Forecasting Delivery Dates

I've long been an advocate to share the delivery dates of a product enhancement with internal teams. In my opinion, there is no such thing as 'too much transparency' when sharing information about when a product enhancement will get delivered. We can certainly talk through the merits of my approach, how transparency is indicative of organization trust, and how it better serves customers. I talked about it some here.

I am curious about anyone using data science and specifically forecasting capabilities to automate delivery date forecasting. My experience has been primarily in Atlassian shops and I've cut my teeth (and sworn a lot about) Jira. That said, using Jira with even basic levels of consistency across teams gives us a good bit of metadata about remaining work load (stories in the Epic), level of effort for each item (the lifecycle of a ticket, time between stages, story point estimation, etc), and historical risk of scope creep (new tickets added per epic over the last...10 epics closed by the team).

If this isn't an off-the-shelf product now, it certainly will get released by Atlassian or otherwise soon. This ability to automatically enable your downstream teams is incredibly valuable, and opens the doors to working more collaboratively with other teams, and even asynchronously as teams have a pulse on what and when to prepare for.

I am excited for the day when a PMM team gets notified that a new increment is coming in 42 days, and they should add work on their calendar to begin planning their messaging and enablement. For when smoke test QA teams can accurately change their resource planning and build better testing methods when fewer items are released. For when Tech Support agents get hot-off-the-presses alpha versions to begin testing, learning, and understanding features and better help customers when they call.

Are you using any AI generated forecasting dates? Have you built a model on your product deliver behaviors? How have you used them?