Creditice

Sector: financial services · Scope: nationwide

Creditice places mortgage and business loans through a broker network spread across the country.

The problem

A credit broker doesn’t sell a single product: they handle mortgages, business loans and several variants, each with its own assignment rules, requirements and exceptions. When a broker is sitting with a client and a question comes up — does this profile qualify? what documentation does this product require? — they need the answer right then, not the next day.

With the network spread across Mexico, resolving those questions depended on someone from the internal team being available. That slows down placement and burns hours answering the same thing over and over.

How we solved it

We implemented AI agents trained on Creditice’s own knowledge: their credit types, the rules a broker must follow to assign a loan to a client, and their internal processes.

• Knowledge base (RAG) built on the company’s real documentation, not generic knowledge.
• Conversational lookup: the broker asks in plain language and gets the answer with the company’s own criteria.
• Available 24/7, without depending on the internal team’s schedule.
• Continuous training: the agent also works as an onboarding tool for new brokers.

Benefits for the broker

• Resolves questions on the spot, with the client in front of them.
• Reinforces product knowledge without formal training sessions.
• An agile tool to check rules and requirements in the field.

Benefits for the business

• Less internal team time spent answering repetitive questions.
• Standardized knowledge across the network: every broker gets the same criteria.
• Faster onboarding of new brokers, with no training bottleneck.