Author: Michaela Lundberg, May 12, 2026
Getting Started with AI Commerce: A Practical Guide for Brands
This month, we sat down with experts from Shopify and Nosto for a roundtable lunch in Sydney. We had one rule going in. Skip the predictions and the theory, and focus on what brands can actually do with AI right now. What came out of the room was a clear picture of where most brands sit today, and a practical path forward.
Here's what we covered.
Before we get into it, if you've been finding it hard to keep up with all the AI terminology flying around right now, we've put together an AI Dictionary for eCommerce at the bottom of this blog. It covers the key terms and concepts worth knowing, without the jargon. Scroll to the bottom before reading on, or keep it as a reference as you go.
Where are you on the journey?
The first question worth asking is an honest one. Where are you today?
Most brands fall into one of four stages.
- Foundation. Your data exists, but it lives in the wrong places, in the wrong formats, or both. Before AI can do anything useful for your business, this needs to be resolved.
- Accelerate. Your data is in reasonable shape. Now it's time to connect it. To your systems, your tools, and to the large language models that can actually act on it.
- Automate. Manual tasks that your team runs week in, week out start coming off the list. AI handles the repetitive work so your people can focus on the decisions that matter.
- Acquire. Your products and brand start showing up in AI-powered search results, in ChatGPT, Perplexity, Gemini, and across Shopify's own AI agents. Shoppers find you before they even get to Google.
Most brands are sitting somewhere between Foundation and Accelerate. A few are further along. The advice from the room was consistent. Figure out where you are, then focus on the one thing that moves you to the next stage. Trying to do everything at once slows everything down.
Start with the data. Everything else follows.
This was the most consistently raised point across all organisations in the room. Data is not the unsexy part you deal with later. It is the part to get right now.
There are two data layers worth thinking about separately.
Product data is how AI systems understand and recommend what you sell. This includes your product titles, descriptions, attributes, taxonomy, and metafields. If your product data is incomplete, inconsistent, or buried in spreadsheets rather than living natively in systems like Shopify, AI agents struggle to read it. That means lower visibility in AI-powered search, fewer recommendations, and less reach, regardless of how good your products actually are.
Shopify's Universal Commerce Protocol (UCP) is the open standard that makes product data usable by AI agents across any channel. It is built on the idea that if your product data is properly structured and machine-readable, any AI agent, on any platform, can find it, understand it, and recommend it. The quality of that data, though, is on the retailer. Taxonomy needs to be set up correctly, descriptions need to be relevant, and important information should live inside the structured data model, not in a separate spreadsheet or PDF.
Operational data is everything else your business runs on, such as orders, inventory, supplier information, and reports. This data typically lives across your ERP, your 3PL, e-commerce platform, in Google Sheets (yes.. we know you’re still doing it) and a collection of tools that don't talk to each other. Getting this connected up is what makes the Accelerate and Automate stages possible.
At MindArc, we do this using MCP (Model Context Protocol). MCP connects Claude directly to your live Shopify data and your other business systems, so your team can query inventory, pull reports, and take action from one place, without switching tabs or waiting on a developer. This is what our service AI Prompt is built on.
Connect the layers, then remove the work
Once your data is in order and your systems are connected, two things become possible.
The first is giving your team real leverage. With AI Prompt in place, your team asks a plain-language question and gets a real answer from live data. Stock levels, order status, and product performance are surfaced instantly, without having to build a custom report or chase someone across three departments.
The second is removing the manual tasks that eat up your team's time. Weekly reconciliations, supplier emails that need to become purchase orders, and documents that need to be validated before they go into your system. These are the jobs that nobody enjoys, and everybody does, week after week.
Our AI Workflow service builds automated pipelines that take those tasks entirely off your team's plate. Documents go in, clean data comes out, and your systems stay in sync. The time your team gets back is real and immediate.
The Shopify AI Toolkit, recently released, connects AI coding tools directly to Shopify's live documentation and API schemas, providing access to over 300 resources and more than 10,000 data points. It is primarily a developer enablement tool, but it signals clearly where Shopify is heading. Commerce infrastructure that AI can read, act on, and build on top of. It is worth asking your development partner how they are using it, and what it means for your roadmap.
Get found in AI search
ChatGPT, Perplexity, and Gemini are already fielding product questions from shoppers. Shopify's AI agents are already recommending products inside the platform. The question is whether your brand appears in those results or if you’re being bumped down because your competitors are doing things better.
Getting found in AI search is not the same as getting found in Google. Traditional SEO and AEO (Answer Engine Optimisation) are different disciplines. AI models need structured brand content they can cite, a clear taxonomy, and product data that follows the right schemas. If those things are missing, your brand is invisible to AI, regardless of how strong your Google rankings are.
If you want to learn more, we've written a detailed guide on exactly what this means for your Shopify store, read our guide on UCP and Product Data here.
Our AI Brand service builds and monitors your brand's presence across AI chat platforms. We create structured content that AI models can reference, track where you do and don't appear, and close the gaps. Visibility in AI search is infrastructure. Building it now matters because brands that establish an early presence benefit as AI-driven discovery continues to grow.
We audit your entire product catalogue across 100+ checkpoints, fix structural gaps, and align your taxonomy so every product you sell is readable by AI. If your data is clean, AI systems can recommend you. If it isn't, they can't.
Personalisation only works when the foundations are right
Jim Lofgren, CEO at Nosto, made this point clearly, and it is worth repeating.
Personalisation is not a feature you switch on. It depends entirely on having two data layers working together: clean product data and quality customer behavioural data. When both are in place, AI can personalise across the entire customer journey. Recommendations, search results, merchandising, and email. When they are missing or broken, poor personalisation leads to lower conversion, higher bounce rates, and reduced customer lifetime value.
The good news is that the foundation's work is not a waste of effort. Getting your product data clean and structured, getting your systems connected, and getting your workflows running automatically directly enable personalisation. It also directly enables AI search visibility. The investment compounds.
Key takeaways from the room
Start small, test, learn, then build. Trying to roll out AI across your entire operation at once is a reliable way to get nothing done. Pick the highest-impact problem, solve it well, and build from there.
Start with the data. Clean, structured, machine-readable data is the foundation on which everything else sits. It is also the most durable investment you can make.
Do one thing and do it well. One well-implemented AI workflow that removes a real task from your team's week is worth more than five half-built integrations.
Build for the long term. The brands that are investing in data structure and AI readiness now are building a compounding advantage. The work you do today will keep paying off as the tools around it continue to improve.
Not sure where to start?
MindArc offers AI Readiness Reviews. We look at where your data lives, how your systems are connected, and where the biggest opportunity is on your AI commerce journey. We then give you a clear, practical starting point to build from.
Getting started with AI can feel like a big lift. It doesn't have to be. The right starting point and the right partner make it straightforward, and the progress compounds fast once you're moving.
If you want to know where your brand sits and what to focus on first, get in touch.
