AI chat: more value in support or product discovery?

Shopify Apps
1

Hello everyone,

I’ve been checking out different chat / AI agent apps on the Shopify App Store. From what I’ve seen, most of them mainly focus on customer support.

Many include features like ticketing, live chat, WhatsApp, and other tools for stores with support teams. Some also help with product discovery, but this part often seems more limited.

For example, some tools do not use all product details, like metafields, so they may not answer detailed product questions very well.

We’re building in this area, and so far, we see people using it more for product search, product questions, and shopping help than for classic support. Roughly, it looks like 70% product discovery / sales help and 30% support.

But usage is only one side. From a business point of view, I’m more curious about which creates more value:

AI for support, by saving time and reducing support work, or AI for product discovery and sales, by helping shoppers find and choose the right products?

Would love to hear from Shopify store owners and ecommerce managers.

Thanks a lot! :slight_smile:

2

as per my understanding, some partners are already working on it and it will help for merchant too. AI will definitely change the way of thinking and shopping :shopping_bags: experience

4

@mastroke thanks for sharing your view!

I agree, AI will probably change the shopping experience a lot, and it’s good to see more partners working on this.

From your experience, where do you think AI creates more value for merchants today: in customer support, or in product discovery and helping shoppers choose the right products?

Would be curious to hear your thoughts.

5

In my experience, customers use it for product discovery, but support is also important. However, I haven’t found the Agent that knows it all.

6

@david.kis Yes.. preference the product discovery, buyers don’t have time to spend to read unnecessary things.. they just want quality product.. if AI will help.. then it’s best way ..

7

Yes :+1: @stokeniko you are right.. we cannot ignore the support system too

8

Worth distinguishing volume from value. Your 70/30 usage split feels right — shoppers engage with chat pre-purchase more readily than post-purchase support cases, which still mostly route to email and help desks. But usage volume isn’t the same as business value.

Discovery AI value is real but hard to attribute. A shopper asks “do you have this in size 12?” and converts. Would they have converted without the AI? Without controlled A/B testing, most claims are vibes-based.

Support AI value is easier to measure. Deflected ticket = X minutes saved. Most merchants track it as a cost line and the numbers add up cleanly.

The category that’s underrepresented in your framing: post-purchase queries. Order status, shipping ETA, returns, exchanges. High volume, high intent, clear answers. Often the easiest place to prove ROI because the conversion already happened and the question is well-bounded. It’s probably the best place to invest right now, at least. Easy answers, high volume.

9

Weirdly, I think the biggest value of support AI is reducing customer anxiety more than support

A lot of tickets are basically:

Did my order disappear?
Is this delayed?
Is someone actually there?

Even when operations are fine, silence makes customers nervous really fast

That’s why post-purchase AI feels more immediately valuable to me than discovery AI right now

11

I have seen an option like Shop with AI in most of the Shopify stores. It has been automatically enabled in our store. It helps users find the questions and details about products in your store.

12

Thanks for the thoughtful breakdown - really appreciate it.

Totally agree on A/B testing being the source of truth for measuring discovery value. I spent 11 years building and marketing a product search tool, and A/B testing was always the best way to prove the benefits - but getting it right (picking the A/B test tool, the right period, setting it up, analyzing & inderstanding the results) can be overwhelming for the e-commerce team.

What we do instead is track add-to-cart events directly on the AI interface through a product quick view. We then attribute revenue based on those cart events. It’s not bulletproof, but when someone searches, clicks, reads the details, adds to cart, and orders - that journey justifies the attribution.

Does that sound like a logical approach to you, or would you recommend something different?

On the support side, we’re handling order tracking, cancellations, restocks, and refunds, plus policy questions like shipping, returns, and payment options.

Curious though - what other post-purchase support topics do you think would be most useful?

13

Thanks for sharing! Out of curiosity, what post-purchase support topics would you find most valuable to have covered?

14

Thanks for sharing your perspective!

We do cover order tracking, cancellations, refunds, and restocks, and the system can answer pretty much anything that lives on the store like policies, pages, blog articles, category descriptions, you name it. But our main focus has been product discovery - helping users find the right product, answering product-related questions so they can decide quickly, plus upselling based on the conversation context.

We haven’t gone deep on the support side yet, so really curious what else would be valuable.

One thing I’ve noticed is shoppers asking for discounts in the chat - I’ve been thinking about letting the AI offer discounts based on active Shopify promotions.

Would that be something you’d find useful?

Also, how crucial is live chat (human handoff) for you - is that a must-have or not really?

15

Thanks! I checked out the AI solution on alhalalcosmetics.nl, it looks really cool!

I did some research and as far as I can tell, Shopify doesn’t offer anything like this natively, so it must be a third-party app.

Would you mind sharing which app you’re using?

16

It is an App called Agentic, which Shopify Automatically install in our stores.

17

We have just launched an App that partially covers Product discoverability in Shops through Agentic Commerce. The Agent adds to cart and initiates Checkouts. We would be glad to get some feedbacks, knowing the app in still being enhanced and developed. More features are planned for release. AI Sales and Conversion Patrol - AI that finds products, adds to cart and initiate checkouts | Shopify App Store

18

Hi David,

I developed an app that discovers contextual products like “evening dresses,” replacing Shopify’s native search. Within the same app, it includes a sales and support agent that recommends products, which are then added to the cart directly from the chat. It also provides order tracking information, allows customers to track their orders, and integrates with the associated brand context through easy setup. The search function is already operational in a pharmacy, and we are currently configuring it for direct chat support (we provide direct support for faster delivery). My app is called SmartAgent AI Sales & Search (SmartAgent AI Sales & Search - Tracks revenue from AI sales and support | Shopify App Store).

19

Thanks for sharing. I found it, it’s the Microsoft Clarity app. The Agent is running in beta.

I installed the app, and the storefront UI looks nice, but the settings seem quite limited. I don’t see how it supports multi-market setups, and as far as I can tell, it’s only available on your NL market, but not on the others.

Is there a specific reason why you haven’t enabled it on the other markets yet?

20

Thanks for sharing. I checked it on the demo store and searched for “snowboard,” since the demo store has products like that, but it wasn’t able to return any results.

I assume it’s still quite a fresh release and may need some further improvements. Good luck with the app.

One more question: does it support Shopify Markets / multi-market setups? For example, can it handle different languages, currencies, and product availability across markets?

21

I recommend use https://farmex.cl to check de search.

Good feedback, the demo store it’s not configured yet. Yes, the app handles different languages ​​without problems. I haven’t had a client yet who operates in different markets. I’d be happy if you have any information on someone who needs piloting.

23

If anyone’s looking for a solution, Talkom is worth a test.

It’s more focused on product discovery (an actual AI shopping experience) than most similar apps.

What stands out: it measures the revenue, orders, and carts attributed to the app (with product quick view & add to cart right in the UI), so the impact is actually trackable. From what we’ve seen, stores using it are getting a solid chunk of revenue through the app.