How replies are written

Everything about how replies sound lives in Reply Preferences, which has five tabs: Brand voice, Per-app voices, Keywords, Language and Automation. At the top you choose what drafts your replies:

Brand Voice is the guided way. You write a brief and personalize the replies with examples, tone, emoji, length and words to avoid. Most teams should use this.

Custom Prompt hands you the instructions directly. Write them yourself, or start from a template (Warm & concise, Support-first, Upsell on 5-star, Local business). A saved custom prompt replaces the brand voice drafter entirely.

Only one is active at a time, marked Running. Clearing the custom prompt, or choosing Switch to Brand Voice, goes back to the brand voice. Only the Owner and Managers can edit these settings; Developers can view them.

A draft is written for one review, using:

• the review text, its rating, and the language it was left in
• what the classifier found: sentiment, urgency, intent, and what the reviewer actually mentioned
• your brand voice or custom prompt
• the app it belongs to, including its voice, features and keywords

That last one matters more than it sounds. A reply that answers the complaint the person actually made reads like a person wrote it. A reply that thanks them for their valuable feedback reads like nobody did.

The Brand voice tab has three parts.

Brand brief. How you talk to customers, in up to 2,000 characters. Starters (Warm and friendly, Professional, Playful, Caring support, Short and direct) give you something to edit instead of a blank box. With exactly one store app connected, Suggest with AI fills the brief from your store listing; nothing changes until you save.

Business context. Your primary goal (more customers, a better rating, or saving time), up to 6 services, and a location.

Personalize replies.

• Example replies: up to 3 replies you were genuinely happy with. They teach it more than any setting.
• Tone: From my brief, Warm, Neutral, Formal or Playful. Playful is used on happy reviews only; others get Warm.
• Emoji: Never, Rarely, or If they use one.
• Reply length: Short (200 characters), Standard (350) or Detailed (900), always capped at the platform's own limit.
• Words to avoid: up to 20 phrases your team has banned. They will otherwise turn up.

The Per-app voices tab gives each app its own brief, which is what you want if you run a banking app and a puzzle game off the same account. Generate with AI drafts one from the app's listing. An app's brief replaces the account brief for that app; tone, emoji, length, words to avoid and examples still come from the account. Apps you leave alone use the account voice.

Each app can also list its features (short phrases up to 60 characters). Suggested features are added when an app connects, and you can edit them. A reply mentions at most one.

On the Keywords tab you tell the drafter which app store (ASO) keywords matter, for the whole account or per app. It works them in where they fit a sentence and leaves them out where they do not. Suggest with AI proposes keywords from your listing and reviews, and needs Starter or above. One-word generic keywords such as "good" are rejected.

Rules that keep this honest:

• One promotional mention per reply, at most: a keyword or a feature, never both.
• Never in the opening sentence.
• Never on negative or urgent reviews, two stars or below, abusive reviews, or reviews about money.

Stuffing keywords into review replies is obvious to readers, and it is the fastest way to make a reply look automated. The point is to sound like the person who built the thing.

A review in another language can be read in the original or translated. The Language tab sets the language the console is shown in.

You can read and edit a reply in your language with Show translation, then use Translate to the reviewer's language. The reply always goes out in the reviewer's language, and you see that version before it sends.

Nothing is locked. Open a draft and edit it like any other text box. Once you have touched it, nothing rewrites your words.

If a draft is close but wrong, Regenerate. If it reads stiff, Humanize rewrites it at the level you pick: Light (small touch-ups), Natural (sounds like a real person typed it) or Casual (relaxed and conversational). Undo rewrite takes you back. Regenerating or humanizing the same review does not use another AI draft from your allowance.

Edits do not quietly retrain anything. If a phrasing keeps coming out wrong, fix it in the brand voice so the next hundred drafts come out right.

AI drafts, including regenerated and humanized ones, are checked before they are offered to you, and written to fit the length limit of the platform they are going to.

The drafter is told never to invent a feature, a fix, a price or a timeline. Promises such as a release date or "we've fixed it" are filtered out. If a reviewer asks for something only a person can promise, the reply acknowledges the problem rather than committing you.

Templated corporate gratitude is filtered out too. "Thank you for your valuable feedback" is not a reply.

Your own edits are yours: what you type is sent as you wrote it, so read it before you press send.