A customer reports a stove showing error E4. Your after-sales team has already fixed this problem, maybe twice. But who, when, and how? The answer sleeps in a closed ticket nobody can find. Odoo 20 compares every new ticket with the old ones, suggests the solution citing its sources and drafts the email to the customer. Here is how to enable it, what to know beforehand, and who it is worth it for.
1. The problem: an after-sales service with no memory
In most after-sales departments, knowledge lives in the heads of the longest-serving technicians. The tickets are neatly filed, but nobody rereads closed tickets: you don't know what to search for, or with which words the customer at the time described the fault. The result: you start from scratch, the customer waits, and the day the experienced technician is on leave, the knowledge is on leave too.
💡 Key takeaway
Your after-sales team's knowledge already exists in Odoo. What was missing was a way to find it without knowing the right keywords.
2. What Odoo 20 does
The feature comes from the standard ai_helpdesk module (Odoo Enterprise), installed automatically with the Helpdesk and AI apps. It works in three steps.
- Find. Each ticket is compared with the others by the meaning of its title and description, not by keywords. A smart button shows the number of close tickets ("Similar") and, among them, the number of closed tickets ("Closed"). These labels are in English: the module is barely translated in V20.
- Suggest. In the ticket's AI panel, the "Find a solution based on similar tickets" button (also in English) asks the AI to read how those tickets were resolved. It summarises the solution in a few steps and cites the tickets they come from. If no close ticket was resolved, it says so and reports what has already been tried.
- Reply. The AI then offers to draft the email to the customer. It only writes it if you accept, and you proofread it before clicking "Send".
The instructions shipped by Odoo are explicit: the AI addresses the operator, not the customer, and never suggests a solution that does not come from the tickets, unless you ask it to, in which case it must flag this.
💡 Key takeaway
Every suggested step points to a ticket you can open. You check the source instead of taking the AI's word for it.
3. Enabling it on a team
The setting is made per helpdesk team, in the team configuration, "AI" section. All the labels there are in English:
- "Look for Similar Tickets": the checkbox that enables the feature;
- "Include Tickets from": how far back to go, from "Last Week" to "Forever" (default value);
- "Agent": the AI agent used, filled in automatically;
- "Similarity Threshold": the degree of resemblance required, 0.45 by default.
On activation, Odoo starts indexing the tickets in the background. Allow some time before the "Similar" counters fill up on the history.
💡 Key takeaway
Keep the default settings at first, then adjust the threshold after a few weeks: too many "similar" tickets that have nothing in common, raise it; no results, lower it.
4. And if someone has to go on site
In Odoo 20, on-site interventions have joined the Planning app. If the "Field Service" option is ticked on the team, the ticket carries a "Plan Intervention" button: it opens the planning with the ticket and the customer's address already filled in. All that remains is to book the technician's slot.
5. Limits and prerequisites
- The quality of the history decides everything. The AI reads how tickets were closed: closing messages, operator actions, customer confirmation. A ticket closed without a word teaches nobody anything. This is often the first project.
- Odoo AI credits. Searching for a solution and drafting the email consume IAP credits. In Odoo 20, there is no personal AI key any more: everything goes through Odoo's service.
- Odoo Enterprise, with the Helpdesk app; plus Planning for interventions.
- A partly English interface until the module's translation catches up with the version. Plan for this in your team's training.
Who is it for? SMEs with an after-sales or support service and some history: installers, stoves and heating, machines, IT. The richer your base of resolved tickets, the more the feature pays off.
⚔️ The Match: by hand vs Odoo 20
| Step | By hand | Odoo 20 |
|---|---|---|
| Finding a similar case | Keyword search, or a colleague's memory | "Similar" counter, closed tickets included |
| Extracting the solution | Rereading the exchanges ticket by ticket | Steps summarised, source tickets cited |
| Replying to the customer | Email written from scratch | Draft suggested, proofread before sending |
| Scheduling the visit | Re-entering customer and address in the planning | Planning opened from the ticket, prefilled |
Behaviour observed in the Odoo 20 Enterprise code (ai_helpdesk, helpdesk_planning_field_service). The real gain depends on the quality of your closed tickets.
The Verdict: your after-sales service finally has a memory, if you have written it down
Odoo 20 turns your ticket history into a base of solutions, with sources to back it up. The condition is a history that tells how each problem was fixed.
Do your closed tickets say how the fault was resolved? We check this with you during an
audit of your processes,
and we train your team in good closing habits with an
Odoo training.
A customer reports a stove showing error E4. Your after-sales team has already fixed this problem, maybe twice. But who, when, and how? The answer sleeps in a closed ticket nobody can find. Odoo 20 compares every new ticket with the old ones, suggests the solution citing its sources and drafts the email to the customer. Here is how to enable it, what to know beforehand, and who it is worth it for.
1. The problem: an after-sales service with no memory
In most after-sales departments, knowledge lives in the heads of the longest-serving technicians. The tickets are neatly filed, but nobody rereads closed tickets: you don't know what to search for, or with which words the customer at the time described the fault. The result: you start from scratch, the customer waits, and the day the experienced technician is on leave, the knowledge is on leave too.
💡 Key takeaway
Your after-sales team's knowledge already exists in Odoo. What was missing was a way to find it without knowing the right keywords.
2. What Odoo 20 does
The feature comes from the standard ai_helpdesk module (Odoo Enterprise), installed automatically with the Helpdesk and AI apps. It works in three steps.
- Find. Each ticket is compared with the others by the meaning of its title and description, not by keywords. A smart button shows the number of close tickets ("Similar") and, among them, the number of closed tickets ("Closed"). These labels are in English: the module is barely translated in V20.
- Suggest. In the ticket's AI panel, the "Find a solution based on similar tickets" button (also in English) asks the AI to read how those tickets were resolved. It summarises the solution in a few steps and cites the tickets they come from. If no close ticket was resolved, it says so and reports what has already been tried.
- Reply. The AI then offers to draft the email to the customer. It only writes it if you accept, and you proofread it before clicking "Send".
The instructions shipped by Odoo are explicit: the AI addresses the operator, not the customer, and never suggests a solution that does not come from the tickets, unless you ask it to, in which case it must flag this.
💡 Key takeaway
Every suggested step points to a ticket you can open. You check the source instead of taking the AI's word for it.
3. Enabling it on a team
The setting is made per helpdesk team, in the team configuration, "AI" section. All the labels there are in English:
- "Look for Similar Tickets": the checkbox that enables the feature;
- "Include Tickets from": how far back to go, from "Last Week" to "Forever" (default value);
- "Agent": the AI agent used, filled in automatically;
- "Similarity Threshold": the degree of resemblance required, 0.45 by default.
On activation, Odoo starts indexing the tickets in the background. Allow some time before the "Similar" counters fill up on the history.
💡 Key takeaway
Keep the default settings at first, then adjust the threshold after a few weeks: too many "similar" tickets that have nothing in common, raise it; no results, lower it.
4. And if someone has to go on site
In Odoo 20, on-site interventions have joined the Planning app. If the "Field Service" option is ticked on the team, the ticket carries a "Plan Intervention" button: it opens the planning with the ticket and the customer's address already filled in. All that remains is to book the technician's slot.
5. Limits and prerequisites
- The quality of the history decides everything. The AI reads how tickets were closed: closing messages, operator actions, customer confirmation. A ticket closed without a word teaches nobody anything. This is often the first project.
- Odoo AI credits. Searching for a solution and drafting the email consume IAP credits. In Odoo 20, there is no personal AI key any more: everything goes through Odoo's service.
- Odoo Enterprise, with the Helpdesk app; plus Planning for interventions.
- A partly English interface until the module's translation catches up with the version. Plan for this in your team's training.
Who is it for? SMEs with an after-sales or support service and some history: installers, stoves and heating, machines, IT. The richer your base of resolved tickets, the more the feature pays off.
⚔️ The Match: by hand vs Odoo 20
| Step | By hand | Odoo 20 |
|---|---|---|
| Finding a similar case | Keyword search, or a colleague's memory | "Similar" counter, closed tickets included |
| Extracting the solution | Rereading the exchanges ticket by ticket | Steps summarised, source tickets cited |
| Replying to the customer | Email written from scratch | Draft suggested, proofread before sending |
| Scheduling the visit | Re-entering customer and address in the planning | Planning opened from the ticket, prefilled |
Behaviour observed in the Odoo 20 Enterprise code (ai_helpdesk, helpdesk_planning_field_service). The real gain depends on the quality of your closed tickets.
The Verdict: your after-sales service finally has a memory, if you have written it down
Odoo 20 turns your ticket history into a base of solutions, with sources to back it up. The condition is a history that tells how each problem was fixed.
Do your closed tickets say how the fault was resolved? We check this with you during an
audit of your processes,
and we train your team in good closing habits with an
Odoo training.