Guides · Engineering decisions

Practical guides to software repair.

Understand the evidence and decisions behind a repair, then choose a workflow that fits your application and engineering process.

By Altari Systems · Reviewed 12 September 2026

See how Remedy works →Request early access

Start with the engineering question

These guides cover diagnosis, recovery metrics and deployment boundaries. Examples are illustrative unless explicitly identified as measured product or customer evidence.

How AI bug fixing worksEvidence, diagnosis, source changes and verification explained.MTTR and AI incident responseMeasure where time is spent and where automation helps.Self-hosted AI bug fixingData boundaries and deployment questions for private infrastructure.

Go deeper into the repair process

Root-cause analysisTurn observations into a testable hypothesis.Candidate code repairExact source, focused changes and project checks.What counts as verified?Understand the difference between a tested change and an observed recovery.

Evaluating a product?

Use the AI bug-fixing tools comparison to distinguish incident workflows from security autofix and pull-request repair. Check Remedy’s architecture and availability labels before defining an early-access pilot.

From bug report to verified fix.

See how Remedy connects evidence, diagnosis, repair, deterministic checks and production verification.

Explore the complete workflow →

Continue exploring