What you can build with your AI tool, product by product
The requests worth making in each part of the store, and how a good one becomes a play
Published
With your AI tool connected to Hello Retail, every product in the store can be looked at and changed from a request in your own words.
This guide goes through the products one at a time. For each it says what you can build, and shows requests the connection documents an answer for today. The requests are read from the live server, so the list grows as new tools ship.
Under every request there is a note on whether it only looks or also changes something after your yes. If that distinction is new, start with our guide to what your AI tool can change in your store.
Search
Search is where shoppers tell you what they want in their own words, so it is the richest place to start. You can find the searches that lead nowhere and fix them, decide what comes first for a given query, choose the filters shoppers get, and change how the search box looks.
- “Which searches find nothing, and can you fix them?”Looks, then changes after your yes
- “When people search for X, show them this product first”Looks, then changes after your yes
- “Push our own brand higher in search results”Looks, then changes after your yes
- “Are we offering the filters shoppers actually use?”Looks, then changes after your yes
- “Show something useful before people start typing”Looks, then changes after your yes
Product: Search. Background: What is ecommerce site search?
Recommendations
Every recommendation box has a job: Show alternatives, complete the basket, bring back what the shopper looked at. You can ask what each box is set up to do, find the one that underperforms, change its strategy, give all boxes a new look, or add boxes to a new landing page.
- “Which recommendation box is underperforming, and what should we change?”Looks, then changes after your yes
- “A recommendation disappeared from our product page”Looks, then changes after your yes
- “Give all our boxes the new look”Looks, then changes after your yes
- “Add recommendations to our new landing page”Looks, then changes after your yes
Product: Recommendations. Background: Product recommendations for ecommerce
Pages
Category and campaign pages are where merchandising ideas become visible. You can ask why a category page converts badly, put chosen products first on a sale page, build a page for a campaign, or copy a setup that works to another category.
- “Our category page converts badly - why?”Looks, then changes after your yes
- “Show these products first on our sale page”Looks, then changes after your yes
- “Build a page for our summer campaign”Looks, then changes after your yes
- “How are our pages doing this month?”Looks only
Product: Pages
Product Agents
Product Agents send emails on their own when the moment is right for one customer, such as a price drop on something they looked at. You can decide which agents run and for which products, adjust how the emails sound, and read the results agent by agent.
- “Start emailing customers about price drops, but only for our own brand”Looks, then changes after your yes
- “Make the agent emails sound more like us”Looks, then changes after your yes
- “Never let the agents email about cheap or clearance products”Looks, then changes after your yes
Product: Product Agents
Retail Media
Retail Media lets the brands you sell pay for sponsored placements in your store. You can see which campaigns run and how they perform, and have product and banner campaigns drafted for a brand. A drafted campaign stays in review until a person makes it live in My Hello Retail.
- “Which sponsored campaigns are running, and are they working?”Looks only
- “Set up a sponsored placement for a brand we sell”Looks, then changes after your yes
Product: Retail Media. Background: What is Retail Media?
Product Intelligence
Product Intelligence is the data that tells you what is selling and when. On its own it answers planning questions. Combined with the other products it becomes action: The tool reads the top products and then places them in search, on a page or in a newsletter.
- “What should we promote this weekend?”Looks only
- “What should we push for Christmas, or in Q3?”Looks only
- “Which products should we discount, and which never need it?”Looks only
Product: Product Intelligence
Data and housekeeping
Everything above depends on product data that is correct and current. The same connection lets you check the feed, trace a wrong price to its source, and see who changed a setting.
- “Why does this product show the wrong price?”Looks only
- “Connect our product feed”Looks, then changes after your yes
- “Did last night's feed run go wrong?”Looks, then changes after your yes
- “Who changed this, and when?”Looks only
- “Show me the failed API calls from the last day”Looks, then changes after your yes
From request to play: How Agentic Playbooks fit
A request that works once is worth keeping. Written down with what the tool did and what came back, it becomes a play: Something you run again next month, hand to a colleague, or schedule for Monday morning.
Agentic Playbooks is Hello Retail's library of such plays. Each one gives the request to copy, the steps the AI tool takes with it, and the reason it matters, written up from a run that actually happened.
Browse the plays in Agentic Playbooks.
A request is the smallest unit. A play is a request that has proved itself. A store that keeps its plays builds something no other store has, because the ideas in them are its own.
For the thinking behind this way of working, see Build your own ecommerce personalization.
Frequently asked questions
What can I build with an AI tool connected to Hello Retail?
Anything the products can be configured to do, described in your own words. Typical examples are fixing searches that find nothing, deciding which products lead a category or sale page, setting up recommendation boxes on a new landing page, designing product tiles for a newsletter, tuning the emails Product Agents send, and drafting sponsored placements for a brand you sell.
Do I need all the products to get started?
No. The AI tool works on the products your agreement includes. If you ask for something that needs a product you do not have, the tool can pass your interest on to Hello Retail so someone contacts you.
Where do the example requests in this guide come from?
From the Hello Retail connection itself. It carries a list of customer questions it documents an answer for, and this guide reads that list from the live server. The wording is the server's own.
What are Agentic Playbooks?
Agentic Playbooks is Hello Retail's library of plays: A request to copy, what the AI tool does with it, and why it matters. Each play is something a team has actually run, written up so another store can run it too.
Can one request use several products at once?
Yes, and that is often where the best ideas are. A request about what to promote this weekend can read the top products from Product Intelligence and then place them in search, on a category page and in the next newsletter, with a confirmation for each change.
What should I build first?
Start where you already know something is wrong. Searches that find nothing and a recommendation box that underperforms are the two most common starting points, because the data shows the problem and the fix is small.
What will you build next?
Connect Claude, ChatGPT or OpenAI Codex to Hello Retail and try one request from this guide.
Start building with our MCP