Wholesale automation: orders, product data, price increases
Orders and reconciliation on retailers' Web-EDI, new product data, and price-increase requests: per-customer work, automated with AI and human review.
- Runs on your own PCs and servers
- Use Azure OpenAI, Amazon Bedrock, Google Vertex AI, or an in-house model
- Its engine is Descarty's open-source Dagu (GPL-3.0)

Common problems
Each morning, collecting orders from the Web-EDI portals, faxes, and emails of customers not on Ryutsu BMS
Re-entering new product data to fit each retailer's registration screens and templates
Price-increase requests and checks on receipt and payment differences multiply with every customer
Use cases in this industry
Enter an Excel list into a web system
For each row of an Excel list of orders, AI enters it into a web system and writes the receipt number back into the row. Only failed rows are redone.
Enter records into a desktop accounting app
For each row of an Excel list, AI enters the record in a desktop app. A person approves before it is saved, and the slip number goes back into the row.
Read Excel purchase orders from email and enter them
AI reads Excel purchase orders in each customer's own layout, checks them against the product master, and enters them. A person checks orders that do not match.
Take in orders from retailers' Web-EDI
Each morning, sign in to each retailer's Web-EDI, download orders, and convert them for the core system. Staff check unusual quantities before import.
Send shipping notices and reconcile receipts
Send shipping notices through each retailer's Web-EDI, then reconcile next morning's receipt data. AI suggests causes of quantity differences; staff confirm.
Match payment notices to receivables
Get payment notices from each retailer's Web-EDI and match them to receivables. AI suggests why amounts differ, such as rebates or returns; accounting confirms.
Turn supplier price increases into customer requests
Read suppliers' price-increase emails and list the items, dates, and margin impact. AI drafts each customer's revision request for the sales manager to approve.
Register new products on retailers' systems
AI reads your product data and spec sheets and fills each retailer's registration screen or Excel template. Staff check unclear fields and approve.
Revise and resubmit food spec sheets
When labeling rules or ingredients change, AI finds likely affected products. After QA confirms, it fills each customer's template for review and submission.
Every customer has its own channel and format
Wholesalers and consumer goods makers deal daily with supermarkets, drugstores, home centers, restaurant chains, and others: orders, shipments, receipts, invoices, and payments. Ryutsu BMS is the retail industry’s standard EDI in Japan. As of June 1, 2025, more than 21,600 wholesalers and makers had adopted it, according to the Ryutsu BMS council’s estimate. The council also notes that companies are moving over from fax and Web-EDI.
Phone and fax orders remain, though. METI estimates that in 2025, 43.9% of wholesale trade was e-commerce, meaning orders placed over computer networks. Orders by phone, fax, or email outside a fixed format are not counted. With customers not on Ryutsu BMS, orders also come through each retailer’s own web-based EDI portal, known in Japan as Web-EDI (Collecting orders from each customer’s Web-EDI in one list).
| Work | Screens and systems | What stays manual |
|---|---|---|
| Orders | Ryutsu BMS, each retailer’s Web-EDI, fax, email | Signing in to each Web-EDI to download data, and typing in orders that arrive by fax or email |
| Shipping notices and receipts | Each retailer’s Web-EDI, warehouse system | Sending shipping notices by each deadline, and finding out why received quantities differ |
| Payments and clearing | Each retailer’s Web-EDI, accounting system | Working out why payments fall short of invoices after discounts, rebates, and distribution center fees |
| New product data | Industry databases, each retailer’s registration screens, Excel templates | Re-entering the same data to fit each customer’s fields and formats |
| Food spec sheets | Spec-sheet services, customers’ own templates | Resubmitting to each customer whenever ingredients or labeling change |
| Price-increase requests | Sales management system, Excel, email | Estimating the impact on each customer and writing each request |
Customers connected through Ryutsu BMS exchange order-to-payment data directly with your core system. Industry databases carry product data, and services exist for exchanging food spec sheets in a common format. What gets automated is the share of customers and formats these don’t cover.
New product data is re-entered for each customer
METI says makers, wholesalers, and retailers enter product information individually into each company’s own forms, relying on what it calls a manual “bucket brigade.” It puts the work of handing over product information at 300,000 person-months a year, or 820,000 including shelf planning and e-commerce listings.
A study commissioned by METI found that although product data is gathered in industry databases such as Planet, FDB, and Inforex, the final hand-off to each retailer is still done individually, which weighs on makers, wholesalers, and retailers alike. Among the causes, it lists formats that differ by customer and wholesalers supplying product information to retailers on makers’ behalf.
The industry is moving toward shared data. GS1 Japan, Japan Inforex, Product Registry Service, and Planet are jointly building a cross-industry registry of product information. METI’s guidelines on exchanging product information, set in March 2026, aim for basic data on new products to be registered on a product information platform six weeks before the launch week. They also ask users of the data not to request it from each supplier individually, as a rule.
METI’s materials, however, say information should ideally be published 12 weeks before launch, when buyer meetings with retailers start, so those meetings begin before the registration target. Until every customer takes data from the platform, entry for each customer remains.
Food spec sheets are resubmitted whenever labeling changes
In the food trade, suppliers give customers product specification sheets covering ingredients, allergens, origin, and more. On April 1, 2026, the Consumer Affairs Agency revised Japan’s Food Labeling Standards to add cashew nuts to the allergens that must be labeled. It asked food businesses to recheck ingredients and production methods, with a two-year transition period. Products that use the ingredient need revised spec sheets, resubmitted to every customer that received one.
Services exist for exchanging spec sheets in a common format. Infomart’s BtoB Platform Spec Sheets supports a standard format developed under a Ministry of Agriculture, Forestry and Fisheries subsidy program, and Infomart says more than 10,000 companies use it. Where customers use such a service, resubmission happens there. For customers that specify their own templates, copying into each template remains.
Price-increase requests need grounds for each customer
Japan’s Small and Medium Enterprise Agency surveys small suppliers after its Price Negotiation Promotion Month each March and September. In the survey after March 2026, suppliers had passed on 54.2% of their cost increases to prices. By the buyer’s industry, the rate was 54.3% where the buyer was a retailer and 54.6% where it was a wholesaler. The agency says the overall trend is improving but still only halfway there.
Japan’s amended Subcontract Act, renamed the Toriteki Act and in force since January 1, 2026, bans deciding prices unilaterally without negotiating, such as ignoring a small supplier’s request to discuss prices or not giving the necessary explanation. It covers manufacturing consignments, transport consignments, and similar deals that meet capital or headcount thresholds. A retailer asking a supplier to manufacture goods to its specifications may be covered. The Japan Fair Trade Commission explains, however, that ordering a maker’s own branded goods is not a manufacturing consignment when the goods are general-purpose and the retailer has no changes made for itself.
Either way, a request needs grounds: which of raw materials, labor, and energy went up and by how much, and how that affects margin with each customer, worked out for every product and customer. An example that automates reading supplier notices, estimating the impact, and drafting requests is Turn supplier price increases into customer requests.
What to check before automating
- Whether a data link is enough: customers you can reach through Ryutsu BMS, an API, an industry database, or a spec-sheet service should be handled that way
- Whether automated screen operation is allowed: check the terms of each customer’s Web-EDI and registration screens
- How each site signs in: IDs and passwords, one-time passwords, and client certificates differ by customer. Sites that turn automation away with image challenges are not automated
- Whether it finishes before cut-offs: work back from order cut-offs and shipping-notice deadlines to set run times and who responds when a run stops
- Who decides: staff decide on unusual orders, how differences are settled, allergen labeling, and price negotiations
What Kitewell does
Descarty’s Kitewell is integration and automation middleware that connects systems through APIs, files, databases, and screens.
- Operates each customer’s screens: AI reads and operates Web-EDI and product registration screens. Operations that worked are remembered and repeated the same way on later runs
- Keeps going when screens change: when a remembered operation no longer fits the screen, AI reads the screen again, does the step, and remembers the new operation. If it still cannot proceed, the run stops, and AI looks into the cause and suggests the fix
- Handles sign-in hurdles: for sites that ask for a client certificate, it presents the certificate registered on the computer for that site. Saved browser sign-ins are reused by later runs, and when a one-time password is required, the run pauses until staff enter it
- Passwords stay out of AI’s hands: Web-EDI passwords are referenced by name and typed in without showing the value to the AI
- Runs on a schedule or on email: it starts at set times, such as weekdays at 8:00, or when email arrives. Mailboxes are checked at set intervals, and attachments such as order forms are saved
- Reads documents: AI steps read order forms, spec sheets, and payment notices as PDFs, Excel files, or images, and classify or draft text. The AI model is chosen per step
- People check where it matters: approval can be set on any step, such as before import or registration. Requests reach Slack, Microsoft Teams, and email, and sending one back restarts the run from the chosen step
- Runs from an Excel list: register a list of products or customers as a batch sheet, and it runs row by row and writes each result back to its row
- Feeds older systems: an Excel file can be converted to a Shift_JIS CSV for an older system’s import screen
- Alerts on failure: failures are sent to Slack, Microsoft Teams, email, and more
You can start from one piece of work, such as taking in orders from customers’ Web-EDI through registering them in the sales system and confirming them. A demo shows how it runs, a free connection test confirms Kitewell reaches your systems, and a one-month PoC measures the result against success criteria agreed up front, before you license it for production. See From evaluation to production for details.
Sources
- LOGISTICS TODAY: Ryutsu BMS adoption spreads to smaller firms, passing 21,600 companies (July 8, 2025; in Japanese; estimate by the Ryutsu BMS council)
- METI: FY2025 e-commerce market survey report (October 2026; in Japanese)
- METI: Making supply chains more efficient with a common product master (March 14, 2025; in Japanese)
- METI: Product information exchange: discussions so far and next steps (March 18, 2026; in Japanese)
- SIGMAXYZ: FY2024 survey of product information exchange in consumer goods distribution, report (March 2025; in Japanese; commissioned by METI)
- GS1 Japan (Distribution Systems Research Institute): GS1 Japan cross-industry registry (March 18, 2026; in Japanese; material for METI’s product information council)
- METI: Guidelines on exchanging product information (March 2026; in Japanese)
- Consumer Affairs Agency: Labeling of foods containing allergens (April 1, 2026; in Japanese)
- Infomart: BtoB Platform Spec Sheets (in Japanese)
- Small and Medium Enterprise Agency: Follow-up survey results for the March 2026 Price Negotiation Promotion Month (June 26, 2026; in Japanese)
- METI: Results of the follow-up survey for the March 2026 Price Negotiation Promotion Month (June 26, 2026; in Japanese)
- Japan Fair Trade Commission: From January 2026, the Subcontract Act becomes the Toriteki Act (August 2025; in Japanese)
- Japan Fair Trade Commission: Toriteki Act FAQ (in Japanese)
Frequently asked questions
- Is it used with customers connected through Ryutsu BMS or other EDI?
- No. Where data arrives through Ryutsu BMS or EDI, that is more reliable. Kitewell suits the parts with no data link, where people read a screen or document and type: each customer's Web-EDI and product registration screens, or order forms that arrive by email.
- May a tool operate customers' Web-EDI and product registration screens?
- It depends on each customer's terms. The PoC checks each customer and automates only what its terms allow. Sites that turn automation away with image challenges or similar are not automated.
- Does AI decide price negotiations or how differences are settled?
- No. AI reads price-increase notices, estimates the impact on each customer, drafts request letters, and suggests reasons for differences. Staff decide how much to pass on to each customer and how each difference is settled.
Use cases for any industry
Read invoices into the accounting system
AI reads incoming invoices, checks them against purchase orders, and records them in the accounting system. Invoices whose amounts differ wait for approval.
Create invoices and send them by email
Invoices are filled into your template from an Excel billing list and emailed once a person approves. AI first points out amounts far from last month's.
Research a list of prospects
For an Excel list of hundreds of companies, each one's business, locations, and size are collected from its website into its row, with a priority for sales.
Find missing clock-ins
Collect yesterday's clock-ins from the attendance system and, once HR approves, ask employees with missing entries to fix them in Teams.
Take in emailed invoices and draft the entries
Find unread email with an invoice attached, have AI read the amount, and record a draft ledger entry. Uncertain entries wait for accounting to confirm.
Review SaaS accounts
A SaaS account review matches each service's users to current staff. Collect each console's users, match them, and disable leavers only after approval.
Partner registration that starts on approval
A kintone approval starts the workflow by webhook. It looks up the tax registration number, AI matches it to the request, and a match enters the sales system.
Collect KPIs from SaaS tools into a weekly report
Collect figures from advertising, sales, support, and other SaaS tools, and produce a weekly report with week-on-week comparisons and likely causes of change.
Check supplier invoice registration numbers
Check your supplier master's registration numbers monthly through the NTA's Web-API, and list suppliers that were revoked, lapsed, or differ in name.
Prove Kitewell on your own systems
Try it free for 14 days on your own PC. To see it working first, ask for a 30-minute demo. There is also a one-month onboarding (¥500,000 before tax), where we build your first job together.
What the onboarding checks
- Which method connects each target system: API, file, database, or screen
- Whether Kitewell can work a system with no API through its screens, as a person would
- Whether AI relearns a step and processing continues when a screen changes
- Whether errors are handled as agreed: stop, notify, or retry
- Whether it runs in your environment, with the AI model you choose
- Whether processing time and AI usage cost fit daily operation
How the onboarding runs
- Week 1Requirements and success criteria
- Weeks 2 to 3Build
- Week 4Verify on production screens
- Final dayReport
One-month onboarding
- Length
- One month
- Cost
- ¥500,000 (tax excluded). Your staff build, with the developer beside them
- If you go on to a license
- The onboarding fee is credited to the first year's contract
- Without the onboarding
- At Team scale, start on Team monthly (¥14,800 per computer a month, tax excluded), and cancel in any month
You can try it first on your own PC, in the free 14-day trial.