AI that works from your reviews. And asks before it acts.
Summaries, shopper answers and screening are all generated from your own published reviews. The assistant that can change things waits for your approval first.
Customers consistently praise the soft fabric and true-to-size fit, and many mention the colors holding up after washing. A few reviewers found the neckline looser than expected.
Four features, one source of truth.
Nothing here is generated from thin air. Every output points back at reviews your customers wrote and you published.
- Review summaries
- A short per-product summary written from the published reviews, shown above the review list in the widget. It regenerates as reviews arrive, and you can revert to the previous version if a regeneration reads worse.
- Shopper answers
- A shopper types a question on the product page and gets an answer drawn from that product's published reviews, with the sources shown underneath. Answers are private to the asker, and you see every question in a dashboard list. Available on request for any plan.
- Review screening
- An optional classifier reads each arriving review before the publish decision is made. A flagged review never goes live and comes back down, because the decision waits rather than retracts.
- The dashboard assistant
- A chat panel that answers questions about your own store and makes the changes you approve. It signs in as you, so it can only see and change what you can.
An assistant with your permissions, not its own.
The assistant drives the same admin API you do, signed in as you. It cannot see a store you cannot see, and every change it makes waits for an approval you click.
Approved changes land in the audit log under your name, because you approved them. There is no shadow actor to untangle later.
You
Reply to Dave’s review and thank him for the photos.
Assistant
Here is the reply I would post on his four-star Classic T-Shirt review. Nothing is sent until you approve it.
Answers for shoppers, receipts for you.
The ask box answers the question a size chart cannot: what did people like me actually experience?
“Does it run small?”
Most reviewers say the fit is true to size, though two suggest sizing up if you are between sizes. Drawn from 12 reviews that mention fit.
Answers come only from that product's published reviews, and the reviews used are shown under the answer. If the reviews do not cover a question, the box says so instead of improvising.
Every question lands in a dashboard list, which is a running record of what your product pages are failing to say. The most-asked question is your next product photo, size note or FAQ entry.
The rest of Cascade.
Every feature runs on the same customer record and the same REST API, so each one you turn on makes the others worth more.