Digital Bites March 2026
In this guide, we’re going to look into how AI is redefining product discovery and provide a roadmap for optimising your store for AI search and autonomous shopping agents.
AI searchGuides
LLMs and ecommerce: How to boost your sales
Search has evolved. With the arrival of LLMs, the journey from query to purchase has been redefined. Once viewed purely as informational answer engines, these models now serve as a vital transactional layer for ecommerce. This guide breaks down how to leverage this shift to boost sales and engagement.
Intro and overview of landscape
Over the past few years, the search landscape has changed dramatically, and more so recently with the arrival of LLMs. Once thought of as purely answer engines, they now play a vital role in the search journey for ecommerce sites.
How do AI assistants influence product discovery before users reach your site?
Let’s pretend I want to buy a new sofa, but I have pets, and as it’s an expensive purchase, I want to make sure I choose one with durable material.
There are three example user intents we could group this into:
- Learn about pet-friendly sofa materials and features (top of the funnel)
- Find recommendations for the best pet-friendly sofas (mid-funnel)
- Browse and purchase pet-friendly sofas (bottom of funnel)
Starting with the first intent group, this is top-of-funnel awareness informational content. Users have questions around materials, fabric types, durability, and pet-proofing against the inevitable scratching, shedding, and staining.
But the messy middle has evolved and users are not just doing their research in Google anymore. They are jumping from platform to platform, comparing and contrasting, looking at images and reading reviews.
LLMs are becoming central to this research journey, actively working to keep users in-platform by offering follow up suggestions.
For example, when I use the prompt ‘what are the best pet friendly materials when buying a sofa’ in ChatGPT, the platform provides lots of useful information, before strategically prompting me to pivot to a more commercial search:
When I type a more commercially focused prompt, the LLM generates a list of results from different retailers, linking through to product pages that align perfectly with my specific pet-friendly criteria.
While we’re seeing that most organic conversions still come from search engines, the growth of LLM assisted revenue is a trajectory that brands can no longer afford to ignore.
The trick is to appear everywhere along the modern search journey and be available in the right formats at every touch point to build brand recognition and increase the chances of converting.
Because LLMs have become a permanent fixture in this ecosystem, optimising e-commerce sites for AI discovery has become absolutely essential.
How to optimise content for AI search
Before you start creating an LLM search strategy, you need to do your research. That means finding out what keywords and prompts users are searching for and assigning them to different groups of intent. This will help you to plan your content across each stage of the funnel and be visible across those different touchpoints.
Optimising your commercial content for LLMs
Building LLM -friendly product content that actually converts is crucial for any LLM ecommerce SEO strategy.
- Ensure your PDPs have plenty of information that answer natural language questions. The more information about the materials, durability, colours… the better. And as for our pet-friendly example? If the material is scratch friendly, easy to wipe clean and easy to remove fur, make sure you get that information in there.
- Include detailed specifications that the models can cite.
- Include clear policies likes returns and shipping information.
- Use winning trust signals like review ratings and testimonials.
- Avoid thin or duplicate content that reduces AI visibility.
- Incorporate several high-quality images on a neutral background with good lighting, that show your products at various angles. If you’re selling clothes, make sure they’re model-worn.
Optimising your informational content for LLMs
As with traditional search, top-of-funnel content remains the cornerstone of brand discovery. Solving problem–based queries through expert-led content is crucial for brand awareness, and being front of mind for your audience can lead to purchases later on down the line; making it a crucial part of any GEO optimisation strategy.
- Create content that answers specific questions and ensure this is well structured and clearly formatted to help LLMs extract concise information.
- Use unique data and stats where possible to enhance your chances of being cited by LLMs.
- Use schema mark-up to help LLMs interpret your content faster.
- Build authority by using expert-led content and quotes.
- Include relevant internal links throughout content to not only show users other relevant informational content, but to push them down the funnel to convert.
What about agentic shopping for ecommerce sites in LLMs?
With the introduction of agentic shopping in LLMs, the retail landscape is entering an exciting new phase.
We conducted a recent poll and 75% of people who took part thought agentic shopping will take off, but only in certain verticals. And this sentiment reconciles; AI agents will search, compare and purchase on behalf of users.
In the context of our pet-friendly sofa example; would we really want an agent to make that purchase for us? Probably not. However, low-consideration, recurring purchases (like pet food) are the ideal use case for agentic shopping, where the AI can autonomously optimise for price and delivery across multiple platforms.
Is agentic shopping in LLMs fully rolled out?
Not yet. But 2026 is being called the ‘breakout year’ for agentic shopping. Remember, it’s only going to get bigger in the future, so it’s time to prepare!
The current state of play is:
- OpenAI (ChatGPT/Operator): instant check out in ChatGPT is currently only available in the US, for single purchase items and only for certain partners like Shopify sites and Etsy. It means shoppers can buy directly from ChatGPT powered by Agentic Commerce Protocol (ACP), both interfacing with the buyer and securely transacting with the merchant.
- Google Gemini and AI Mode: available only in the US.
- Perplexity: also currently only available in the US.
How do you optimise for agentic shopping?
This is where shopping feeds suddenly become super important.
For OpenAI:
Step 1. Sign up as a merchant to get approved for feed ingestion and enable checkout functionality if eligible – currently just Shopify and Etsy.
Step 2. Use ChatGPT’s product feed specification in CSV, TSV, XML, or JSON to, create and submit your feed.
Step 3. Refresh your feed. ChatGPT allows for refreshes every 15 minutes. It’s best to automate this task or risk shoppers buying a product that is out of stock.
Final thoughts, and conclusion
If 2025 was the year for testing and seeing the role of AI in users’ search behaviour, 2026 is the year learnings need to be put into action. Whilst our clients are currently seeing anywhere between 0.5-5% of their total traffic come via AI; with LLMs pushing towards more commercially-focused queries, we expect to see this number rise.
If you haven’t already, make sure you have LLM reporting set up in your GA4, so you can see traffic and revenue; and invest in tools that enable you to understand the scope of opportunity – these are our pick of the best LLM visibility tracking tools.
Ultimately, the user journey has become increasingly fragmented across traditional and conversational touchpoints; and so as you optimise your commercial – and informational – content to be seen across LLMs, it’s important you have an accurate overview of the impact your organic strategy’s having on your sales and revenue.
FAQs
My store isn’t on Shopify or Etsy, can I enable my store for instant checkout?
You can make your checkout agent ready by implementing the Open-Source ACP specification.
Can I list my products if my store isn’t enabled for Instant Checkout?
Yes, you can still submit your feed to appear in ChatGPT results and get discovered.
What keywords and prompts should I optimise for?
Optimise for long-tail conversational keywords. In-market shoppers are likely to be looking for something specific. For Google Gemini and AI Mode: This is all about merchant centre feed and schema optimisation. Ensure your attributes are all fully optimised for. For Perplexity: Mainly reliant on well optimised structure data including detailed attributes like materials, dimensions, and care instructions.
Lizzie is a skilled SEO professional with over 10 years’ experience operating at a senior level and has a background in driving organic performance for a global client base. She has a penchant for data-led, commercial strategy and in addition to her vast experience in international SEO, she has also worked with many well-known national and local brands, across multiple industries.
Lizzie Lewington
Head of SEO Performance – Connective3
Let’s get started