AI Shopping Isn't a Future Behaviour. Retailers Are Already Measuring It
By Malin Gaertig · September 22, 2026
Two measurements from the ARC AI Summit — Bunnings' AI assistant converting at roughly twice the rate, and Shopify's AI-agent traffic tripling with 58% higher conversion — show the retail goal moving from earning the click to earning the selection.
- —AI shopping is already measurable: Bunnings reports Buddy users converting at roughly twice the rate of non-users.
- —Shopify data shows AI-agent traffic and orders tripling in 12 months, with 58% higher conversion than organic search.
- —One example measures AI inside the shopping experience, the other measures AI agents arriving from outside it — the same shift from different angles.
- —The journey is moving from search, click, compare, decide to ask, shortlist, select, buy.
- —Earning the selection requires product information a system can understand, compare, explain and trust — and a measurement of whether it selects you.
| Traditional search | AI shopping | |
|---|---|---|
| The journey | Search → click → compare → decide. | Ask → shortlist → select → buy. |
| Where AI sits | After the search, inside the retailer's site. | Before the click, researching and comparing on the customer's behalf. |
| The goal | Earn the click from a results page. | Earn the selection from a model's shortlist. |
| What you optimise for | Ranking, traffic and conversion rate. | Whether the system understands, compares, trusts and selects your products. |
AI shopping isn't a future behaviour. Retailers are already measuring it.
At the ARC AI Summit, two examples made the point from very different angles.
What did Bunnings find when customers used its AI assistant?
Vivek Pradhan shared early results from Buddy, its AI shopping assistant. Customers using Buddy were reported to convert at roughly 2x the rate of those who didn't.
That is an on-site feature result. It doesn't prove Buddy caused the lift — customers who choose to use an AI assistant may already be more engaged. But it is a useful signal: AI assistance can be associated with materially different customer behaviour.
What is Shopify seeing from AI-agent traffic?
Shopify data pointed to AI-agent traffic to retailer websites had tripled in 12 months, with orders from that traffic also tripling. AI-agent shoppers were converting at a 58% higher rate than organic search traffic.
Admittedly, to date AI-agent traffic still accounts for a small share of total traffic, directionally it points to a telling story.
One example shows what happens when AI is introduced inside the shopping experience. The other shows what happens when AI agents begin sending customers to retailers from outside it. Different measurements from the same emerging shift.
How is the customer journey actually changing?
Increasingly, the customer journey looks less like: Search → click → compare → decide — and more like: Ask → shortlist → select → buy.
The AI may do more of the research and comparison before the customer ever reaches the retailer. And that changes what retailers need to optimise for.
What does earning the selection actually require?
Traditional search was largely about earning the click. AI shopping is increasingly about earning the selection, which means: Can the system understand your products? Can it compare them accurately? Can it explain why one is better for this customer? Can it trust the information it finds? And, ultimately, does it select you?
What should retailers do with this now?
This is where the commercial opportunity starts to become interesting. The retailers getting into AI now don't need to predict exactly how shopping will work five years from now. They can put AI into the journey, measure what changes and learn from the behaviour.
That is also the shift MARK is monitoring: from measuring whether a brand gets seen to understanding whether AI selects it — and why.