How to create Sensitive Information Types(SIT), custom SITs and Exact Data Match a deep dive session

This will be a deep dive(L400+) about this topic, the idea is help to you to understand all the capabilities and some concerns related to use, I been working for many years with several customers in this technology and a lot of times the issues are the same, and are related to misunderstanding the patterns and the logic in the background, I'll try to explain all of that in an easy way.
Know you what's means confidence level? the difference between Low, Medium or High? how to affect the proximity windows? Know you that not all the built-in Sensitive Information Types have all the confidence level?
If you create your owns, know you that you can reuse? or clone current SITs to create new ones? What's a pattern? how can I backup my SITs?
Having a clear procedure, this technology will not be a rocket science.
How can I reduce false positives? Add exceptions on the DLP rules is not the best approach to do that, we need to work over SITs to reduce those rates, improving our SITs permit start to trust on them and using it over DLP, MIP, MIP Scanner, MDCA and other services as well.
Exact Data Match is not a big deal if we go step by step, that will be permit to use easily in our environment.
Know you where you can use SITs on Microsoft 365 products?
Having this trusted SITs, permit to use easily over products like:
Data Loss Prevention
MIP Scanner
Sensitivity Labels
Auto-Labeling
Microsoft Defender for Cloud Apps
And several other services, for that reason is important to know all about SITs and use them in the right way.
Know you that you can export and import your Sensitive Info Types?
How to we can reuser regular expressions, keywords and dictionaries?
You can visit my LinkedIn profile at MyProfile.kazn...

[@iharveer]

This is a gold mine! Thanks a lot for sharing !

[@PIrateTomski]

This session was fantastic - I had no idea that the "DLP Test Harness" tool even existed! It will save me a lot of time creating sample data in the future.