What if your biggest competitor this year isn't another business — it's an AI doing the shopping for your customers before they ever see your website? That's not a hypothetical. It's starting to happen right now, and it's going to change how we think about selling online.
Let me explain what's happening and why it matters for your business.
The Big Shift: Selling to AI Instead of Humans
Here's something that would have sounded mad a year ago: increasingly, you're not just selling to people anymore. You're selling to AI agents that are doing the shopping on behalf of people.
Search Engine Journal has published a complete guide to what they're calling "agentic commerce", and it's worth paying attention to. These AI shopping agents - think ChatGPT, Google's Gemini, and similar tools - are starting to research products, compare options, and make purchase recommendations for users.
What does this actually mean for your e-commerce business?
Right now, when someone wants to buy a product, they might search Google, click your ad, browse your site, read reviews, compare prices, and eventually make a decision. But increasingly, they're asking an AI assistant: "Find me the best wireless headphones under £150 for running." And that AI agent is doing all the research, comparison, and recommendation work before a human ever sees your website.
The kicker? These AI agents don't browse websites the way humans do. They don't see your beautiful product photography or clever ad copy. They're reading your structured data, your product descriptions, your technical specifications. They're analysing information in ways that are completely different from human shoppers.
What This Means for Your Product Data
This is where it gets practical. If AI agents are going to recommend your products, they need to understand what you're selling. And they understand through data - specifically, structured data.
Your product feeds suddenly become absolutely critical. Not just for Google Shopping campaigns (though they're still vital), but for being discoverable by AI agents. Those product titles, descriptions, attributes, and specifications? They need to be comprehensive and accurate in ways that go beyond just appealing to human readers.
Think about it: when a human looks at your product page, they can see context. They can interpret. They can forgive a vague description if the photos look good. An AI agent? It's working purely from the data you provide.
Going forward, this means:
- Your product data needs to be more detailed and structured than ever
- Technical specifications matter more, not less
- Accurate categorisation and attributes become critical
- Your product descriptions need to answer the questions AI agents are likely to ask
I'm not saying you need to overhaul everything tomorrow. But this is the direction we're heading, and it's worth thinking about now rather than playing catch-up later.
The Creative Question: Should You Let AI Generate Your Ads?
While we're talking about AI, there's another question that keeps coming up: should you actually use auto-generated creative in your campaigns?
Google's been pushing automated ad creation for a while now. Responsive Search Ads, Performance Max, auto-generated assets - the platform wants to create your ads for you based on what it thinks will perform best.
And here's the thing: sometimes it works brilliantly. Sometimes it's utter nonsense.
The reality for your business:
Auto-generated creative can be a massive time-saver. Instead of writing dozens of ad variations manually, you provide some basic assets and let Google's AI mix and match to find what works. For businesses without dedicated marketing teams (or where your marketing manager is juggling a dozen different tasks), this can be genuinely helpful.
But - and this is important - you can't just set it and forget it.
I've seen auto-generated ads that completely miss the point of what makes a product special. Or that combine headlines in ways that are technically correct but commercially nonsense. The AI doesn't understand your brand voice, your unique selling points, or what actually motivates your customers to buy.
Finding the Balance
So should you use it? My take: yes, but strategically.
Auto-generated creative works well when:
- You're testing new products or campaigns and need to move quickly
- You've provided strong, clear asset inputs that guide the AI properly
- You're actively monitoring performance and weeding out poor combinations
- You're using it alongside human-crafted core messaging
It works poorly when:
- You just dump any old text in and hope for the best
- Your products require nuanced positioning or technical explanation
- Brand voice and messaging consistency are critical
- You're not regularly reviewing what's actually being shown
The practical approach:
Use auto-generated creative as a tool, not a replacement for strategy. Provide it with quality inputs - good product descriptions, clear value propositions, accurate headlines. Then monitor what it produces. Pin the headlines and descriptions that must appear. Exclude combinations that don't make sense.
Think of it like having an enthusiastic but slightly dim assistant. Give them clear direction and check their work, and they can be genuinely helpful. Leave them completely unsupervised and you might find they've told your customers something utterly baffling.
What This All Means Going Forward
These two developments - AI shopping agents and auto-generated creative - are part of the same trend. AI is becoming a bigger part of the entire customer journey, from discovery through to purchase.
For your business, this means adapting how you think about online selling. It's not just about having a good website and running some Google Ads anymore. It's about making sure your products and your messaging work in an AI-mediated world.
The businesses that will thrive are those that:
- Keep their product data comprehensive, accurate, and well-structured
- Use AI tools strategically rather than blindly
- Understand that both human and AI "shoppers" need to find and understand their products
- Stay on top of these changes rather than ignoring them
I'm watching this space closely because it's moving fast. We're in the early stages of a significant shift in how online commerce works. Not overnight, but steadily. The foundations you build now - quality product data, smart use of automation, clear messaging - will serve you well as these changes accelerate.
None of this means traditional paid search is dead or that everything you're doing needs to change immediately. But it does mean the landscape is shifting, and it's worth understanding which direction it's heading.