It knows what's
actually on the shelf.

OnHand is an AI assistant for retail floor staff. Ask it a customer's question — typed or spoken — and it answers from your store's own inventory: the right product, how many are left, which aisle. And when you don't carry it, it says so instead of making something up.

Runs on a store device Voice or text Live stock & aisle Won't invent stock

Where this came from

When a customer asks something hard, staff quietly pull out their own phone and ask a chatbot that has never seen the store. — observed on a retail floor, not in a strategy deck

That workaround is everywhere, and it half-works. A general chatbot knows a lot about products in the abstract. It knows nothing about this store.

  • It can't tell you if it's in stock right now.

  • It can't tell you which aisle to walk the customer to.

  • It will confidently recommend something you don't carry.

  • It happens on a personal device, off the record, outside the store's control.


How it works

Three steps, no IT project.

OnHand reads your product list the way a very fast new hire would — then never forgets it.

STEP 01

Your catalog goes in

Export your products from whatever POS you already use. A setup tool matches your columns — SKU, price, quantity, aisle — and loads them. Most stores are live the same day.

STEP 02

Staff ask in plain words

"Customer wants to hang a 45 lb mirror on drywall." No SKU, no menu, no training. Typed on the floor, or spoken hands-free while carrying something.

STEP 03

It answers from your shelf

The right product for the actual job, with on-hand count and location. If the first choice is out, it offers the best in-stock alternative — from your catalog, not the internet.


The part that matters

The most useful thing it does
is admit what it doesn't have.

A tool that invents stock is worse than no tool — because staff believe it, repeat it to a customer, and the trust is gone. So OnHand is built to refuse.

Generic chatbot

"Need N95 masks for sanding drywall."

Yes — you'll want a NIOSH-rated N95. Most hardware stores stock them in the safety aisle, usually near the respirators. A 10-pack typically runs about $20.

Sounds right. Store is out of stock. Customer sent to an empty shelf.

OnHand

"Need N95 masks for sanding drywall."

Out of stock — the 3M N95 10-pack (Aisle 1 · Bay C) is at zero. There's no direct substitute in stock right now, so let the customer know rather than sending them to the aisle.

Grounded. Honest. The associate keeps their credibility.

Every answer is tied to a real SKU on a real shelf. If it isn't in your inventory, OnHand won't say it is.


For a small store

What actually changes on the floor.

Not "AI transformation." Four specific things a small retailer feels in a week.

Ramp-up

A new hire stops being a liability

The gap between someone's first week and their fifth year is where sales quietly leak. A new person can answer like the veteran standing next to them, on day one.

Floor time

Nobody walks away to go check

The answer arrives while you're still standing with the customer. No trip to the back, no "let me find someone" — which is where a lot of sales are lost.

Attachment

The obvious add-on gets offered

Sold a drill? It surfaces the bits, the case, the second battery — from what's actually in stock. The kind of prompt a great associate does automatically.

Control

Off personal phones, onto your device

Staff are already asking AI. This puts it on a store device, grounded in your data, answering with your products instead of a stranger's guess.

Honest notes

What I found when I checked.

I built OnHand because I watched the problem happen on a sales floor. Then I did the thing that's easy to skip: I checked whether anyone had already solved it.

They had. Ace Hardware, Lowe's and Walmart have all shipped associate-facing assistants grounded in store inventory — at a scale a solo build can't approach. That closed one door and opened a smaller, more honest one: those tools serve their own stores. The independent retailer with a few locations and no in-house team still has the problem and no way to buy a solution.

So OnHand isn't a pitch to displace anyone. It's a working answer to a narrower question — what does this look like for a store that will never get an enterprise rollout? Built to run on a real catalog, deploy in an afternoon, and be honest about its limits.

See it answer
a real question.

The demo runs on a sample hardware catalog. Ask it something a customer would actually ask — then try to catch it inventing something.