Structured order lines
Each order broken into SKU, quantity and unit, ready for the ERP.
Customer orders from email and chat into your ERP
For businesses that take orders as PDFs, spreadsheets, photos or plain text over email, Telegram and WhatsApp. Each order is read, matched to your catalogue, checked against stock and entered into your system without retyping.
Intake
Map
Validate
Execute
Confirm
What this fixes
The situations we hear about in the first call.
Orders arrive in every format
Some customers send a neat Excel sheet, others a photo of a handwritten list or a paragraph of free text in a chat. Someone has to read each one and copy it in by hand.
Product names never match the catalogue
Customers use their own names, abbreviations and old product codes. Staff have to guess which SKU they mean, and mistakes only surface after the order ships.
Stock gets checked too late
An order is confirmed before anyone checks whether the item is actually in stock, so the customer is told about a shortage days later.
Confirmations are slow or skipped
Customers wait hours for someone free to acknowledge the order, or get no confirmation at all until the invoice shows up.
How it works
Five stages, the same for every stream we build. Anything that fails validation goes to a person, in the tool you already use.
- Intake
Orders are picked up from email inboxes, Telegram and WhatsApp as they arrive, whatever the format: PDF, Excel, a photo, or a plain-text message.
- Map
Product names and codes in the order are matched to your SKU catalogue using fuzzy matching. Anything that cannot be matched confidently is listed with its closest alternatives.
- Validate
Quantities and matched SKUs are checked against current stock in MoySklad or 1C before anything is created, so shortages are caught before confirmation.
- Execute
A clean order is created directly in your inventory or ERP system, with line items, quantities and the original message attached for reference.
- Confirm
A confirmation is sent back to the customer on the same channel they used, and any order with an unmatched item or stock issue is routed to a person instead of going through automatically.
What you get
Closest-alternatives list
For any unmatched product name, a short list of likely catalogue matches for a person to pick from.
Stock check result
Each line marked against live stock in MoySklad or 1C before the order is created.
Customer confirmation message
An automatic reply on the same channel confirming what was received and entered.
Review queue for exceptions
Orders with unmatched items, stock shortfalls or unclear text sent to a review queue in the tool you already use.
Order log
A record of every order processed, matched and any corrections made, for later checking.
What the pilot measures
We do not quote accuracy percentages before we have seen your documents. These three numbers are agreed at the audit and reported at the end of the pilot.
Straight-through rate
Share of orders created in the ERP with no manual correction, tracked across the pilot stream.
Minutes per order
Time from an order arriving to it being entered in the system, compared with the current manual process.
Exceptions per 100
How many orders per 100 need a human decision on matching or stock, out of the pilot volume.
Systems we connect for this
MoySklad1CZoho CRMPipedriveHubSpotGoogle Sheets
Processed in the EU. Original files and extracted data stay in your own systems; a data-processing agreement is available on request.
From audit to autopilot
Pilot scope for this stream: One order channel (for example, one Telegram group or one order inbox) and its matching against your live catalogue and stock.
Step 1
Audit
We map one document or enquiry stream end-to-end, count volumes and name the systems it touches. You get a written scope and a fixed pilot price.
Free · 3 working daysStep 2
Pilot
One stream goes live on your real documents. Acceptance is measured, not felt: straight-through rate, minutes per document, exceptions.
About two weeks · fixed priceStep 3
Rollout
The remaining streams follow the proven pattern. Monthly report, support, and changes as your systems change.
Stream by streamBuilt before
Case · Dairy distributor, Kazakhstan
111 dairy SKUs online with stock levels that are never more than 15 minutes old
111 SKUs, two languages, stock refreshed every 15 minutes from MoySklad, 222 product pages and a shopping feed — for a dairy distributor in Kazakhstan.
15 minmaximum age of stock levels on the site
Frequently asked questions
What does this cost?
There is no set price, because it depends on your order volume, channels and how your ERP is set up. The free audit tells you what is realistic; the pilot price is fixed and agreed before any work starts.
How accurate is the product matching?
Fuzzy matching handles abbreviations, typos and old codes well, but it is not perfect. Anything below a confidence threshold is shown as a shortlist of closest alternatives rather than guessed, and a person makes the final call.
What happens when an order cannot be matched or stock is short?
It goes to a review queue in whichever tool your team already uses, such as Telegram, email or a CRM task, rather than being created incorrectly or silently dropped.
Where does our order and customer data go?
Everything is processed in the EU. The original messages, files and extracted data stay in your own systems. A data-processing agreement is available on request.
Does this work with our existing ERP and inventory system?
We build against the system you already use, such as MoySklad or 1C, rather than asking you to move to a new platform. If you use something else, the audit will confirm what is possible.
Start with the free audit
Three working days, one stream mapped end-to-end, a fixed pilot price on paper. No commitment.
Get a free automation audit