Infrastructure and Operations

• The internal AI strategy is rooted in a four-year-old event-store system that processes millions of SDLC events daily, which the firm built to debug workflows long before the current AI boom.28:48
• Verification is no longer just code review; it has evolved into a multi-layer control stack incorporating security scanners, app-store regulatory checks, and financial compliance agents.9:49
• Cost has become a primary bottleneck similar to cloud adoption, leading the firm to prioritize secure, fast-spinning container runtimes while scouting model-routing strategies to manage inference expenses.14:03

Governance and Metrics

• The firm rejects vanity productivity metrics like lines of code or commit counts, instead utilizing investigation-focused dashboards that allow business leaders to define success based on specific initiatives.34:17
• An unusual 92–93% response rate on internal surveys—sent to over 3,000 employees—has served as a primary driver for product direction and infrastructure changes.63:47
• Operating models follow a hybrid design: a centralized platform provides a common registry and security guardrails, while domain-expert teams retain ownership of the specific skills and agents required for their unique tech stacks.6:12