The AI agent is the heart of Billabex. It acts as a dedicated virtual collaborator for tracking and following up on your invoices. Its goal: maximize your collection success while preserving the customer relationship.
What does the AI agent actually do?
As soon as an invoice is imported or synchronized, the AI agent springs into action. It:
1. Analyzes each invoice and credit note: amount, due date, payment status, customer history
2. Creates an intelligent follow-up plan: tailored to the customer profile, appropriate tone, and context (aging, dispute, promise...)
3. Communicates automatically: via email and letter based on available data (soon, via SMS or voice message)
4. Tracks responses: and adapts its behavior in case of payment promise, dispute, special request, etc.
5. Collaborates with your team: reaching out only when human intervention is needed (via a task)
Invoice lifecycle
The follow-up plan is adaptive. At each stage, the AI agent:
- checks the payment status
- reschedules follow-ups based on aging
- adjusts the tone (courteous → professional → firm)
- stops actions when the case is closed or manual intervention occurs
Each communication is scheduled on business days, adjusted according to the customer account's country.
Email deliverability tracking
The AI agent includes an automatic deliverability tracking mechanism to ensure messages reach their intended recipient.
When an email cannot be delivered (e.g., non-existent address, hard bounce), the agent:
- automatically marks the contact as invalid in the account record
- creates a task to request a valid contact from the team
- suspends follow-ups to that contact until the information is updated
- tracks the request from the Tasks module
This logic protects your sending reputation and prevents compromising the effectiveness of future follow-ups.Communication language
The AI agent automatically adapts the language used in its messages, whether for contacts (customer accounts) or members of your organization.
The language is determined in the following priority order:
1. The language set at the contact level of the account
2. If none is set, the language is automatically inferred from the country specified in the account
This ensures each customer is addressed in their own language, with no additional configuration required.
For your organization members
When the AI agent communicates with organization members (e.g., task notifications), it follows these rules:
1. If a member has a preferred language set in their settings, it is used.
2. If no language is set, the default language is English.
When multiple users are recipients of the same communication:
- If all share the same preferred language → that language is used.
- If multiple languages are set:
- A majority language is used if possible
- If no clear majority → communication defaults to English
When the AI agent interacts with you
The AI agent operates autonomously but reaches out to you in the following cases:
- Ambiguous customer response or dispute
- Missing or inconsistent data (e.g., missing contact)
- Detected roadblock (explicit refusal, invalid email, callback request, no response after 4 follow-ups)
- Request from a team member
A task is then automatically created to ensure human follow-up.
Does the agent learn?
The AI agent is not self-learning but relies on an intelligent contextual engine. It:
- adapts its follow-ups based on past interactions
- avoids following up with a customer already in an active conversation
- respects your organizational and language preferences
- incorporates specific case management rules such as email address validity or due dates
Regular updates ensure continuous improvement without affecting its operational stability.