Skip to main content
Back to Work

5-month items. Shipped in 2 weeks.

10×

delivery rhythm

Advisory · Forward Deployed · Ongoing · 2025–2026

way2.com.br

The Problem

Way2 runs Brazil's smart-grid Meter Data Management — a regulated, multi-service platform on Oracle with CCEE integrations and a 30-person product team. Local environments took days to spin up. Seeding was manual. Cross-service tests slowed everything else down.

Roadmap items routinely sat at 3–5 months. Support tickets piled into manual triage. Way2 wanted AI-assisted everything — engineering, QA, product, support. No one had embedded AI in a legacy-plus-modern platform deep enough to stick.

The Approach

We embedded across the product team and ran weekly 1:1 mentoring — engineers, PMs, QA, support. Engineers and product managers learned to plan with AI, write specs with AI, ship with AI. We standardized how stories get created — every demand grounded in context, knowledge, and references before the first commit.

Per-PR on-demand environments (Helm + Terraform on GKE) spin up the full platform when a PR opens — ready to test, not days of local setup. AI-pre-qualified support: every Freshdesk ticket arrives pre-classified with answer suggestions and references pulled from the live knowledge base. way2-ai-plugin centralizes team knowledge so every engineer reaches the same AI surface.

The Outcome

30
1:1 mentees
<1d
PR to preview env
5mo→2wk
roadmap velocity

5-month roadmap itemsshipped in 2 weeks

Roadmap items quoted at 5 months ship in 2 weeks — spec-driven, TDD, fully automated. Per-PR environments collapse the test loop from days to clicks. Support tickets land already pre-classified, with AI-suggested answers grounded in the knowledge base. 30 engineers and product people work AI-native by default. The 5-month sprint is gone.

Start a similar project.

Talk to us