear trail of what will happen.
Apple has historically been good at designing interfaces that feel calm and controlled. If this Shortcuts AI feature leans into that strength—showing the proposed workflow clearly, offering sensible defaults, and making it easy to correct errors—it could earn trust faster than more chaotic “type-to-generate” approaches.
But if it feels too magical, users may hesitate to deploy workflows that run automatically. The difference between “cool