For anyone running Planning Analytics or TM1, the decision to modernize and migrate to TM1 Version 12 (TM1 v12, or PA SaaS) isn’t about taking advantage of new features. It’s about preparing for what breaks along the way.
At IBM TechXchange last week, our own Michael Cowie presented a behind-the-scenes look at an experiment we’ve been running internally:
Can an AI agent make the migration process faster, cleaner, and more predictable?
Instead of more slideware, Mike is demoing a working prototype, built using Watsonx Orchestrate, leveraging retrieval-augmented generation (RAG) in a vector database (Milvus) to leverage IBM and QueBIT knowledge on v12 migrations, as well as TM1-specific tools used by AI to analyze live TM1 v11 environments in order to find potential migration issues and help guide administrators through resolution steps automatically.
The Real Challenge with Migration
IBM provides a command-line tool (TM1 Migration Assistant) to convert and package up environments from TM1 v11 to TM1 v12. It does its job, but it also produces massive log files filled with every action the migration tool performs. It’s incredibly hard for administrators to find potential migration issues and cleanup opportunities, such as:
The tool converts what it can, then leaves the rest up to you to decipher from log files and follow-up testing, post-migration.
Teams often lose time and miss migration deadlines by not having a clear understanding of what they’ll be required to fix before or after migration.
Our Experiment: AI as a Migration Copilot
Instead of manually scanning logs line by line we provided IBM Watsonx Orchestrate with a TM1 v12 Migration knowledge base in Milvus and a set of tools that allowed the AI agent to search for, and catalogue migration action items required for potentially any TM1 v11 environment.
Using Watsonx Orchestrate, we built an agent that can:
✅ Reach into a live TM1 v11 database to query information about your application
✅ Quickly search thousands of TM1 objects for known v12 migration issues
✅ Recommend remediation paths based on identified issues, or in some cases fix them on the spot!
✅ Answer freeform questions like “Why won’t Process X work after migration?” or “What are alternatives to ExecuteCommand references in my processes?”
This agent is not positioned as a product (yet), but the framework is reusable across multiple PA tasks, not just migration:
Why This Matters for PA Teams
If migration to Version 12 is on your roadmap, whether six weeks or six months out, being able to assess readiness in minutes instead of days makes a difference.
You don’t have to wait for a full rollout to start building intelligence around your environment.
And that’s really the point of Mike’s session at TechXchange:
AI doesn’t replace upgrade planning or the people responsible for performing upgrades, it accelerates expert decision-making.
Want to See It in Action?
If you’re exploring modernization or just curious how AI agents can support Planning Analytics administration, we’re happy to share more after the conference.
Just reach out, even if it’s just to compare migration war stories.
https://www.linkedin.com/in/cowiemichael/