spotIndustry
Scale
104,000 products
Modules
Years on market
18
Integrations
IdoSell
Results
Key results
Measured in AskSpot over a 30-day window. A typical chat-driven order runs to 2.3 items — a kilo of beans and the filter or the descaler that goes with it. More than one chat in five ends in a purchase, the highest rate we have measured on any merchant.
€47,600
in chat-driven orders a month
29.1%
of chats fill a basket
21.6%
of chats end in a purchase
68%
of chats resolved
“AskSpot doesn’t just answer questions — it sells. The numbers we’d hoped for from a chatbot, we got from AskSpot in the first month.”
Before AskSpot
Two shops sharing one basket
Konesso sells coffee, and coffee is chosen by taste: sweet, no bitterness, no acidity, and it still has to hold its own under milk. Nothing on a product page answers that. It is the question you ask a person behind a counter, and answering it well takes someone who has actually drunk the thing — which meant an adviser, one shopper at a time.
The other half of the catalogue does not work like that at all. Machines, grinders, moka pots, and behind them a long tail of parts that either fit or do not: a water filter for a Jura E8 in the EC revision, a 63 mm gasket for a four-cup Bialetti Venus, a milk tube set for a C8. That is a fitment problem, closer to car parts than to coffee, and getting it wrong means a return. Both kinds of question arrived on the same widget — and most of them after the shop had closed.
What decides a coffee sale
taste
What decides a parts sale
the model number
Where both answers lived
a trained adviser
Help after closing
none
Rollout
How they rolled it out
One module, three jobs, in the order a coffee purchase actually happens: pick the beans, pick the machine, then keep the machine alive.
Step 01
Taste-Profile Grounding
Choosing coffee the way a barista would
The 103,730-product IdoSell catalogue was indexed on the things that actually decide a bag of coffee — roast level, acidity, body, tasting notes, the brewing method it was blended for, and the roast date. The agent also carries one service fact it can act on: Konesso grinds to order, so it asks how you brew before it recommends.
What shoppers ask
- A sweet coffee, no bitterness, no acidity, for an automatic machine
- Can you grind it for a V60 before you ship it?
- Decaf beans for a dripper
Step 02
Machine Advisory
The four-figure decision
Espresso machines are where the money is and where the questions get hardest — Jura against Melitta, a milk jug against a steam wand, one revision of a model against the next. The agent was given the comparison axes that genuinely separate them, plus the two facts that close the sale: warranty length, and where the nearest authorised service point is.
What shoppers ask
- Nivona 820 or Melitta Barista SE F83 — which is better?
- A machine for an office of 20, up to €1,150
- What is the difference between the 820 and the 821?
Step 03
Parts & Fitment Matching
The part that has to fit
Filters, gaskets, milk tubes, descalers and cleaning tablets were mapped to the machines they fit, down to the revision — a Jura E8 in the EC revision takes a chipped CLARIS Smart PLUS and will not accept the older cartridge. It is the same job a car-parts catalogue does, and it is why 57% of the chats where the agent recommends something end in a click on one of its picks.
What shoppers ask
- Which water filter fits a Jura E8 EC?
- Do you have gaskets for a 4-cup Bialetti Venus?
- Do Kalita Wave 155 filters fit a Timemore Crystal Eye B75?
From real conversations
What Konesso shoppers actually ask
Half of everything asked is one shopper choosing one thing — finding it, checking the spec, comparing two models, confirming a part will fit. Order status is 3.9%. This is a selling desk, not a support desk. The surprise is third place: discount codes and loyalty points are 15% of all conversations, and most of those are not "give me a code" but "why isn't my code coming off the total?" — a checkout problem the agent can actually solve.
About 8 messages per conversation · 54% arriving outside business hours.
More case studies
See how other stores use AskSpot
More stories from teams that automate support and grow sales with AskSpot.
How Morele.net cut customer service costs by 90%
Morele.net cut support costs 90%, from 50 agents to 5. AskSpot resolves 65% of its 37,600 monthly chats with no human and answers Allegro questions too.
Read the story →
How Militaria resolves 72% of chats in 9 languages
Militaria.pl runs AskSpot across three domains in nine languages: 27% add-to-cart after a chat, 72% resolved with no human, 16,600 conversations a month.
Read the story →
How 4FIZJO resolves 61% of customer chats with AI
AskSpot's AI Chat resolves 61% of 4FIZJO's 21,360 monthly chats with no human across 6 markets; its Inbox Agent auto-resolves 25% of email tickets.
Read the story →
How Senetic turns 19.5% of technical IT chats into cart-adds
Senetic's AskSpot agent answers port, PoE and compatibility questions across seven European storefronts. 19.5% add to cart, 8.1% buy, 62% resolved by AI.
Read the story →
How Mamabrum serves customers in 20 languages with one team
How Mamabrum reduced customer service inquiries by 30% and serves customers in 20 languages across 20+ EU storefronts with AskSpot — chat and email answered from one catalogue.
Read the story →
How Moliv answers 65% of skincare questions with no human
Moliv's creams are built on live probiotics, so shoppers arrive with clinical briefs. The AskSpot agent answers 65% with no human, in three languages.
Read the story →
See every story
Every merchant we have measured — what shoppers asked, what the agent did, and the numbers it moved.
Browse them all →