The Challenge: Credit Growing Exponentially, Collections Team Staying Lean
- Bold’s credit product was scaling exponentially with the business
- Demand on the collections team was rising fast
- More clients needed help normalizing obligations and accessing relief
- The traditional response would be to grow headcount linearly with demand
- The team wanted the opposite: stay small and well-trained, and lean on technology to absorb the growth
Bold is a fast-growing Colombian fintech, and its credit product was scaling exponentially. Every new credit customer is a future collections customer, and the team was running into the classic collections trap: more credit volume means more clients who need help normalizing obligations, which usually means more agents. Miguel and the team wanted the opposite. The plan was to keep a small, well-trained collections team and use technology to absorb the growth, instead of hiring through every demand spike.
The Solution: Voice AI That Reaches Clients Before They Fall Into Arrears
Dapta gave Bold a voice AI layer that doesn’t wait for delinquency. Bold uses predictive models to identify which clients are likely to run into trouble, and Dapta’s agents reach out before there’s a problem to talk through relief and prevention options.
- Voice AI agents reach out to clients identified by Bold’s predictive models
- Calls happen before delinquency, not after
- Agents walk clients through normalization options and relief tools
- Capacity scales up and down with credit volume without adding agents
- First agent went live in about 30 minutes
- Weekly improvement cycles with the Dapta team keep the agents getting better
The collections flow no longer starts on day 1 of delinquency, when the client already has a problem. It starts before that, anticipating client needs and offering help in advance. The human team stays small and focused on the cases that genuinely need a person.
The Results: A Small Team Powered by AI, Ahead of Client Needs
- The original goal of keeping a small, stable collections team was met
- 0 added agents despite the credit product growing exponentially
- The first AI agent was built in about 30 minutes
- Bold runs a small operation powered by AI instead of a thousand-agent department
- Collections shifted from chasing past-due clients to anticipating client needs
- Weekly improvement cycle with Dapta keeps the agents getting better
With Dapta, Bold turned collections from a headcount problem into a capacity problem solved by AI. The team stayed small, the credit product kept growing, and the collections flow shifted from chasing 30-day past-due clients to anticipating client needs before delinquency ever starts. Miguel’s view of where collections is headed is exactly this: intelligence, data, technological tools, and an immediate response, all in service of helping clients at the right moment.