How dLocal autonomously resolves 45% of engineering tickets with Port - driving 20% performance gains

dLocal unified services, repos, teams, and policies in Port’s Agentic SDLC Platform, then built an agentd on top of that foundation to take tickets from requests to deployed PR.
dLocal (NASDAQ: DLO) processes payments for companies like Google, Uber, Microsoft, and Amazon across 40+ emerging markets in Latin America, Africa, and Asia. They run more than 500 million transactions per month, growing +50% YoY, with 99.99% uptime SLA - all supported by 400 engineers across 60 teams.

Today, an AI agent autonomously resolves 45% of engineering tickets, end-to-end. Developer productivity is up 20% - the equivalent of adding 20% headcount with zero hiring.

How did they get there?

The challenge:

Before Port, dLocal faced a familiar problem: Engineers spent hours onboarding to scattered tools, stitching together incomplete data, and context-switching to answer questions that should have been a single query. CTO Alberto Almeida saw it for what it was: not a tool problem, but a context problem. And he knew it'd only get worse as AI agents entered the picture. Agents working off scattered, outdated data doesn't just slow you down; it makes confident, costly mistakes.

The solution:

To address these challenges, dLocal turned to Port’s Agentic SDLC Platform to provide the foundation to allow them to scale their AI-SDLC. An Agentic SDLC Platform provides the Context Lake, Workflow Orchestration, Agent Management, and guardrails needed to enable organizations to build, govern, and operate AI across the SDLC without losing control. 

First, dLocal unified their entire engineering ecosystem - services, infrastructure, repositories, teams, policies, and more - under Port’s context lake as a single source of truth. Then they built dCoder on top of that foundation - an AI agent orchestrated through Port that handles routine engineering work, while maintaining complete visibility and control.

When an engineering ticket lands, dCoder is triggered to evaluate its scope and pulls context from Port (what services are involved, who owns them, where the code lives, what patterns this team follows). It generates a plan that respects organizational standards, writes code, runs security and QA checks, and opens a PR. Engineers review and either approve or send it back.

The results:

• 45% of engineering tickets autonomously resolved by the agent
• Teams using dCoder saw 20% performance gains
• Deployment frequency up 60–70%
• Lead time down 50%
• MTTR down 50%
• Developer onboarding 80% faster

Developers are still accountable. They just don't have to do the grunt work.

Watch the video:

‍

In this short video, Almeida shares why they chose Port, and the key capabilities that enable them to scale agentic coverage across the SDLC.

"Today, the agent does 45% of all our tickets. The single metric I care about the most is end-to-end agentic coverage. I go to Port and see the agent's work, rejection rates, and how engineers code differently. My entry point for all of that is Port." Alberto Almeida, CTO, dLocal

Summary

dLocal built autonomous workflows on top of a stable, unified foundation that makes agent adoption sustainable instead of chaotic. Port serves as the context layer, governance system, and the interface developers and managers use daily to monitor the SDLC.

Read the full case study to learn more about dLocal’s journey and the design decisions that enabled them to scale their agentic SDLC.

{{read_case}}