AGENTES ADI · COLLECTIONS
Negotiates, closes and logs the commitment
Outbound voice that runs the negotiation inside your policy and writes the agreement to the CRM with no human typing.
The operational jump
From dialler to logged agreement
Both models start from the same mailing. What changes is where the conversation ends.
Robot dialler
Coru collections agent
Results in production
Impact that is measured
Aggregated platform results, measured in production. Not attributed to any specific account.
How it runs
From mailing to CRM, in one call
Load and segmentation
Debtor data arrives by CSV, API or console, with a defined time window and retry rule.
Human or voicemail
The agent tells the two apart before starting the conversation. Voicemail does not consume an attempt.
Deterministic validation
Confirmation before any amount is mentioned. Verification runs outside the model.
Correctly pronounced amounts
Principal, interest and due date read from a dedicated dictionary, as a trained operator would read them.
Free speech, fixed policy
Discounts and instalments within the limits you define. The debtor interrupts without breaking the flow.
Logged in the CRM
Date, amount and terms extracted from the transcript and written. No human typing.
Capabilities
Negotiating without pushing
Configurable discount policy
Write-off and instalment bands per portfolio profile, applied at conversation time.
Commitment extraction
Fields defined in Data Collection, extracted from the full transcript at close.
Cross-channel resumption
The debtor misses the call and replies on WhatsApp. The context comes along.
Collections metrics
Payment promises, agreement rate, committed amount and D+1 fulfilment.
Strict topical focus
The agent does not offer opinions outside the negotiation. Out of scope, it routes to an operator.
Escalation with the full thread
The operator arrives with the complete conversation and the data already collected.
Feasibility
Connects to what you already have
SIP trunk, CRM and collections systems connected through a documented interface. No core changes.
Onboarding in 1 to 3 weeks. With an existing SIP trunk, the voice channel goes live in 2 to 5 days.
Inside the rules
Collecting without leaving policy
Guardrails in code
Deterministic filters before and after the model. Prompt injection and PII exfiltration blocked at the edge.
Encrypted call
SIP TLS 5061 with SRTP. Voice travels encrypted, including the telephony layer.
Traceability
Every interaction reconstructable: what the agent said, on which version, under which controls.
Closed catalogue
The agent executes only actions registered as enums validated outside the model.
The full cycle
One engine, five agents
One compliance approval covers all five stages. You start with the one that hurts most and expand with no new project.
Origination
Captures, pre-qualifies and converts in free speech, on web or voice.
View →Onboarding
Sign-up and activation with deterministic identity validation.
View →Transactions
Voice-authenticated in-app operations, in seconds.
View →Support
Resolves on first contact, with context across all three channels.
View →FAQ
Frequently asked questions
Can the agent offer a discount without approval?
No. Discount and instalment bands are defined by you and applied as policy. The agent negotiates inside them and has no path to exceed the limit — the action catalogue is closed, with enums validated outside the model.
How does the commitment reach our system?
Through a webhook signed with HMAC-SHA256, at the close of the conversation. The fields are defined beforehand in Data Collection, with a type and an instruction per field, and extracted from the full transcript regardless of the order the information appeared in.
Do we need to change telephony providers?
No. You point your existing SIP trunk at the provided endpoint, over TLS. For outbound, the platform dials against your trunk or operates its own. With an existing trunk, the channel goes live in 2 to 5 days.
What happens when the debtor asks for a human?
Escalation follows the rule you define. The operator receives the complete conversation and the data already collected, and arrives reading the case instead of asking again.
How does the Risk team audit the calls?
Every interaction is recorded and reconstructable, with the agent version and the controls active at the time. Supervision coverage reaches 98% of interactions. Audit logs contain no personal data — PII is tokenised before any call to a model provider.