Where it started
Group NDL operates on two fronts: selling stainless steel kegs to Canadian craft breweries with factory sourcing in China, and business development between Canadian manufacturing SMEs and Asian industrial suppliers. Two markets, two languages, two pricing logics.
Every quote meant redoing the exchange rate by hand, applying the freight markup, picking the right commission rate for the region, filling in an internal purchase order, then writing to a carrier. Alongside that: a high volume of bilingual correspondence with clients and suppliers, a web presence to build from nothing, and a monthly commission calculation rebuilt in a spreadsheet each time.
None of it is hard. It is repetitive, time-consuming, and at that volume every manual step is a chance to get a number wrong on its way to a client.
The challenge
Tool up a complete sales role without changing the business rules it lives on: same prices, same margins, same templates, same suppliers, but with no manual recalculation and no rebuilding of documents.
What we built
The first front covers the keg sales cycle. The business rules were encoded once: currency conversion off the official Bank of Canada rate, freight markup, a commission rate that varies by delivery region, tax logic depending on which entity invoices. The quote comes out in DOCX and PDF, laid out to a standard template, ready to send.
The same system then chains the purchase order to the factory, prefills the Excel template the supplier expects, and prepares the email to the right carrier for the destination. A change in business rules propagates everywhere: since the local warehouse closed, the system systematically prompts to confirm the lead time with the client before any pricing work starts.
The second front is correspondence. A directory of clients, project managers and part references feeds the drafting: French to clients, English to suppliers and factories, with the file's context already in place. Customs holds, container tracking, air versus sea calls, non-conformities, blanket orders: each recurring case has its own template.
The third front is digital presence. The multi-division website was designed and built with AI assistance plus custom code, put online, then wired to advertising conversion tracking, with two campaigns launched in parallel on the French-speaking and English-speaking markets.
The fourth front is monthly commission invoicing. Raw reports are read, freight lines excluded from the calculation by a precise rule, conversion applied, and the output delivered as two files: the PDF invoice and a detailed spreadsheet recap for traceability.
Then the infrastructure, because nothing could be plugged into what was there: seventeen domains brought up to date, twenty years of email moved to Microsoft 365 without a minute of downtime, three security flaws spotted along the way, outside the mandate, reported and fixed. Only then, a secure link that reads the mailbox and pulls the business data out of it.
No internal rate, margin or supplier price is shown here. The calculation rules live in the system, not in the documents sent to the client.
How it unfolded
Sales cycle
Quote, price grid, factory purchase order, carrier email.
Correspondence
Client and supplier directory, bilingual drafting by recipient.
Web presence
Multi-division site, conversion tracking, two ad campaigns.
Invoicing
Report parsing, commission calculation, invoice and recap.
The results
Manual calculation errors are gone from the path: exchange rate, freight markup, commission rate and taxes are applied by the system from a single source of rules. Two quotes prepared two weeks apart use exactly the same logic.
Monthly invoicing, which meant rebuilding a spreadsheet line by line, now generates in one pass: PDF invoice and detailed recap, ready to send together.
The digital presence was built entirely in-house: multi-division site online, working conversion tracking, two campaigns live across two language markets.
Close to 240,000 emails were made readable by AI. The client, part and price databases were rebuilt from that history, and the CRM now updates itself with every email received.
The inbox is triaged every day and reply drafts are prepared ahead. Nothing goes out until a person sends it. Twenty years of email were migrated without a minute of service interruption.
The inbound request pipeline is designed and in development. It targets bringing lead handling from about fifteen minutes down to two or three: intake, classification by division, language detection, contextualized draft reply.
Handling an inbound request
The plan, delivered and ahead
What makes this case different from the others is that it was not built in one go. Each front came out of a task that recurred too often, and got tooled up when the volume justified it. The result is a coherent system, but it was assembled piece by piece, in production.
It is also the clearest demonstration of the method: start with the most repetitive task, make it reliable, then move to the next one reusing what already exists.
Cover the full commercial cycle, from price request to commission invoiced
- Quotes generated and laid outDocuments produced in DOCX and PDF on a fixed template, from pricing rules encoded once.
- Price grids computed automaticallyCurrency conversion on the official rate, freight markup, commission rate by delivery region. Internal rates stay inside the system.
- Purchase orders prefilled for the factoryThe template the supplier expects is filled automatically, with the tax logic specific to the invoicing entity.
- Carrier emails by destinationThe right carrier is picked by region or shipping mode, with the same pickup address and the same request format.
- Bilingual correspondence tooled upA directory of clients, project managers and part references, with templates for the recurring cases: customs, container tracking, non-conformity, blanket order.
- Monthly commission invoicingRaw reports parsed, freight lines excluded, conversion applied, output as a PDF invoice and a detailed recap.
Close the loop at intake, on requests coming from the site
- Automatic intake of site requestsEvery form submission is caught at send time, with no hand-watched inbox in the middle.
- Classification by divisionThe request is routed to the right division by its content, rather than sorted on reading.
- Language detection on the requesterThe reply goes out in the language of the request, without the question coming up.
- Contextualized draft replyThe rep receives a reply already written with the file's context. They approve, adjust and send.
Three custom systems and a site, around one role
Each piece handles one specific task and reuses the others' rules. That sharing is what keeps a price or policy change from having to be applied in five different places.
From the price request to the payment, with no break
The path is tooled end to end. Each step produces the document the next one expects, in the format the recipient already looks for.
The next piece is already mapped: inbound requests from the site, classified and pre-drafted before they even land in an inbox. The rest of the path is already waiting for them.
What the company gets out of it
One source for every pricing rule
Currency, freight, commission and taxes are computed in one place. Change a rule and it changes everywhere at once.
Correspondence goes out in the right language first time
French to the client, English to the factory, with the file's context already loaded. The rep reviews instead of composing.
A digital presence built without an agency
Multi-division site online, conversion tracking configured at the source-code level, campaigns live on two markets.
Invoicing that no longer gets rebuilt each month
Raw reports go in, the invoice and the recap come out. Traceability sits in the file, not in someone's memory.
Twenty years of email turned into a database
Close to 240,000 emails made readable by AI. The client, part and price databases were rebuilt from that history, and the CRM updates itself with every email received.
Infrastructure brought up to date before anything was plugged into it
Seventeen domains brought up to date, email migrated to Microsoft 365, three security flaws spotted outside the mandate, reported and fixed.
The thinking behind this mandate is laid out here: 5 signs your company is ready for AI automation.