Testing the Limits of ShopifyAI: One Review Box at a Time
I’ve been building an e-commerce platform for NeevJivan Foundation, a nonprofit in India I’ve partnered with for years. I knew exactly what they needed: a custom review/comments section for each product that felt human, conversational, and functional. Instead of taking the usual plugin-heavy route, I decided to test whether ShopifyAI could speed up the process and let me build the feature directly- cleaner, faster, and more custom. This was my first attempt at using AI to replace a traditionally manual block-coding workflow.
What Pleasantly Surprised Me
To my surprise, ShopifyAI delivered an 80% accurate block of the review component I described- from just one prompt. The structure, styling, and logic were all aligned with what I envisioned. It felt like a glimpse into a future where we could prototype Shopify blocks from scratch without navigating Liquid files or downloading plugins. That spark of possibility was genuinely exciting.
What Held Me Back
But the more I worked, the more I noticed limitations.
1. I couldn’t preview the block and return to the same AI chat that generated it, making iteration harder than it needed to be.
2. The AI couldn’t build a stable backend to store and manage customer reviews: it only managed the front-end display. This meant the feature looked real but wasn’t functional beyond aesthetics. These gaps slowed down the workflow that AI initially accelerated.
Small Tweaks ShopifyAI Could Make
To truly empower users, a few adjustments would make a world of difference:
• Introduce “linked chats” where the AI remembers the exact block you generated and allows iterative tweaking after previewing.
• Provide backend scaffolding templates, even simple ones, so AI-generated components aren’t front-end illusions.
• Enable data model creation via prompts (ex: “Create a review schema with name, rating, comments”).
• Offer debugging context so users can understand why a block fails or how to enhance its functionality.
• Allow AI to manage metafields dynamically, turning static blocks into real, functional features.
Why ShopifyAI Still Feels Like Progress
Despite the friction points, I still think ShopifyAI marks a massive leap forward. I’ve been using Shopify for four years, and never before has the platform allowed this level of hands-on, generative customization. Even if it’s not fully mature, ShopifyAI is pointing in a direction where creators, nonprofits, and small brands can build faster, smarter, and with far fewer technical barriers. And that- bugs and all; is exciting.