Sep 3, 2026
Optimizing a Process for Target KPIs with AI: A Job Analysts Used to Do by Hand
Two processes. Clear KPIs. An honest look at the results, with real prompts, diagrams, and the AI's own reasoning.
By Maria Alish
Category
Engineering
All Camunda blog posts tagged with Engineering.
Aug 27, 2026
How to Test an AI Agent That Never Does the Same Thing Twice
Testing an AI agent means testing the process around it, which is deterministic and free, the model's behavior, which is neither, and whether it stays consistent over time.
Aug 26, 2026
From Weeks to Days: What Camunda's AI Skills Actually Saved Me
A first-hand account of building a real Camunda 8 process with the official AI Skills, and exactly which skill fixed which problem, from gateway conditions to agent tool wiring.
Aug 5, 2026
The Camunda Migration Skill: A Faster Way to Move from Camunda 7 to Camunda 8
If you're looking for the fastest way to migrate from Camunda 7 to Camunda 8, start with the migration skill.
By Lana-Sophie Stawowski, Peter Bojtos
Jul 21, 2026
Your Contact Center Got an AI Upgrade. Your Back Office Didn't.
Don’t settle for a faster version of the same broken process. Transform it with Camunda and NICE.
By Mike Fuller
Jul 16, 2026
What Native Agentic Architecture Actually Looks Like
A technical deep dive into native agentic architecture, using a KYC onboarding example available in code.
Jul 7, 2026
Why Bolting LangChain onto a Workflow Engine Backfires at Scale
Bolting LangChain onto your existing workflow engine may look like it's the path of least resistance...
Jun 29, 2026
Camunda 8.10-alpha2: More Consistency for Builders, More Control for Operators
Camunda 8.10-alpha2 is live. Highlights include a unified document source dropdown for connectors, start form support in the Desktop Modeler, Helm v4 validation, and Bitnami sub-chart removal.
Jun 17, 2026
Key Principles in Designing Agents for Production – Part IV: Design Patterns for Guardrails and Safeguards
Learn how to protect your Camunda agents in production with BPMN-based guardrails, tool gatekeeping, agent success evaluation, and timing restrictions.
By Niall Deehan
