Bring personal shopping to your own site
5 min read
Every good shop has someone on hand to help.
In store, that's the person who knows the products inside and out, asks what you're looking for, and helps you find the right thing for you. Online, that person either doesn't exist or is rarely as available, so the shopper is left to work things out from product descriptions.
Today we're launching Shopping Agent, DigitalGenius's AI shopping assistant, to emulate that store assistant for your own site. Shopping Agent answers product questions, compares and recommends products, and guides shoppers to checkout, using your product catalog, policies, and business rules. More shoppers get the help they need to find what they're looking for, without you having to staff every conversation.
The help stops where the decision starts
Think about the questions that come up just before someone buys:
How does this fit compared with the brand I usually wear?
Which of these two is better for what I need?
It's a gift, so how do I get the size right?
What goes with it?
Each one is a moment where a sale is won or lost. If the customer is lucky, the answer is buried in a product description - but otherwise they have to wait until an agent picks up their ticket. If they are prepared to wait that is. More often than not, they'll find the information and support they need somewhere else. Customer service teams are often offline in the evenings and at weekends, when most shoppers are browsing. To get a question answered, they have to leave a message and wait till the morning, or later.
In no world is that a good customer experience. And the cost to a business of that bad customer experience is real. 44% of consumers have abandoned a purchase over insufficient product information (Syndigo, 2025). You've done the hard work of bringing an engaged shopper to your site and onto the product page, then they fall at the last hurdle. That’s money left on the table, or worse to a competitor.
We started after the sale
We've spent more than a decade building AI customer service for ecommerce brands, and most of that work happens after the sale: tracking orders, handling returns, changing orders. That's where brands first trusted our AI to act on their behalf.
Over that time, the same chat window kept filling up with a different kind of question. Will this fit? Which one is best for me? What goes with it? Those are buying questions, and they deserve the same quality of response as a returns request.
Shopping Agent takes what we've learned about acting for brands after the sale and brings it to the moment before it. For the shopper, it should feel like one conversation with your brand, from the first question about a product to the question about where the order is.
We believe that's what conversational commerce should be: a brand's own AI, on its own site, giving the kind of advice a good sales assistant would.
One shopper, one gift, one conversation
Picture someone on a jewelry site late on a Sunday evening, looking for an anniversary gift. They know roughly what they want, but not which piece, and the customer service team went home hours ago.
They ask for help. Instead of scrolling through every collection, they ask for "a timeless gift". Shopping Agent asks what they have in mind, then recommends a ring from your catalog and shows it in a carousel, with a clear route to the product.
Shopping Agent recommends a ring from the catalog for "a timeless gift".
They compare. They're torn between two rings, so they ask Shopping Agent to put them side by side. It lays out the shape and style of each, so the shopper can see which one suits them better.
Two rings side by side, compared on shape and style.
They get the answer that decides it. The last thing stopping them is size, because it's a gift and they can't ask. Shopping Agent explains how to measure a ring they already own against the size guide, and offers the team if they'd rather talk to a person.
The answer that settles it, with a route to the team if they'd rather talk to a person.
That's a question, a recommendation, a comparison and an answer, all on your site and all in your brand's voice. This is the conversation a good sales assistant would have had in store, but done online in minutes and seconds.
Shopping Agent doesn't have to wait hiding in the chat widget, either – like a hungover teenage shopping assistant trying to avoid customers. You can embed it on product pages with suggested questions you choose, or invite shoppers to chat when they spend time on a product page. Think about when you’d most like to jump in and help a customer – that’s where Shopping Agent can live, creating a seamless customer experience.
Suggested questions on a product page open the conversation where the shopper is already looking.
Shoppers have started asking
Shoppers have got used to asking, rather than looking. They ask AI assistants to research a purchase, compare options and shortlist what to buy, and they arrive on your site further along the shopping journey than ever. Adobe found that traffic from AI tools converted 90% better than non-AI traffic in September 2026. That makes sense: someone who has done the research has a higher propensity to purchase.
But research isn't the same as deciding. Only 11% of consumers say they would let AI make the purchase decision for them (Gartner, 2026). They still want reassurance about the final details, a comparison with the product next to it, and information that only you and your brand can give them. As personal AI agents get better, more shoppers will buy from you without ever visiting your store. The ones who do visit will expect to get answers there too. Brands need to create online experiences that will keep customers wanting to come back.
What it means for ecommerce and CX leaders
We built Shopping Agent for two teams who often share a chat window but measure it differently.
For ecommerce leaders, it's about helping more shoppers buy. Shopping Agent answers the question holding a shopper back at the moment they're deciding, rather than hours later by email. It also suggests relevant additions from your range, based on what the shopper has told it, so the basket grows with complementary products. And because it's always on, it doesn’t matter when customers are actually shopping – they get the best assistant in the store.
For CX leaders, it's about time and focus. Routine buying questions about fit, materials and care no longer join the queue. That gives your team more time for the customers who need a person. When someone does need help, Shopping Agent hands over with the full context, so nobody has to start again. You can also see which product questions keep coming up, which shows where your product pages could be doing more.
For both, shopping is connected with care. Shopping Agent works alongside Care from DigitalGenius, our post-purchase customer service agents, so the brand that helped someone choose is the same one that helps them after they buy.
Your team stays in charge
An AI that sells on your behalf has to sound like you and stick to your rules. We've built Shopping Agent so your team stays in charge of both.
Train it on your knowledge. Shopping Agent uses your product knowledge from the sources you give it, including your product catalog and knowledge base, and pulls from the latest source before answering.
Give it your voice. Set the tone to match how your team advises customers.
Test before launch. Review realistic answers and refine the guidance before any shopper sees it.
Set the limits. Choose which actions it can take and when your team steps in.
Underneath, our hybrid AI approach pairs generative AI with deterministic decisions. Generative AI handles the conversation, while the business rules you set decide what it can say and do. That helps keep the agent on topic and prevents it from giving advice that contradicts your rules.
What we've seen so far
The evidence that pre-sale conversations sell is growing. In randomized trials, researchers at Columbia Business School found that pre-sale chat lifted sales by 16.3%. Our own results back that up, with brands seeing double-digit conversion rates and uplift.
DreamCloud is a good example. Its customers often shop late in the evening. With Shopping Agent, DreamCloud has seen a 14% chat conversion rate, and a 293% return on investment from product recommendations.
"To get 24/7 coverage has been so helpful for our customers who we know shop late in the evening." Ren Kerekere, Senior CX & Systems Manager, DreamCloud
See it on your own site
We think every brand's site should be able to give the kind of advice a good sales assistant would, at any hour, in its own voice. Shopping Agent is how we're building towards that.
Shoppers are going to keep asking before they buy. The answer should come from you.
If you'd like to see what it would look like for your products, bring the questions your team hears most and book a Shopping Agent demo.
Frequently asked questions
What is the difference between an AI shopping assistant and an AI shopping agent?
An AI shopping agent usually works for the shopper. Tools such as ChatGPT and Gemini research, compare and sometimes buy products across many retailers. An AI shopping assistant works for the brand. Shopping Agent sits on your own site and answers from your product catalog, policies and business rules.
How is Shopping Agent different from an ecommerce chatbot?
A traditional ecommerce chatbot follows scripted paths and keyword matches. Shopping Agent answers product questions from your own sources, asks follow-up questions before recommending, and compares products side by side. Generative AI handles the conversation, while deterministic rules make sure each answer follows the business rules you set.
Does Shopping Agent work outside office hours?
Yes. Shopping Agent answers product questions 24 hours a day, including evenings and weekends, when many shoppers browse and many customer service teams are offline. DreamCloud uses it to cover customers who shop late in the evening.
Where can Shopping Agent appear on my site?
Shopping Agent can run inside your chat widget, or you can trigger it elsewhere on the site. For example, you can invite shoppers to chat after they spend time on a product, or embed it on product pages with suggested questions you choose.
What happens when a shopper needs to speak to a person?
Shopping Agent hands the conversation over to your team when a person is needed, so the shopper doesn't have to start again. You decide which actions the agent can take and when your team steps in.
