Wholesale and online retail · a free guide

Less typing, and fewer sales lost at the checkout

For wholesalers, distributors and online shops.

What AI actually does in a wholesale or online business, where it helps, and where it does not. Ten short slides, a free check you can use tomorrow, and three questions that put a number on it. No sign-up.

Prepared by 25 Tera Watts · free to use, no sign-up
Click or → to move on
Industries I serveWhere the week goes

Where the week goes

New stock

Every product, by hand

“both our main suppliers, every week is new parts, every week.”
The owner of an online model shop
Each new product needs a photo, a name buyers know and a few lines of text, typed in one by one.
Checkout

The sale lost at the end

“I don't have confidence in spending a lot of money on advertising and marketing when I know that the product I have is not good enough because of the checkout.”
The owner of a flower importer
If the checkout loses the sale, money spent on adverts is wasted.
Selling

Typing after every call

“after every call, they have to type up their notes from the call.”
A manager at a plumbing and heating distributor
Notes, enquiries and supplier data get typed in again, in hours meant for customers.
You can see all three without new software. Someone just has to count them.

More sales go online, and most baskets are left

%
of retail sales in Great Britain were made online in September 2026
ONS retail sales, Oct 2026 · 28.8%, rounded
%
of online baskets are left before paying, on average
Baymard Institute, Oct 2026 · 50 studies, not UK · Baymard sells checkout research
%
of UK retail and distribution firms used AI in 2026, up from 10% in 2025
DSIT, AI adoption research, Oct 2026 · 5+ staff
One in six shoppers who leave a basket blame a long or awkward checkout. That part you can fix.
Industries I servePractical use 1 of 3

From supplier list to shop page

When a supplier sends new lines, AI applies your written rules, finds the maker's photo, drafts the name and text, and queues each one for a yes.

Example · one supplier list00:00
What stays human

The yes on every product, and every photo. AI pictures of real products failed our test: a foliage stem came back as a chrysanthemum.

Where we have done this Prototype

Tried on a flower importer's real supplier list of about 8,000 lines. Not in use.

Industries I servePractical use 2 of 3

Answers from your own stock list

A customer asks on the website or Instagram. AI answers from your own product, price and stock list, and passes anything else to a person.

Example · one customer question00:00
What stays human

Complaints, damage claims and pre-order prices. If the answer is not in your list, it says so and hands over.

Where we have done this In build

In build for a flower importer's shop and Instagram. Not live yet.

Industries I servePractical use 3 of 3

The call notes nobody types

After a sales call, AI writes the notes into your customer records. Each new enquiry arrives summed up and goes to the right person.

Example · one sales call00:00
What stays human

The call, the price and every promise. A person reads each note before it is saved, because AI can mishear a name.

Where we have done this Prototype, not built

Shown as a prototype to a plumbing and heating distributor. Not built.

Industries I serveHonest limits

What AI gets wrong, and what stays human

  1. 1It cannot follow unwritten rulesWhat never goes live, which supplier names you change for buyers. If a rule lives in one head, write it on one page first.
  2. 2It must never draw your productsAI pictures of a real product look close and are wrong. Use the maker's photos or your own, and check you have the rights.
  3. 3It gets names and numbers wrongA misspelt name or a wrong price costs trust. A person checks every word AI drafts before it reaches a customer.
  4. 4Faster listing may not be the fixIf buyers are not waiting for new products, more listings will not help. Look at the checkout first.
Start with the leak that costs most. Check it by hand for a month. Then let it run.
Industries I serveFrom our work

Where we have done this

From our own work

A flower and plant importer

Wholesale flowers and plants, with its own online shop
In build, not live yet
  • About 8,000 supplier lines arrive and about 1,500 reach the shop. For five years, each one was listed by hand.
  • A paid discovery found the checkout mattered more than loading products faster.
  • A listing tool, tried as a prototype on the real supplier list. It taught us AI must never draw real products.
  • Now in build: a shop platform upgrade, with an assistant for the website and Instagram.
What stays human

Which products go live, every photo, and every complaint or damage claim.

What we have not measured

Nothing is live yet, so nothing has been measured. We do not quote a figure.

Every project, honestly

Delivered, in build and scoped, on our Work page →

Industries I serveFree · use it tomorrow

Before you hire another pair of hands

Eight questions to answer from memory. They show where the typing goes and where sales slip away.

  1. How many new products did your suppliers send last week, and how many went live?
  2. Are customers asking for products you have not listed yet?
  3. Are your listing rules on paper: what never goes live, which names you change?
  4. Which suppliers could send a spreadsheet, so nobody retypes their list?
  5. When did you last buy from your own shop, on a phone, as a normal customer?
  6. How many hours a week go on typing up calls, enquiries and orders?
  7. Which question do customers ask most by email or Instagram, and who answers it?
  8. If your trade customers find your shop, will they feel you are competing with them?
Bonus

Test your own shop

  • Buy one item on a phone, start to finish
  • Pay once with every method you offer
  • Count the steps before the pay button
  • Search using the name your customers use
  • Check every photo shows the real product
Print this check →
Time one week of typing and buy once from your shop. Start where it hurts most.

Twenty minutes, then one page

What I will ask
  1. Where does typing by hand eat your week?
  2. Where do sales slip away: the checkout, the reply or the follow-up?
  3. Who runs your shop and stock systems, and what can they change?
How it works
Bring your number

A 20-minute call. I ask a few plain questions about how the work runs today. Within 48 hours you get one page: where the time goes, and the smallest thing that would fix it. If it is not a fit, I will say so.

Not ready to talk? Keep the check. It works without us.

If someone else signs off on things like this, bring them along. It saves a second call.
01 / 10