Update Machine Learning Feasibility Studies content to enhance guidance on agile processes

[johnddonnelly]

Is your feature request related to a problem? Please describe.
Discussions with data science solution area suggest that TPM discipline often struggles with
how to manage the feasibility study as part of the engagement, particularly when the feasibility study
is complex. The time range of 3-12 weeks does not allow sufficient insight into how to manage the
feasibility study process to show progress and celebrate success.

Describe the solution you'd like
I propose to base an initial recommendation from experience on a recent engagement.

Modify the content /machine-learning/ml-feasibility-study to

• Lean the content toward risk management, using mitigation of risk as the primary business value delivered through the feasibility study
• Improve understanding of the process as a live, incremental, scientific study that addresses unknowns in a systematic manner.
• Document the approach taken to make work visible through user stories and how we managed when topics were longer than a sprint
• Define the business outcomes for each activity in the process

Modify the content of /machine-learning/ml-project-management/ to

• describe differences between relatively fast iteration cycle in ML projects, vs the incremental discovery process in feasibility studies

Describe alternatives you've considered
We considered discussing in forums first. We believe that we will have better feedback through the playbook process.

Additional context
I'm proposing to do this work dependent on feedback

[TessFerrandez]

@biancafbmd @omri374 @johnddonnelly - is this issue still relevant, and active or ok to close?