How to make Siri write like you

Updated June 11, 2026 · ~4 min read · Ilura Technology

Siri can set a timer, send a text, and answer a trivia question. What it cannot do on its own is write the way you write. Ask it for an email and you get something polite, generic, and slightly off. The problem is not your request. The problem is that the assistant has no memory of your style.

That memory can be built. This guide walks through the four steps, using ILURA as the working example: pick a role playbook, apply it to a real message, correct what sounds wrong, then call the result by name wherever you write.

What does “write like you” actually mean?

Your style is not one setting. It is a stack of small preferences. How you open a message. How direct you are with bad news. Whether you sign with your first name. Which words you would never use.

A generic assistant guesses at all of this, and every new chat starts the guessing over. You repeat the same instructions, get a slightly different voice each time, and lose everything when the conversation ends. A playbook replaces that guessing with memory: a saved set of writing instructions for one role you play, improved through correction instead of re-prompting. The steps below build one from scratch.

Step 1: how do you pick a role playbook?

Start from a role, not a blank page. A role playbook is a starting profile for one job: following up with clients, briefing your manager, replying to customers, updating investors. Pick the single role you write in most often. One sharp playbook beats five vague ones.

The role matters because it carries context a one-off prompt never does. Who you are. Who you are talking to. What a good outcome looks like. A follow-up to a client and a status note to your boss need different instincts, so they belong in different playbooks.

Step 2: how do you apply it to a real message?

Do not test with toy text. Take something you actually need to send today, such as a follow-up on a client thread that went quiet. Give the playbook the context and let it draft.

The first draft will be close, but it will not be you. That is expected. The gap between the draft and what you would actually send is the raw material for the next step. Read the draft the way an editor would, and notice exactly what you would change before sending.

Step 3: how do you correct it and save the preference?

Edit the draft directly. Cut the greeting you never use. Shorten the second paragraph. Change the sign-off. Then save the correction.

The app records each edit as a learned preference — the “learned this preference” moment — such as no hedging openers, short paragraphs, or a first-name sign-off. Preferences are versioned, so you can review what has been learned and roll back anything that does not fit. This is training by correction. There is no settings screen to configure and no prompt to engineer. A handful of corrections across real messages is usually enough for drafts to start sounding familiar.

Step 4: how do you call it by name?

A trained playbook is only useful if it is reachable where you write. Because the actions behind it are built on App Intents, the playbook shows up in three places.

In the app, you pick the playbook and write. From the Share Sheet, you select text in Mail, Safari, or Notes, tap Share, choose the tool, and the playbook applies on the spot. Through Shortcuts, the playbook becomes an action you can chain into one-tap flows or run by speaking its name to Siri where supported.

Everything runs on the device with Apple Intelligence. There is no account to create, and the App Store privacy label is Data Not Collected. The app and the Share Sheet flow work today on iOS 18 and later.

Why does App Intents matter for Siri?

App Intents is the framework that lets apps describe their actions to iOS. That is what makes a playbook usable from Shortcuts, the Share Sheet, Spotlight and Siri instead of being trapped inside one app screen.

For writing, the important habit is name-based reuse: ask for a draft in your Sales playbook, run a manager brief action, or clean up a selected paragraph. Future Siri improvements may make those actions feel more conversational, but the useful behavior starts with the playbook you train today.

Where should you start?

One role, one real message, one correction. Pick the playbook for the writing you do most, run it on something you need to send today, and fix what sounds off. The style you save is the style ILURA can reuse whenever that playbook is invoked.