Octane AI Launches Phoenix for AI-Powered Shopify Product Quizzes

On October 21, 2026, Octane AI introduced Phoenix, a new version of its product quiz platform for Shopify stores. The release brings together AI-generated questions, product recommendation logic, automated checks of customer paths, sales analytics, and A/B testing.

Phoenix can build a quiz from a written prompt or an existing document, select products from a Shopify catalog, and configure the design through a conversation with an AI assistant. After publication, the platform helps teams identify problematic steps, investigate revenue, and test changes. Phoenix is available through the Octane AI app in the Shopify App Store, and existing customers can migrate their quizzes to it.

AI Quiz Builder Creates Product Quizzes From Prompts and Documents

Building a quiz in Phoenix starts with a natural-language description of the task. For example, a cosmetics store might request a quiz that identifies a shopper's skin type, primary concerns, and care preferences before recommending suitable products.

The AI assistant builds questions, answer choices, result pages, and recommendation rules using the Shopify product catalog. Instead of writing a prompt, a marketer can upload existing materials in PDF, Word, Excel, CSV, Markdown, JSON, HTML, or plain-text format. Phoenix examines the document and uses the available information to construct the quiz.

A brand could upload a product spreadsheet with attributes and matching rules for different shopper needs. The assistant would use those inputs to establish the initial recommendation logic. After generation, users can refine the quiz in chat or a visual editor.

Phoenix supports separate conversations for logic, design, and copy. That helps teams change one part of a quiz while retaining the context of other tasks. The assistant also lists completed changes and allows users to compare quiz versions.

AI Issues Finder Tests Thousands of Possible Customer Paths

One of the principal additions is AI Issues Finder. A product quiz can include dozens of questions, conditional branches, and combinations of recommendations. As the number of possible paths grows, checking them manually becomes increasingly difficult.

Phoenix analyzes the quiz structure and simulates answer sequences. It looks for paths that never reach a result, unreachable pages, conflicting rules, and products that cannot appear in recommendations.

It also checks the Shopify catalog connection. For example, a quiz may recommend a product that has been deleted, archived, or is out of stock. Phoenix can identify these situations and suggest corrections. Users can accept proposed changes and rerun the check.

This is particularly relevant for stores with large assortments and frequently changing inventory. However, automated validation does not establish whether every recommendation makes commercial sense. Teams still need to evaluate product fit and the accuracy of marketing claims.

Personalized Product Recommendations and Result Pages

Phoenix can select products based on customer responses. A skincare quiz, for example, may consider skin type, specific needs, preferred ingredients, and the number of steps in a routine.

Recommendation configuration includes branching rules, scoring, product ranking, and AI-powered recommendations. Result pages can use Smart Copy and Smart Properties to tailor explanations to individual answers. Instead of a generic product description, the shopper might see why a specific product matches the needs they reported.

The platform also supports image analysis. In certain quiz flows, a shopper can upload a photograph for use alongside their answers in product selection. A makeup retailer might use this approach to help suggest a shade.

Photo analysis should not be treated as medical diagnosis or a guaranteed assessment of skin characteristics. Its practical accuracy depends on lighting, image quality, and the underlying model.

AI-Generated Images and No-Code Quiz Design

Teams can adjust quiz design through written instructions, asking the AI to change colors, typography, backgrounds, or the appearance of individual screens. Changes can apply to one page or the entire quiz.

Phoenix also generates images, such as question-page backgrounds, illustrations for answer choices, and decorative elements. For existing brand assets, it can help resize images for different quiz components. Generated and uploaded assets are stored in a media library for reuse.

This allows marketers to build and style product quizzes without necessarily involving a developer. Brand images and product descriptions still need review before launch, especially when they must accurately represent the merchandise being sold.

Shopify Revenue Analytics Connects Quiz Behavior With Orders

Phoenix expands the analytics available for product quizzes. By connecting customer interactions with Shopify orders, the platform can measure not only quiz completions but also subsequent purchases.

Reports include conversion rates, revenue, and audience drop-off at individual pages. For example, a marketer may find that shoppers frequently abandon a quiz when asked about their budget. Teams can also compare answers with associated orders to examine which customer needs are more common among higher-value buyers.

Phoenix accounts for purchases made after the initial quiz session. According to Octane AI's analytics description, orders may be associated with a shopper's profile for up to 90 days after quiz completion, extending the analysis beyond purchases made immediately after recommendations are shown.

Quiz-attributed revenue is not the same as incremental revenue. Some customers would have purchased even without the quiz. Estimating additional revenue requires a controlled experiment or another appropriate incrementality method.

AI Experiments Tests Alternative Quiz Versions

Phoenix includes AI-powered Experiments. Based on quiz performance, the assistant can suggest a change, create a variant, and prepare an A/B test.

For example, if many customers drop out on an email signup screen, Phoenix may propose changing the form placement or copy. The marketer approves the experiment and selects a primary metric. Available metrics include Completion Rate, Email Opt-in Rate, Order Conversion, and Revenue per Visit.

Teams can also specify guardrail metrics. A retailer optimizing revenue per visitor, for example, might want to avoid a material decline in email signups. Phoenix takes that constraint into account when evaluating an experiment.

Traffic is split between the original and test versions, with each visitor remaining assigned to their initial variant. The minimum experiment duration is seven days. The platform includes protection against a single unusually large order determining the outcome. Experiments can end manually, on a specified date, or under configured automatic conditions; the selected winning variant can then be published.

Experiments is available on Plus and Enterprise plans.

Zero-Party Data Builds a More Complete Shopper Profile

Quiz answers are stored in the customer profile as zero-party data: information shoppers voluntarily provide to a brand. This might include product type, intended use, style preferences, or desired product characteristics.

When users can be reliably matched, Phoenix combines responses from multiple quizzes into a single customer profile. A shopper might complete one quiz during an initial purchase and another several months later. The newer answers enrich the existing profile.

These signals can inform both on-site recommendations and subsequent marketing. For example, a cosmetics brand might segment customers who expressed an interest in a particular care routine and prepare a corresponding email campaign.

Zero-party data represents stated preferences; it does not guarantee future purchases. Collection and use of contact information must also comply with applicable consent and privacy requirements.

Integrations Connect Quiz Answers to Marketing Platforms

Phoenix retains Octane AI integrations with major Shopify marketing tools. Klaviyo is particularly relevant: quiz responses can be sent as Custom Properties, and quiz completion can trigger automated messages.

After finishing a quiz, for example, a shopper might receive an email with tailored product recommendations. If they do not purchase, marketers can use the available data in a subsequent lifecycle campaign.

Other supported integrations named in the product materials include Attentive, Postscript, Recart, Google Analytics, Meta Ads, and Gorgias. Subscription businesses can connect tools such as Recharge, Skio, Stay AI, and Loop.

These integrations allow quiz answers to inform segmentation and workflows. Passing an audience attribute to an advertising platform does not guarantee that every customer will be matched: results depend on available identifiers, consent, and the platform's policies.

MCP, REST API, and Webhooks Support External Automation

Phoenix provides interfaces for connecting third-party applications and AI assistants. Through Model Context Protocol (MCP), a compatible client—including Claude, ChatGPT, Codex, or Cursor—can interact with Octane AI data using natural language.

For example, a user might request the quiz questions with the highest drop-off rates or find customers who selected particular answers. The connection involves authentication and selection of permitted stores and actions.

REST API and webhooks support programmatic integrations. The API covers quizzes, analytics, customer profiles, exports, and experiments. Webhooks can communicate events such as quiz starts and completions, contact capture, and export readiness.

A company could, for example, send quiz-completion data to its own CRM or data warehouse. A Storefront JavaScript API also supports interactions with the quiz from the website. Phoenix can therefore serve as both a visual quiz builder and a component of a broader marketing stack.

Octane AI Performance Figures Are Vendor-Reported, Not Incrementality Results

In its Phoenix announcement, Octane AI cited historical performance data from its product quizzes. According to the company, approximately 15% of shoppers who complete a quiz make a purchase, the average email opt-in rate is around 29%, and average order value has increased by as much as 62% in certain cases.

Octane AI also reported that its quizzes have been associated with more than $1.4 billion in revenue and 15 million orders in aggregate.

These figures describe historical results across the Octane AI platform, not isolated tests of the new Phoenix product. They are not an independent estimate of incremental revenue generated by quizzes. Outcomes for a particular Shopify store will depend on its assortment, recommendation quality, traffic sources, and baseline conversion rate. The vendor figures should not be treated as guaranteed implementation results.

Phoenix Availability and Pricing

Phoenix was introduced on October 21, 2026 and is available through the Octane AI application in the Shopify App Store.

According to pricing information cited at launch, the base plan starts at $50 per month, Plus at $200, and Enterprise at $500. Plans differ in the number of available quizzes, credits, analytics capabilities, and experimentation features. AI functionality can consume additional credits.

A 14-day trial is offered for new customers. Existing Octane AI users can move their quizzes to Phoenix, while merchants using other quiz tools can request migration assistance. Exact commercial terms should be checked when subscribing because fees depend on usage and the chosen plan.

Octane AI — Product Quiz and Personalization Platform for Shopify

Octane AI is a Shopify-focused platform for creating product recommendation quizzes, capturing zero-party customer data, and connecting quiz responses with commerce and marketing systems. Phoenix expands this model with AI-assisted building, validation, experimentation, and performance analysis, while keeping marketers responsible for recommendation quality, consent, and measurement of incremental impact.