Short answers to the questions merchants ask us most. The same list is searchable from Support inside the app. If yours is not here, the chat and support@craftshift.com are one click away.
Octolift is a Shopify app that scores how well each of your products can be found, understood and recommended by AI shopping assistants such as ChatGPT, Gemini and Perplexity, and then shows you exactly what to change.
It runs inside your Shopify admin and reads your products through Shopify. Nothing is written to your store until you approve a specific change.
Read more: Overview
Answer engine optimization (AEO) and generative engine optimization (GEO) are the practice of preparing product listings so AI assistants and answer engines can quote and recommend them, in the way search engine optimization prepares pages for Google. Octolift measures that readiness per product and helps you improve it.
Read more: Overview
No. Scoring, the AI catalog checks, the ranking checks and the agentic sales figures are all read only. A product is only updated when you press Save, Apply category or Change category, or Apply all suggestions and confirm, and only the fields that change is about are written. Title and description changes can be undone from View changes on the Products page.
Read more: Improve with AI
No. Scores, suggestions, visibility checks and agentic sales all work through Shopify's API. Octolift does not require an app embed or any change to your theme.
No. Octolift does not ask ChatGPT, Claude, Gemini or Perplexity directly and does not report a position inside them. A single API call answers differently from the consumer product, changes between runs and varies by country, so such a number would look precise while meaning nothing.
Instead, Octolift queries Shopify's global AI catalog, the catalog AI agents read through Shopify's agentic commerce tools, and reports whether your product comes back for a shopper question and which product wins.
Read more: AI visibility
Yes. Suggestions are written in the language of your existing listing, and the shopper questions used for ranking are written in the main language of the country being checked. The app interface itself is in English.
Press Analyze product on the Home page and choose a product. You get a score, a breakdown of what to fix, how AI assistants see the product and where it ranks for shopper questions.
Then press Improve with AI on the analysis page to generate suggestions, review them, and save the ones you want. The Free plan is enough for a first round; change plan any time from Plan and model.
Read more: Getting started
Start with the products you most want AI assistants to recommend: bestsellers and the ones shoppers ask about. The Products page opens on Needs work, the analyzed products that scored weak or critical and have nothing applied yet, weakest first. Not analyzed shows what is left, and Improved shows what you have already changed.
Usually a few seconds to about half a minute. The app fetches the product, runs the checks, has the AI rater score the text and asks the AI catalog whether the product is discoverable, all at the same time. Suggestions take a little longer because each draft is scored and polished before you see it.
Yes. Anyone you give access to the app in your Shopify admin can use every feature, including applying changes. The allowance belongs to the store, not to a person.
A score from 0 to 100 built from two halves. Listing quality covers what you write and structure: the description, title, category, category fields and variant attributes. Agent readiness covers what Shopify's AI catalog returns for the product: whether it is found at all, the core fields, variant prices, policy links, availability, media and reviews.
Each check is worth a fixed number of points. The score breakdown shows the points you earned on each one and what would earn the rest.
Read more: How scoring works
They are rating bands on the 0 to 100 score. Strong means the listing is well optimized, Good means solid visibility, Weak means it is missing key signals, and Critical means it needs attention. The thresholds are the same for every store and are listed on the How scoring works page.
Read more: How scoring works
One fixed model rates every store, and it is not affected by the suggestion model you pick on the Plan and model page. That is deliberate: scores are only comparable over time and between products if one rater produced them all. Measured on the same listings, different models disagree with each other by a wide margin.
Read more: How scoring works
The agent readiness half is live: each time you open the product, Shopify's AI catalog is asked again, so a product that just synced, went out of stock or gained reviews moves. The AI-judged text checks are cached and only re-run when the listing text changes.
If the catalog could not be reached during a scan, the scan is marked incomplete and the score is computed from the remaining checks only, which can also shift the number.
Read more: How scoring works
Shopify's AI catalog did not return your product when Octolift searched for it in your store. Common reasons: the online store is password protected, the product is new or was just changed and has not synced yet, the product is not active, or it is not available in the checked market.
Confirm the storefront is public and the product is active, give Shopify some time to sync, then analyze again.
Read more: AI visibility
The app could not reach the AI catalog while scanning, so the agent readiness checks were left out and the score was computed from the checks that did complete. This is a temporary problem on our side or Shopify's; press Try again or open the product again in a moment.
Read more: How scoring works
The check looks for the four policy pages Shopify exposes to AI agents: refund policy, shipping policy, privacy policy and terms of service. Points are proportional to how many of the four are published. Add the missing ones under Settings, then Policies, in your Shopify admin.
Read more: How scoring works
At least three images, each with alt text, earns full points. Fewer images, or images without alt text, earn half. A product with no images earns nothing. Alt text is edited on the product's media in the Shopify admin.
Read more: How scoring works
The check counts the reviews Shopify's AI catalog reports for the product, which are the reviews Shopify itself holds for it. Reviews that live only inside a third-party app and never reach Shopify are invisible to the catalog, and so to AI agents.
Read more: How scoring works
Whether your variant option values are specific enough for an assistant to match a shopper's request. Navy and 42 are specific; Default Title, Option 1 or Style A are not. All values specific earns full points, at least half of them earns half, fewer earns nothing.
Read more: How scoring works
A rewritten description, an alternative title (occasionally more than one), values for the category fields that are still empty and, when the product has no category or the current one looks wrong, a category. Each suggestion appears next to what you have now. You pick what to use and can edit it before saving.
Read more: Improve with AI
It is instructed not to. A suggestion keeps every factual claim already in your listing and adds no materials, measurements, certifications or other facts that are not there. Category values are only picked when the title or description clearly supports them. Read a suggestion before applying it, as you would any copy.
Read more: Improve with AI
Only the fields you changed: the title, the description and the category fields you picked, saved together as one all-or-nothing write; if one part cannot be written, the rest is put back and the product stays exactly as it was, and the message says so. A category assignment is written on its own when you press Apply category or Change category. Nothing else on the product is touched.
Some category fields are saved in pairs (Color together with Pattern, for example), and a save with only half of a pair is stopped with a note. The description is saved as plain paragraphs, so a description that had custom HTML formatting loses that formatting when you replace it.
Read more: Improve with AI
Yes, for titles and descriptions. On the Products page, press View changes on the product's row: every change is listed with its previous value, and a title or description change has a Revert button (Revert all appears when several can be undone), as long as you have not edited that field since. If you have, the undo is refused rather than overwriting your work.
Category and category field changes are shown in the history but cannot be undone from the app; change them on the product in Shopify. History is kept for a month and then deleted.
Read more: Improve with AI
Press Rescan: the Back to analysis button turns into Rescan after a save. Because the text changed, the AI rater scores it fresh and the catalog is asked again. Shopify's AI catalog can take a while to pick up a change, so the agent readiness half may follow a little later.
Read more: Improve with AI
One product improvement on the default model. Each press of Improve with AI is one generation, and regenerating for the same product uses another. A stronger model uses more of the allowance per product; the Plan and model page shows how many improvements each model leaves you.
Read more: Plans, allowance and models
The points that check is expected to earn if you apply the suggestion, based on how the draft scored against the same rater that scores your product. The estimate at the top of the page adds them up. The real score after a Rescan can differ slightly, because the catalog checks are live.
Read more: Improve with AI
The summary Shopify's AI catalog has built for your product: unique selling points, top features, tech specs and attributes. Assistants quote from this, so a Not specified there is a gap worth filling in your description or category fields.
Read more: AI visibility
Octolift writes a handful of search phrases a shopper might type for your product, from its title, product type and category, in the main language of the country being checked. Each phrase is then run as a real search against Shopify's AI catalog, and the top ten results are read to see where you rank and which product wins.
Read more: AI visibility
Ranking mixes many signals: how complete the listing is, popularity, buyer engagement and brand recognition. Listing completeness is the part you can improve directly, so start with the score breakdown and Improve with AI. The rest builds over time.
Read more: AI visibility
The catalog answers per buyer country and shipping destination, and shoppers phrase things differently per market. Octolift checks one country at a time, defaulting to your home market. If you ship to several countries, a country dropdown appears at the top of the panel: pick a market and press Check to rank it there.
Read more: AI visibility
Press Track on any question to keep it, with its rank at the time, on the Tracked tab for that product and country. You can track up to 25 questions per country and untrack any by pressing Tracking. A tracked row keeps the rank it had when you tracked it; to refresh one, untrack it, run it again under Try your own search, and press Track on the fresh row.
Read more: AI visibility
Yes. Type it under Try your own search and press Check rank. It runs the same live catalog search and shows your rank, the winning product and its brand. You can track the result like any other question.
Read more: AI visibility
Orders that an AI assistant either placed for the shopper or referred to your store, as Shopify's own sales analytics attribute them. Octolift reads the totals from Shopify and shows the last 30 days on the Home page, and 7, 30 or 90 days per product on the analysis page.
Read more: Agentic sales
The channels Shopify's sales data credits: ChatGPT, Google AI Mode and Gemini, Microsoft Copilot, and Shop. The By platform list on the product card breaks the total down by channel.
Read more: Agentic sales
They use Shopify's last non-direct click attribution and are an estimate. AI assistants often open your store without a referrer, so some AI-driven orders show up as direct traffic and the real total is likely higher. The figures refresh about once an hour.
Read more: Agentic sales
Only product improvements: each time Improve with AI generates suggestions for a product. Scoring products, the AI catalog checks, shopper question rankings and agentic sales are included and do not use the allowance, however many products you look at.
Read more: Plans, allowance and models
Suggestions pause until the allowance renews or you move to a larger plan; the app shows the renewal date. Scoring keeps working, and everything you already scored and applied stays as it is. Unused improvements do not carry over to the next period.
Read more: Plans, allowance and models
Open Plan and model and press Change plan (See plans, before you have picked one). Shopify's own plan picker opens; pick a plan and your allowance follows immediately. To stop paying, pick the Free plan or uninstall the app. Billing is handled by Shopify and appears on your Shopify invoice.
Read more: Plans, allowance and models
The default model is the cheapest and does this job well. A stronger model uses more of your allowance per product rather than costing more money, so the picker shows how many products each model leaves you this month. Switching is free, does not change any score you already have, and only affects suggestions.
Read more: Plans, allowance and models
There is a Free plan you can stay on for as long as you like. If the paid plan you pick comes with a trial, the Plan and model page shows when the trial ends and when the first charge is taken.
Read more: Plans, allowance and models
A separate, much larger safety ceiling on the included work (scans, the category audit, the shopper questions and their rankings) that exists to stop runaway automation, not to limit you. It only appears once a store has used most of it. The text verdicts of a score are reused for 30 days while the listing has not changed, so normal use never gets near it.
Read more: Plans, allowance and models
Applying a suggestion updates the product's title, description, category or category fields through Shopify, which needs the products write permission. Reading is covered by the same permission. It is only used when you press Save or Apply.
Read more: Data and privacy
The product's title, product type, category, category field values and description, plus, when the app tests how an assistant would answer a shopper, the AI catalog's summary of the product and text excerpts of up to three of your published policy pages (refund, shipping, privacy, terms of service). Customer data, orders and your access tokens are never sent to a model provider.
The providers are OpenAI, Anthropic and Google, depending on the model your store uses. We use their paid API tiers, whose terms state that submitted content is not used to train their models.
Read more: Data and privacy
Access to your store is revoked immediately and all processing stops. Nothing you applied is reverted: the content is yours and stays exactly as it is. About 48 hours later Shopify asks us to erase your store's data and we delete your scores, change history, tracked questions and contact details.
Read more: Data and privacy
No. The only order-related data it reads is aggregate sales totals attributed to AI channels, from Shopify's own analytics. It never reads, stores or displays individual customers or orders.
Read more: Data and privacy
Press Reload app. That reloads Octolift inside the Shopify admin with a fresh session, which clears almost every transient error. If it keeps happening, open Support and start a chat, telling us the product you were on and what you pressed.
The AI rater did not return a complete verdict, and Octolift never shows a partial score as if it were final. Wait a moment and press Try again. If it persists for one product, contact support with the product name.
Read more: How scoring works
Your store has used its product improvements for the current billing period. Scoring keeps working; suggestions resume on the renewal date shown in the message, or right away if you move to a larger plan from Plan and model.
Read more: Plans, allowance and models
Octolift was updated with a permission it did not have when you installed it. Close the app, open it again from the Apps menu in your Shopify admin, accept the permission prompt Shopify shows, then save again.
Read more: Data and privacy
The save is all or nothing, so whenever one field cannot be written the message ends with Nothing was changed. It names the reason: a field that needs its partner too (Color with Pattern, for example: pick both and save again), a field set by a product option (Color when it is a variant option: edit the option in Shopify instead), a field Shopify does not accept on that category, or a field not set up on your store yet.
Fix the named field, or drop it from the picks, and save again. Everything else you picked is still in the form.
Read more: Improve with AI
Octolift limits how many AI generations and scans a store can run per minute and per day, to keep costs predictable and to stop runaway scripts. Wait a minute and try again. Normal use never reaches the daily limits.
Open Support in the app menu and press Chat with us, or email support@craftshift.com. Include your store URL and the product you were looking at. If the chat is offline we reply by email.
Open Support in the app menu to chat with us, or email support@craftshift.com. Tell us your store URL and the product you were looking at and we will pick it up from there.
Install the app, analyze your first product, apply your first improvement and check where you rank in AI search.
The two halves of the AI visibility score, every check and its points, the rating bands, and when a product is scored again.
What a plan includes, what counts against the monthly allowance, when it renews, and how the suggestion model changes it.