The Fintech Playbook for Latin America

Santiago Suarez, founder and CEO of Addy, joins Angela Strange and Gabriel Vasquez of a16z to discuss building a fintech company in Colombia that has grown to serve over 3 million consumers and 50,000 merchant partners. The conversation begins with Suarez's personal background — coming from Colombia on a full scholarship to Yale, working at an early AI startup called Navia, being fired as CEO after 35 days, and then spending six years at JPMorgan in a small strategy group close to Jamie Dimon to learn how great CEOs operate.

Suarez explains why he chose Colombia over Brazil or Mexico as his primary market. He cites two key insights: rapid smartphone adoption (75% penetration, faster than Brazil and Mexico) creating a zero-marginal-cost distribution channel, and a strong cultural tradition of installment payments without adequate infrastructure to support it. He also notes that his relative unfamiliarity with Colombian financial regulations allowed him to see opportunity rather than obstacles.

A significant portion of the conversation covers Addy's foundational technology decisions. Suarez credits his CTO for insisting on a monorepo architecture at a time when microservices was the industry standard, and for adopting an event-sourcing architecture using Kafka, which logs every single event across the platform — over 10 million per day. These decisions, combined with a Databricks partnership for real-time data ingestion, created the infrastructure foundation that later enabled aggressive AI adoption.

On AI, Suarez describes Addy's approach as starting counterintuitively with legal automation rather than customer service. In Colombia, companies face 'tutelas' — emergency constitutional actions requiring 48-hour responses or the CEO faces personal liability. Building AI pipelines for this high-stakes, accuracy-demanding use case forced rigorous development of RAG pipelines and feedback loops that then transferred directly to customer service. Their AI agents now handle 100% of customer service queries with close to 80% full resolution and no human in the loop, and a merchant onboarding agent boards 2,000–3,000 merchants per month with over 20% improvement in conversion.

Suarez discusses a visit to Kazakhstan to meet the CEO of Caspi, a company he views as a comparable model — a highly focused fintech succeeding in a single country. Key lessons from that meeting included prioritizing NPS and customer experience, thinking modularly about product roadmaps, and ignoring equity investor skepticism for years while building conviction around a contrarian geographic strategy.

On organizational culture, Suarez emphasizes running the company in English (to attract global talent and raise internal standards), maintaining a written-first culture for all decisions (which proved prescient for AI agent context), and organizing around a North Star metric rather than cascading OKRs. He describes the company running 150 heads below budget while exceeding growth targets, attributing this to AI-driven productivity gains.

Looking ahead five years, Suarez envisions Addy as a force for financial inclusion across multiple Latin American countries, a mostly AI-native organization where agents and humans collaborate under a unified 'human and agentic resources' function, and a place where a few hundred people do the best work of their lives while pushing technological boundaries from Bogotá.