Prompts & collections

What a prompt is

A prompt is a real buyer question — the kind a shopper would actually type into an AI assistant when they're trying to decide what to buy. Something like best waterproof hiking boots for winter or what's a good gift for a coffee lover under $50.

superAEO runs each prompt through five AI engines — ChatGPT, Gemini, Perplexity, Claude, and DeepSeek — and records whether your store gets mentioned in the answer. Across many prompts and repeated runs, that's how you build a picture of your brand's visibility in AI search: where you show up, where competitors take your place, and which questions you're invisible for.

Because prompts are how the app "sees" you, the wording matters. The closer a prompt is to how a customer really phrases a question, the more accurately your results reflect what buyers experience.

Funnel stages

Every prompt is tagged with a funnel stage, so you can see where in the buying journey you're visible — not just whether you're visible at all. There are exactly three stages:

• Awareness — the shopper is exploring a problem or category and doesn't have a specific product in mind yet. For example, how do I keep my hands warm while cycling in winter. Showing up here puts your brand in front of people before they've chosen who to buy from.
• Consideration — the shopper is comparing options and narrowing down. For example, best winter cycling gloves for cold hands. This is where buyers weigh brands against each other, so visibility here is high-value.
• Purchase — the shopper is close to deciding and looking for the final nudge — price, availability, or a direct recommendation. For example, where to buy heated cycling gloves. Being mentioned here is the closest AI visibility gets to a sale.

Tagging stages lets you filter your results and spot gaps — for instance, strong Awareness coverage but nothing at the Purchase stage, where it counts most.

Adding & editing prompts

You manage everything from the Prompts page. To add a new question:

1. Click "Add Prompt" at the top of the Prompts page to open the editor.
2. Enter the question exactly as a shopper would ask it.
3. Pick a funnel stage — Awareness, Consideration, or Purchase — so the result is filed under the right part of the journey.
4. Optionally assign a collection to roll the prompt's results up under one of your Shopify collections. Leave it unassigned to track at the brand level.
5. Save. The prompt joins the next tracking run and starts collecting results across all five engines.

You can edit a prompt's wording, stage, or collection at any time. You can also pause a prompt — pausing stops it from being tracked on future runs without deleting its history, so you can resume it later and keep the past data. Resuming puts it back into the rotation. Deleting a prompt removes it and its history for good.

How many you can track

The number of prompts you can track at once depends on your plan:

• Pro — up to 10 prompts.
• Max — up to 20 prompts.

A counter on the Prompts page shows how many of your slots are in use, so you always know how much room you have left. The starter prompts that superAEO auto-generates from your catalog during setup count toward this limit — so if you want to add your own, you may need to delete or pause a starter prompt first. Paused prompts still occupy a slot; delete one to free it up entirely.

Organizing by collection

superAEO syncs your Shopify collections automatically — they're read-only here, mirrored straight from your store, so you never have to recreate them. Assign a prompt to a collection and its results roll up under that collection. Prompts you leave unassigned are tracked at the brand level instead.

The Collections page lists every collection alongside its product count, its active prompt count, and its AI visibility. That makes it easy to spot collections with no coverage at all — a collection with dozens of products but zero active prompts is a blind spot, because you have no idea whether AI engines recommend it. Adding a prompt or two to those collections closes the gap.

Reading prompt results

Open any prompt to see how it performed, broken down per engine — ChatGPT, Gemini, Perplexity, Claude, and DeepSeek. For each one you'll see:

• Mention — whether your brand appeared in that engine's answer at all.
• Sentiment — when you were mentioned, how the engine framed you (positive, neutral, or negative).
• Sources cited — which pages and sites the engine drew on to build its answer, so you can see what's shaping the response.
• Competitors — which rival brands showed up instead of, or alongside, you.

Reading results per engine matters because the five engines don't agree. You might be the top recommendation on Perplexity and completely absent on Gemini. The per-prompt view is where you confirm exactly which engines see you, which sources they trust, and who's beating you — the raw material behind every gap report and improvement action.