A customer sends a WhatsApp message asking about products.<\/p>\n
The workflow reads it, works out what they want, finds those products in Odoo, creates a quotation, and replies.<\/strong><\/p>\n What that replaces:<\/strong> a salesperson reading the message, opening Odoo, searching for each product, building the quotation, and typing a reply. Several minutes, every time.<\/p>\n The input is free text written by a person.<\/strong><\/p>\n Not a form. Not a spreadsheet. Somebody typing on a phone.<\/p>\n “Do you have the 40mm brackets, need about 200, and the mounting kits too”<\/p>\n No rule handles that.<\/strong> The product name is approximate, the quantity is buried in a sentence, and the second item has no quantity at all.<\/p>\n This is exactly where a language model earns its place<\/strong> \u2014 extracting structure from text that varies without limit.<\/p>\n FIGURE 1: FROM A MESSAGE TO A QUOTATION<\/p>\n Message arrives<\/p>\n AI extracts<\/p>\n Odoo searched<\/p>\n Quotation drafted<\/p>\n The message goes to an AI model with instructions: pull out the products mentioned and the quantities.<\/p>\n The output must be structured<\/strong>, not prose. A list of items, each with a name and a quantity \u2014 something the next node can work with.<\/p>\n Two things that improve this a great deal:<\/strong><\/p>\n Give it your product vocabulary.<\/strong> A model that knows you sell brackets, kits and fixings interprets an ambiguous message better than one guessing from nothing.<\/p>\n Ask it to say when it is unsure.<\/strong> A model told to flag ambiguity is far more useful than one that always produces a confident answer.<\/p>\n The extracted text is not a product.<\/strong> “40mm brackets” has to become a specific Odoo product.<\/p>\n Search Odoo<\/strong> for each item.<\/p>\n Three outcomes, and each needs a decision:<\/strong><\/p>\n One clear match.<\/strong> Use it.<\/p>\n Several possible matches.<\/strong> Do not guess. Either ask the customer which they meant, or send it to a person.<\/p>\n No match.<\/strong> Tell the customer you could not find it, or hand to a salesperson.<\/p>\n This step is where most of the risk sits.<\/strong> A model can produce a plausible product name that does not exist, and quietly matching it to the nearest thing is how a customer receives a quotation for the wrong item.<\/p>\n Fail visibly rather than guessing.<\/strong><\/p>\n The same pattern as every Odoo integration:<\/strong> search first, create only if not found.<\/p>\n Match on phone number<\/strong> \u2014 you have it, since they messaged you.<\/p>\n Which means phone numbers must be stored consistently in Odoo.<\/strong> Country code included, no spaces or punctuation. Inconsistent formatting is the most common cause of a customer not being found and a duplicate being created.<\/p>\n With a matched customer and matched products, the quotation is straightforward \u2014 Odoo applies the customer’s price list, payment terms and taxes.<\/p>\n One important decision: what state does it arrive in?<\/strong><\/p>\n Draft, for a salesperson to review.<\/strong> Safer, and it keeps a person on anything commercial.<\/p>\n Confirmed automatically.<\/strong> Faster, and it commits you to prices a model helped assemble.<\/p>\n Draft is the right default.<\/strong> The workflow removes the assembly work \u2014 which is most of the time \u2014 and leaves the decision with a person.<\/p>\n FIGURE 2: WHERE THE HUMAN STAYS<\/p>\n The workflow does<\/p>\n A person does<\/p>\n A confirmation back to the customer on WhatsApp.<\/p>\n Because they messaged you, a service window is open<\/strong> \u2014 you can reply freely for a period without needing an approved template.<\/p>\n What a good reply says:<\/strong><\/p>\n What you understood them to want. This matters<\/strong> \u2014 it gives them a chance to correct a misreading before anything goes further.<\/p>\n What happens next, and roughly when.<\/p>\n What it should not do:<\/strong> state prices or commitments the model assembled without a person checking them.<\/p>\n Five things.<\/p>\n Guessing at ambiguous products.<\/strong> The most damaging. A confident wrong match produces a quotation for something the customer did not ask for.<\/p>\n Confirming automatically.<\/strong> A quotation is a commercial document. Keep a person on it.<\/p>\n Inconsistent phone numbers.<\/strong> Customers not matched, duplicates created.<\/p>\n No route to a person.<\/strong> A customer whose request the workflow cannot handle needs a human, quickly.<\/p>\n No logging.<\/strong> When a quotation comes out wrong, you need to see the original message, what the model extracted, and what it matched to.<\/p>\n That last one is what makes this debuggable.<\/strong> Without it, an AI step is a black box.<\/p>\n FIGURE 3: WHAT MAKES THIS SAFE<\/p>\n Fail visibly on ambiguity<\/p>\n Draft, not confirmed<\/p>\n Consistent phone numbers<\/p>\n Log input, extraction and match<\/p>\n The saving is the assembly, not the decision.<\/strong><\/p>\n Reading the message, searching for products, building the quotation \u2014 that is most of the time and all of the tedium.<\/p>\n Checking it and sending it takes a salesperson under a minute.<\/p>\n So the realistic outcome is not “quotations without people”.<\/strong> It is quotations that arrive ready to review, instead of taking ten minutes to build.<\/p>\n For a business handling many enquiries, that is a real change in capacity.<\/strong><\/p>\n Four questions.<\/p>\n Do enough enquiries arrive this way<\/strong> to justify it? A handful a week does not.<\/p>\n Is your product data clean enough to match against?<\/strong> Inconsistent names and missing references make matching unreliable.<\/p>\n Are phone numbers stored consistently?<\/strong> The customer match depends on it.<\/p>\n Who handles what the workflow cannot?<\/strong> There must be a person, and a fast route to them.<\/p>\n The workflow reads a message, extracts what was asked for, matches it to real Odoo products, and drafts a quotation.<\/p>\n The AI step is justified because the input is free text<\/strong> \u2014 no rule handles how people actually write.<\/p>\n Never guess at an ambiguous product.<\/strong> Fail visibly, ask the customer, or hand to a person.<\/p>\n Draft, not confirmed.<\/strong> The saving is the assembly; the decision stays human.<\/p>\n Enquiries arriving faster than quotations go out?<\/p>\n Get in touch. We build WhatsApp-to-Odoo quotation workflows that assemble the draft and leave the commercial decision with your team.<\/p>\n<\/div>","protected":false},"excerpt":{"rendered":" What it does A customer sends a WhatsApp message asking about products. The workflow reads it, works out what they want, finds those products in Odoo, creates a quotation, and replies. What that replaces: a salesperson reading the message, opening Odoo, searching for each product, building the quotation, and typing a reply. Several minutes, every […]<\/p>\n","protected":false},"author":3,"featured_media":9281,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[28],"tags":[],"class_list":["post-9280","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-n8n"],"yoast_head":"\nWhy this is harder than a normal import<\/h2>\n
\n
\n
\n
\n
Step 1 \u2014 Extract the request<\/h2>\n
Step 2 \u2014 Match to real products<\/h2>\n
Step 3 \u2014 Find or create the customer<\/h2>\n
Step 4 \u2014 Create the quotation<\/h2>\n
\n
\n
Step 5 \u2014 Reply<\/h2>\n
Where it goes wrong<\/h2>\n
\n
\n
\n
\n
What it is worth<\/h2>\n
Before building one<\/h2>\n
The short version<\/h2>\n