Memory - remember customers and organisations across tickets
Memory allows the AI assistant to remember specific information about a ticket, customer, organisation or brand - and automatically use it the next time it encounters the same entity.
What is memory?
A memory is a short note attached to a particular entity:
| Ticket | This ticket only | "The customer is very frustrated - be extra caring" |
| Customer | All tickets from this person | "Pia Benne prefers short and precise responses" |
| Organisation | All tickets from this company | "ToyAcademy is B2B - always use a formal form of address" |
| Brand | All tickets for this brand | "Brand X only responds in English" |
Save a memory
Write in the chat field:
- "Remember that this customer is a VIP"
- "Save: the organisation always requires a receipt by email"
- "Remember for this ticket: the customer has already received compensation"
The AI assistant asks about the scope (ticket / customer / org / brand) and confirms.
The AI assistant suggests on its own
When the AI assistant learns something new about a customer or organisation, it can ask (in the agent's language):
"Would you like me to remember this for this customer in future?"
Auto-inject: Memory is shown automatically
When you open a ticket, the AI assistant automatically retrieves all relevant memories and shows them at the top of the context:
Memory from earlier interactions
[Ticket #123] The customer is frustrated about slow delivery
[Customer: Pia Benne] Prefers short responses
[Org: ToyAcademy] B2B customer - always use a formal tone
You do not have to do anything - memories are used automatically in the draft.
Retrieve saved memory manually
Write in the chat field:
- "What do we know about this customer?" - retrieves customer-level memory
- "Is there anything saved about this organisation?"
- "recall_memory" directly to the AI
The difference between memory and rules
| Memory | Rules | |
| Applies to | One specific entity (ticket, customer, org, brand) | All tickets in general |
| Example | "Pia prefers short responses" | "We always use the informal 'du' form" |
| Created via | Agents from the chat | Agents from the chat / admin |
Use memory for person- or context-specific knowledge. Use rules for general policies that apply to everyone.