E-commerce and retail

The cases below are different companies with the same underlying problem: the sales channel did not match the data behind it.

What these companies face

A retailer rarely has one system. The catalog sits in a store, the stock in an ERP, the conversations in WhatsApp and the ad numbers in the platforms. When they disagree, customers are sold things that do not exist, orders get lost between chat and store, and nobody can say which channel actually pays.

The pattern repeats across the cases. A sporting goods brand had complex variants (model, size, color, stock) and sold mostly over WhatsApp, where replies were slow, sizes wrong and some orders never reached the store. An auto parts distributor had real inventory in an ERP that the online store never saw.

What KINTO builds for them

• A sales agent that looks things up. In the sports e-commerce case, the agent works from live stock and a curated spec sheet. KINTO also reconciled every order in the window with its real source and found one ad channel returning about twice as much as the other against the same cash. The owner reported channel sales about 30% above their history in the first full month; the effect isolated from seasonality is not known.
• An ERP as the source of truth. In the auto parts case, over 10,000 products went online and the agent began offering about 26% more real products while no longer offering hundreds that did not exist.
• A store that fits how the company sells. The networking distributor sells by quote, so its headless store hides price and cart when the price is zero and opens WhatsApp with the product prefilled. Its product listing weighs about 80% less than before.
• A rescue before a redesign. The specialty retailer went from a store crashing under bot traffic to a headless one, with a reproducible exclusion policy governing what appears in the channel.

What the cases show, and what they do not

The measured results are about page weight, catalog accuracy, reconciliation and response times. For later sales and conversion, most cases report no evidence yet, and the auto parts case notes that in the first weeks there were few, small orders. The sports case is the only one with a client-reported sales figure, labeled as such.

How the work started in each case

KINTO’s first deliverable is almost always a cross-check. For the auto parts distributor it compared the channel against the ERP item by item, turning the hunch that the bot offered odd things into a quantified list. For the networking distributor it cross-checked the client’s inventory against the store catalog by code and barcode. For the sports brand it reconciled orders with their source. Each cross-check turned an impression into a number, and each number set the scope of what to build.

Where to go next

If your problem is the chat, read AI agents for WhatsApp and its e-commerce cross-page. If it is search visibility for a large catalog, read the e-commerce SEO and GEO page.