Approve
The proposal is published to the store.
Operations
An AI agent for the behind-the-scenes work in your store.
Product pages, translations, images, prices, and the other tasks your team repeats every week. We design each one with the person who does it.
Book a callThe work arrives ready, with its source in plain view.
An AI tool waits for someone to tell it what to do. This agent starts from what your store already has, such as these three sources, and leaves its proposals waiting for your team.
Customer questions
What customers ask customer service that the product page doesn't answer.
The store's photos
The product photos your store already has.
Competitor prices
The stores your team chooses to keep an eye on.
Output
The proposal queue, in the ShopGuru app
Four examples and your team's task
Product pages that answer customers' real questions.
When the same question keeps coming up, the agent writes the answer into the product page, in the words customers use. AI product descriptions start from what your store is missing, not from generic copy.
Sources in plain view: the supplier's spec sheet, your team's answers and the conversations the question came from.
Left blank if the data is missing: the agent leaves the field empty and asks your team.
Written for search: titles and descriptions using the terms customers search for on Google and in AI assistants.
Your catalogue in English and Spanish, with the store's glossary.
AI catalogue translation uses the store's glossary, so that "narrow fit" is worded the same way on every product page and in every language.
Clear product pages, approved in one go: a whole collection arrives as a single proposal.
Doubts flagged: a term that isn't in the glossary or a missing detail comes clearly marked.
The right tone for each market: in Spanish, the store can ask for customers to be addressed as tú.
New product images, made from the store's own photos.
The agent prepares studio backgrounds and lifestyle settings for your product photos. The product is cut out of the original and stays unchanged.
The product stays the same: only the background and the lighting around it change.
Next to the original: your team compares them before approving.
The store decides how to disclose it: the product page can tell customers the image was generated.
Base photos: dayan (CC BY-SA 4.0, manufacturer's logo removed), 12akd (CC0) and Adonis Lin (CC BY 3.0), via Wikimedia Commons. The example images are licensed under CC BY-SA 4.0.
Competitor prices, with the maths done.
The agent tracks the competitors your store chooses, compares the final price including shipping and proposes a price within your team's margin rules.
Minimum margin: no proposal goes below it.
Maximum drop: the store sets how far a price can fall in one go.
The store's rounding: if your prices end in .90, so does the proposal.
The task your team repeats, designed with the person who does it.
The four previous tabs are examples. For supplier orders, Marta checks stock every Monday and orders whatever has dropped below the minimum; the agent now prepares the order for her to approve. With any other task in your store, the path is the same, in three steps.
Who does it today, and when: we talk to the person who does it, to understand what they look at and what they decide.
The rules, in writing: the minimums, the suppliers, the lead times and the cases only your team knows about.
A proposal to approve: the agent prepares the work and your team approves, edits or rejects it, as in the other examples.
If your team does it every week, with rules that can be written down, the agent can take it on.
We map yours out in the Diagnostic.
Nothing reaches your store without your team's approval.
Everything the agent prepares arrives in the ShopGuru app as a proposal. Every decision is logged with a name and a time.
Edit
Your team changes a sentence or the price before publishing.
Reject
A perfectly normal decision. The agent only proposes the same thing again if the data changes.
Log
ExampleProduct specs · Chinelo Gerês
rejectedSónia, 29 Sep, 11:20
"It's leaving the catalogue in October."
Generated images · Sandália Foz
approvedSónia, 28 Sep, 10:15
Product specs · Sapato Ria
approvedSónia, 22 Sep, 15:02
It publishes to the platform you already use (Shopify, WooCommerce, PrestaShop and others) or to your catalogue spreadsheet, whichever way your team already works.
Store report
Monday, 9am: what customers asked during the week.
By email, to your team: a summary of the week's conversations, analysed by the agent.
- What they asked, by number of times
- New questions that went unanswered
- What held up purchases and why items came back
- What's waiting for approval
From operations to customer service
The approved product page reaches the conversation.
On 22 September at 15:02, Sónia approved the narrow fit on the Sapato Ria product page. Since that afternoon, when a customer asks whether the Ria runs small, the agent answers with the size the product page recommends.
See customer serviceQuestions about the operations agent
Can the agent change a price on its own?
No. A price only changes when someone on your team approves it, logged with a name and a time. While the proposal waits, the store keeps selling at the current price.
I already use my platform's AI to write descriptions. Why would I need this?
Your platform's AI helps your team when your team asks it to. This agent starts from what your store already has, such as customer questions and competitor prices, prepares the work without anyone asking and leaves it ready to approve. The two can work together.
Do the generated images change the product?
No. The product is cut out of the original photo and only the background and lighting change. Each image arrives next to the original, so your team can compare them.
Where does the product page data come from?
From the supplier's spec sheet, your team's answers and conversations with customers. If no source has the information, the field stays blank and the question goes to your team.
Does it publish straight to my store?
It publishes to the platform you already use (Shopify, WooCommerce, PrestaShop and others) or to your catalogue spreadsheet. Publishing directly depends on the integration with your platform, which we confirm in the Diagnostic.
Does the agent only handle product pages, translations, images and prices?
No. Those are examples. If your team does a task every week, with rules that can be written down, such as supplier orders, the agent can take it on. We map it out in the Diagnostic with the person who does it.
Not sure where to start?
There are two ways in. If you're still working out whether the agent is right for your store, start with a call. If you already know what you want to hand over to it, go ahead with the Diagnostic.
Call
to get started
Half an hour to see whether it makes sense.
- On video, with the engineer who will look after your store
- Tell us how your store works, what your team repeats every week and the tools you use
- At the end, we tell you whether ShopGuru can help and where to start
Diagnostic
if you're ready to go ahead
Your store's processes, mapped with the people who run them.
- Kick-off meeting: how things work today, which tools you use and who does what
- A video session with each person who runs a process, so we can see how they do it
- The written plan in 10 working days: what can move to AI, where to start and how much it costs
- A final meeting to present the plan to whoever makes the decisions at your store
€490+ VAT · one-off payment
Buy the DiagnosticDeducted from the Implementation if you go ahead within 60 days of the plan. If you don't, the plan is yours to keep.