Salesforce Jul 1, 2026

How AI Learned to Investigate Mobile Build Failures Like an Experienced Support Engineer

Article Summary

Archana Indran's team at Salesforce turned years of mobile debugging expertise into AI that investigates build failures like a senior engineer. Eight people now support 60+ repositories with less firefighting.

The Mobile CI/CD team at Salesforce built Analyze Build Tools, an AI system that replicates how experienced support engineers diagnose mobile build failures. Instead of teaching AI to read logs, they taught it to ask the right investigative questions across multiple systems.

Key Takeaways

Critical Insight

By encoding investigative expertise into AI skills rather than building dashboards, an 8-person team scaled support for 60+ mobile repositories while cutting debugging time by three-quarters.

The breakthrough wasn't in log parsing: it was teaching AI to build a case the way human investigators construct hypotheses and gather evidence.

About This Article

Problem

Salesforce's Mobile CI/CD team had a scaling problem. Eight engineers were manually investigating failures across 60+ repositories. They built Splunk queries and correlated logs across multiple systems, but couldn't tell if issues came from application code, Managed Pipelines, Salesforce infrastructure, or changes from Apple and Google.

Solution

Archana Indran's team built Analyze Build Tools with AI-powered skills like /mp-investigate-build and /analyze-mobile-builds. These tools replicate how experienced engineers investigate problems. They gather evidence from Managed Pipelines, Splunk logs, and historical metrics to find root causes without manual log searching.

Impact

The system let the team handle more repositories without hiring more people. Developers now get answers with context instead of raw data. The eight engineers can focus on real infrastructure work instead of repeatedly investigating the same types of failures.