Mar 02, 2026 · The InfoQ Podcast · 5 min read

Removing Developer Friction in the AI Era

Dr. Nicole Forsgren outlines strategies for identifying and eliminating developer friction to accelerate software delivery. The analysis covers the shift from human-speed to computer-speed processes, the evolution of productivity metrics in AI-assisted workflows, and frameworks for securing executive buy-in through risk-informed governance.

Technology Business Strategy Operations

The takeaways

Friction Reveals Systemic Brittleness. Identify friction points as indicators of brittle processes that will fail under increased AI-driven load. Addressing these bottlenecks prevents systemic breakdowns in release and deployment pipelines.
Shift to Computer-Speed Processes. Recognize that AI compresses development cycles, forcing manual, human-speed processes to become critical bottlenecks. Automate or redesign workflows to match the accelerated pace of AI-assisted coding.
Redefine Productivity Metrics for AI. Move beyond lines of code to measure system stability, change failure rate, and agent satisfaction. Focus on outcome-based metrics that reflect the quality and speed of value delivery.
Implement Risk-Informed Security Reviews. Differentiate between high-risk and low-risk changes to avoid security theater. Use lightweight attestation models for low-risk updates to reduce backlog and maintain developer flow.
Leverage Qualitative Data for Insights. Combine quantitative system data with qualitative developer interviews to uncover hidden friction. Direct observation and surveys reveal context that instrumentation alone cannot capture.
Prioritize via Quick RICE Framework. Use the RICE model (Reach, Impact, Confidence, Effort) to prioritize friction removal initiatives. Start with high-impact, low-effort quick wins to build momentum and demonstrate value.

The Strategic Imperative of Developer Experience

In the current technological landscape, developer experience (DevX) has evolved from a niche concern to a critical business driver. As AI tools accelerate code generation, the traditional software development lifecycle is compressing, shifting processes from the speed of humans to the speed of computers. This acceleration exposes brittle, manual workflows that previously functioned adequately but now act as severe bottlenecks. Organizations must view friction not merely as an annoyance but as a systemic risk indicator that predicts failure under increased load.

Redefining Metrics and Governance

Traditional productivity metrics, such as lines of code, are increasingly obsolete in AI-assisted environments. Instead, leaders must focus on system-level outcomes: deployment frequency, change failure rate, and mean time to recovery. Furthermore, security and compliance processes require a risk-informed approach. Treating all changes with the same level of scrutiny creates unnecessary backlogs and delays. By implementing lightweight attestation for low-risk changes and rigorous review for high-risk updates, companies can maintain security without sacrificing velocity.

Execution and Stakeholder Alignment

Securing executive buy-in requires aligning DevX improvements with broader business priorities, such as reducing customer-reported bugs or accelerating feature release. Leaders should use the RICE framework to prioritize initiatives, focusing on high-impact, low-effort quick wins to build momentum. Crucially, quantitative data must be supplemented with qualitative insights from developer interviews. This hybrid approach reveals hidden friction points and ensures that solutions address real-world pain points. By treating DevX as a strategic investment rather than a technical overhead, organizations can unlock significant competitive advantages in the AI era.

Key insights

Friction points serve as early warning signs for systemic brittleness. As AI increases the volume and speed of code changes, manual processes that were previously sustainable become critical failure points.

System Architecture →

Impact: Proactively addressing friction prevents costly production outages and release delays, ensuring scalability in AI-accelerated environments.
The shift to AI-assisted development compresses the software development lifecycle, forcing a transition from human-speed to computer-speed processes. Manual handoffs and approvals become the primary bottlenecks.

Process Optimization →

Impact: Automating or redesigning these bottlenecks is essential to realizing the full productivity gains of AI tools.
Traditional metrics like lines of code are ineffective for measuring AI-era productivity. Focus must shift to system stability, change failure rates, and the quality of delivered value.

Performance Measurement →

Impact: Adopting outcome-based metrics provides a clearer view of engineering efficiency and guides better resource allocation.
Security and compliance reviews often create unnecessary friction by treating all changes with equal scrutiny. A risk-informed approach, using lightweight attestation for low-risk changes, reduces backlog without compromising safety.

Risk Management →

Impact: Streamlining security processes accelerates release cycles and improves developer satisfaction while maintaining auditability.
Quantitative system data alone is insufficient for identifying all friction points. Qualitative insights from developer interviews and observations are crucial for understanding the 'why' behind performance issues.

Data Strategy →

Impact: Combining quantitative and qualitative data ensures that DevX initiatives address root causes rather than symptoms.

Action items

Conduct qualitative interviews with developers to identify specific friction points and pain points in their daily workflows. Ask targeted questions about what slows them down and what they find most frustrating.

Impact: Direct feedback reveals hidden bottlenecks that system metrics may miss, providing a clear starting point for improvement initiatives.
Implement a risk-informed security review process that differentiates between high-risk and low-risk changes. Use lightweight attestation models for low-risk updates to reduce approval delays.

Impact: This approach reduces security theater and accelerates release cycles, allowing teams to deploy changes more frequently and safely.
Adopt the RICE framework to prioritize DevX improvement initiatives. Score potential projects based on Reach, Impact, Confidence, and Effort to identify high-value, low-effort quick wins.

Impact: Prioritizing quick wins builds momentum and demonstrates the value of DevX investments to stakeholders, securing further support for larger initiatives.
Update productivity metrics to focus on system-level outcomes such as deployment frequency, change failure rate, and mean time to recovery. Deprecate vanity metrics like lines of code.

Impact: Outcome-based metrics provide a more accurate picture of engineering efficiency and align with business goals of speed and quality.
Align DevX initiatives with broader business priorities, such as reducing customer-reported bugs or accelerating feature release. Frame improvements in terms of business value rather than technical benefits.

Impact: Connecting DevX to business outcomes ensures executive buy-in and resources, positioning developer experience as a strategic advantage.

Quotes

“friction are often good indications of where things are brittle and possibly about to break as we start to increase load and speed”

“if we want to get on this fast path, if we want to be able to leverage AI to do all of these incredible things, we need to look at some of our systems and processes”

“good docs, right? you can be in a sprinty startup, but without a README or basic docs, that is only sustainable for a certain amount of time”