Case Study of FIBE

Recovering Stalled Applications Before Intent Becomes Lost Revenue.

Fibe partnered with Fundamento to recover borrowers who had received a credit limit but stalled before disbursal, using AI Voice Agents to identify the point of friction, guide customers through completion and escalate complex cases when required.

Written By

Team Fundamento

Published

Helping Digital Borrowers Move From Credit Intent To Disbursal

Fibe is a digital lending business serving customers through personal loans and BNPL products.

As digital lending journeys became increasingly dependent on customers completing multiple online steps, stalled applications created a clear opportunity to recover existing borrower intent rather than acquire new demand.

Credit Approval Was Not The End Of The Journey.

Customers who had already demonstrated borrowing intent could still stall before disbursal because of unresolved documentation, KYC or checkout friction.

The challenge was to identify where progression stopped, understand why it stopped and intervene while the intent was still active.

Journey Stage Observed Friction
Bank Statement Pending documentation
KYC Selfie / verification failure
Recover the customer at the exact point where the journey stopped — rather than restart the entire acquisition process.

Signal OS™ Identified The Friction Before The Borrower Disappeared.

Fibe’s deployment connected application-stage data with customer behaviour to identify stalled borrowers and determine the appropriate recovery intervention.

Signal What It Revealed
Journey Stage Where the application stopped
Stall Reason Why progression stopped
Application Type Personal loan vs BNPL
Merchant Source Which BNPL ecosystem the borrower came through
KYC / Document Status What needed to be completed
Architecture Layer Deployment Capability
Customer Signals Application stage • Stall reason • KYC status • Document status
Decision Intelligence Journey-stage prioritisation • Recovery eligibility
Journey Intelligence APND completion • BNPL recovery
Execution Layer AI Voice Agent • IVR
Human Intelligence Live transfer to Fibe agents for complex cases
Application Stalls → SignalOS Identifies Friction → Recovery Journey Triggered → AI Guides Customer → Completion / Human Transfer → Disbursal

From Stalled Application To Completed Disbursal.

The deployment orchestrated separate journeys based on the customer’s product and point of friction.

Live Journey Trigger Intervention Outcome
APND Completion Credit limit assigned, journey incomplete AI Voice guidance Application completion / disbursal
BNPL Recovery Application stalled Merchant / stage-specific guidance Checkout completion / transfer
Stalled Application → Identify Stage → Understand Blocker → Guide Next Step → Complete → Disbursal

One Agent. Multiple Friction Points. Context-Specific Action.

The Voice Agent engages customers based on the exact point where their application stalled, rather than delivering a generic follow-up.

Agent / Layer Responsibility
Voice Agent Re-engages stalled applicants
IVR Supporting outreach
Workflow / Routing Triggers recovery based on application state
Human Transfer Escalates complex customer situations

Automation Where The Journey Is Clear. Humans Where Context Matters.

Most recovery journeys were designed for autonomous execution. When the customer’s query moved beyond the agent’s supported context, the conversation could be transferred to Fibe’s team.

AI Handles Human Teams Handle
Application recovery Complex queries
Step guidance Exceptions
KYC / documentation guidance Cases requiring human assistance
BNPL checkout guidance Escalated completion

Capabilities Are Not Projected. They Are Measured.

81%

Customer Connectivity

65%

KYC Completion

24%

BNPL Disbursal Conversion

Live

Deployment Status

From Personal Loan Recovery To Multi-Product Application Intelligence.

The deployment expanded beyond a single personal-loan recovery flow into multiple stalled-application cohorts across BNPL use cases

  • Application-stall cohorts
  • KYC / document friction patterns
  • Merchant-specific BNPL cohorts
  • Product-specific recovery logic
Started With Expanded Into
APND recovery BNPL recovery
Personal loans Merchant-specific BNPL cohorts
Application completion Product-specific friction intelligence