BRQ and BBTS boost negotiation conversions by 306% with AI on WhatsApp

BBTS needed to evolve, in partnership with BRQ, a traditionally complex and sensitive journey: debt negotiation. The process demanded clarity in the terms presented, the ability to run real-time simulations, security in handling information, traceability and adherence to business rules.

The challenge was not just to build an informational chatbot, but to apply conversational AI to a highly critical journey, one capable of understanding different intents, keeping the context of the conversation, querying information in internal systems, supporting the formalization of agreements and enabling operational steps, such as issuing payment slips.

The solution also needed to operate with authentication, stability, performance and governance, meeting the security requirements of a financial institution while keeping the experience simple for the customer and robust in operation.

Together with BBTS and AWS, we structured a credit recovery journey via WhatsApp using generative AI, scalable architecture, transactional integration and layers of governance and security.

The work began by mapping the renegotiation steps, the systems involved, the success indicators, the business rules and the technical requirements needed to enable the solution. From there, we defined the project's architecture, the integrations, the conversational agents, the contextual memory, the prompt engineering engine and the control mechanisms.

The solution guides the customer from identifying the initial intent through simulating personalized terms, querying information, supporting the formalization of agreements and issuing payment slips. With integration to internal systems and banking APIs, the journey connects AI, business rules and operations to make renegotiation smoother, more contextual and more secure.

As a result, BRQ and BBTS expanded the digital capacity of the collections operation, increased negotiation conversions by 306%, enabled 50% of interactions to be completed without human intervention and reduced the average number of installments in agreements from 33.17 to 14.22. The technology foundation also opens the way for new generative AI applications in critical banking journeys.