Overview
Shopify SimGym is a first-party AI simulation tool that lets you test storefront changes with synthetic shoppers before real customers ever see them. Released as part of Shopify's Winter '26 Edition, SimGym represents a new category of ecommerce optimization: simulated A/B testing. This glossary article covers what SimGym is, how it works under the hood, eligibility requirements, pricing, known limitations, and how it fits alongside Shopify's native Rollouts feature and traditional testing tools.
Contents
What is Shopify SimGym?
Shopify SimGym is a Shopify app in AI Research Preview that uses AI-powered shoppers to simulate buyer behavior on your online store. SimGym creates AI shoppers that browse your store the way real customers do. These shoppers interact with your products, navigate through collections, add items to their cart, and provide qualitative feedback about their experience. You can compare your active theme against a draft theme to determine which has a better add-to-cart ratio among the AI shoppers.
Think of it as a focus group that runs on demand, in minutes, without needing real traffic. Shopify describes it as working "like a flight simulator for your online store, letting you get early signals with AI shoppers before real customers see them."
How SimGym works in Shopify
SimGym is a first-party Shopify app, built by Shopify, not a third-party, that sends AI-powered shoppers through your store before you publish changes to real customers. These aren't bots running random click patterns. Shopify trained the shopper personas on behavioral data from billions of real transactions across their platform, then tuned those personas to your specific store's customer behavior.
Hundreds of robots browse the merchant's store in a contained environment. Each has a persona, a budget, and a shopping intent. Each runs in a cloud browser. No mocked DOM, no shortcuts. The robot sees the page, decides what to do, clicks, scrolls, adds to cart. A twin does the same on the alternate theme. A simulated A/B test in minutes instead of weeks, even for stores with zero real traffic.
Under the hood: Shopify + NVIDIA infrastructure
Under the hood live two machines. Browserbase runs the cloud browsers: up to 2,000 concurrent Chromium sessions. The inference cluster runs the model behind the robots. The loop is tight: Browserbase sends the page state, the model returns a JSON action, Browserbase executes it. Repeat 10-13 times. SimGym is a Shopify + NVIDIA collaboration.
Two simulation types
You can run two types of simulations: Theme comparison, which lets you compare your live theme against another installed theme to analyze how theme changes affect shopper behavior; and Single theme analysis, which lets you analyze any live or draft theme on its own to get feedback on what's working and where to improve.
Optionally, you can select a Focus area to direct AI shoppers to a specific part of your store: Home page, Products, Collections, Cart, or Search.
Example: testing a theme redesign before launch
Say you've rebuilt your store's navigation and refreshed the product page layout. Instead of publishing the new theme and watching your conversion rate for days, you install the SimGym app, select your draft theme as the challenger, and run a comparison simulation. Each simulation surfaces what's working, what's not, and recommends which theme performs better so you can make storefront decisions backed by data, not guesswork. Results arrive within 10 to 15 minutes.
What SimGym reports
You get a detailed report covering: navigation and product discovery, add-to-cart flow and friction points, and directional performance indicators for add-to-cart, navigation, and more. AI shoppers browse, navigate collections, add products to their cart, and leave qualitative feedback. Shopify positions it around behavioral signals, such as add-to-cart performance, plus AI-driven shopper feedback on navigation experience, checkout experience, and store layout.
Eligibility and pricing
SimGym works with stores that meet all three of these conditions: a Liquid storefront (Hydrogen and headless storefronts are not compatible), Shopify Network Intelligence (SNI) must be active on your store, and no password protection (your store must be publicly accessible). Your store also needs at least one product listing. For theme comparisons, your store needs at least two themes installed, including your published theme.
SimGym uses a pay-per-use pricing model where each simulation you run costs one credit. During the research preview, you might be allocated free credits. After your free credits are depleted, you can purchase top-ups to continue running simulations. More details on pricing are available in the SimGym app.
As of March 11, 2026, SimGym is available to all eligible merchants with no waitlist required.
SimGym vs. Rollouts vs. traditional A/B testing
The Winter '26 Edition shipped two complementary testing tools. Understanding when to use each one (or both) is key.
SimGym runs synthetic shoppers against your themes in a contained environment. It delivers a simulated A/B test in minutes instead of weeks, even for stores with zero real traffic. It is a pre-validation tool, not proof of how real humans will behave.
Rollouts shipped as part of the Winter '26 Edition, and it means you can split-test theme changes without installing anything. Rollouts lets you stage changes to your online store, control what percentage of visitors see them, and schedule when they go live. This uses real traffic for real statistical confidence.
SimGym is designed to work as the first stage in a two-step workflow alongside Rollouts. The idea is straightforward: use SimGym to catch the obvious problems synthetically, then validate with real traffic through a staged Rollout.
Tip - Use SimGym to narrow choices and fix obvious issues, then validate the highest-impact changes with controlled live A/B testing using actual traffic.
Known limitations
SimGym is currently in AI Research Preview, which means it's real and available, but it's v1, and Shopify has been transparent about its limitations. Right now, simulations can only compare your active theme against another theme from your theme library. You can't test individual page elements, hero copy, or specific section layouts without duplicating your entire theme to change them. Checkout flows, pricing, discounts, and anything outside the theme are out of scope.
Simulations use AI shoppers that mimic real buyer behavior, but results might differ from actual buyer behavior. AI shoppers model human patterns but cannot perfectly reproduce every nuance of real user behavior, emotions, or context-driven decisions. Real-world conversions depend on marketing mix, traffic sources, seasonality, and brand reputation, factors not fully captured in isolated simulations.
Side note - Hydrogen and headless storefronts are not compatible with SimGym. If your store runs on a headless architecture, you cannot use this tool.
Why SimGym matters
Every small to medium Shopify store faces a common problem: you don't have enough traffic to run statistically significant A/B tests. With 1,000 visitors monthly, properly testing a major change requires weeks or months of data collection. SimGym sidesteps that bottleneck entirely by generating synthetic shopping sessions on demand.
Smaller businesses gain insights they'd never access otherwise, while larger merchants leverage their existing traffic patterns for risk-free testing. For brands doing serious volume, the cost of a failed redesign isn't theoretical. It's real revenue lost while you diagnose what went wrong.
More brands running more experiments means more conversion intelligence flowing through the platform, which feeds better AI models, which makes tools like SimGym more accurate over time. This creates a flywheel: the more merchants use SimGym, the better the synthetic shoppers become at predicting real behavior.
For years, rigorous storefront testing was a specialist activity. You needed to know which third-party app to buy, how to set up proper A/B testing methodology, how to interpret results without getting fooled by novelty effects or seasonal variance. Shopify is now building that capability directly into the platform, designed to be accessible to merchants who never would have run a proper test before.
Related concepts
Shopify Rollouts: Shopify's native A/B testing and staged deployment feature. Rollouts splits real traffic between theme variants and measures conversion impact. Use it after SimGym to validate findings with actual visitors.
Shopify Network Intelligence (SNI): A data-sharing setting required by SimGym. SNI lets Shopify use aggregated, anonymized shopping patterns from across the platform to power features like SimGym's AI personas.
A/B testing: A controlled experiment that shows two variants to different groups of real users and measures which performs better. SimGym simulates this process synthetically; Rollouts and third-party tools like Shoplift or Intelligems run it with live traffic.
Conversion rate optimization (CRO): The practice of systematically improving the percentage of store visitors who complete a desired action. SimGym is a CRO tool that focuses on the pre-launch validation phase.
AI Research Preview: Shopify's early-access program for experimental AI features. Products in Research Preview are functional but may have limited scope, and their capabilities can change before general availability.
View on app store