Cortex launches DRIVE to track AI-era engineering health

Cortex has introduced DRIVE, a framework aimed at measuring engineering organizational health as AI speeds up software development. The company says DRIVE focuses on whether an organization can sustainably turn customer needs into reliable software, not just whether individual developers or teams are productive.

The framework is built around five pillars: Delivery, Reliability, Initiatives, Vigilance, and Efficiency. According to Cortex co-founder and CTO Ganesh Datta, the idea is that AI has removed many of the constraints on writing code, while the controls needed to manage faster output have lagged behind.

Cortex argues that engineering work is shifting toward designing and operating the systems that produce software. In that view, the harder questions are moving from pure coding output to operational control, risk, and resource allocation.

The five pillars track a specific set of metrics:

• Delivery: deploy frequency, lead time for changes, and on-call pager volume
• Reliability: functional SLO status and Sev0 and Sev1 incident count
• Initiatives: Tier 1 initiative milestone completion rate and OpEx action item completion rate
• Vigilance: open Critical and High fixable CVEs, assets below the minimum compliance and security bar, and orphaned assets
• Efficiency: cloud spend versus budget, internal AI and LLM token costs, and the percentage of capacity spent on innovation

Operational Excellence review format

Cortex pairs the framework with a recurring Operational Excellence review, or OpEx review, which treats the engineering organization as a system and uses DRIVE metrics to guide decisions on time, staffing, and spending. The company traces the practice back to manufacturing-style operational excellence, where an entire factory is managed as one observable unit.

Recommended reading

Russia’s cloud GPU demand jumps 507% on AI workloads

Cortex says many engineering organizations already run similar rituals. It cites:

• Amazon Web Services: a weekly operational review that reaches the most senior engineering leadership

• Stripe: facilitator-led reviews that shape resource decisions, with API reliability above 99.999%

• Google: blameless production meetings that feed capacity planning

Cortex says the review structure changes with company size. Startup and mid-market organizations run one global review weekly, while enterprise companies split that into local team reviews weekly or biweekly and org-wide reviews weekly or biweekly.

The framework is positioned as something any engineering organization can adopt, while Cortex pitches its own platform as the way to make it practical at scale through scorecards, workflows, and initiative tracking.

Powered by