Can Fusion Automate Project and Task Setup in Workfront Like This?

[JohnMi]

Hello, fellow Fusion users!

I'm exploring how to automate our project setup in Workfront using Fusion and would appreciate your insights on the feasibility and potential challenges. Here’s what we aim to achieve:

1. Project Initiation: When a Marketing Lead submits a detailed Workfront request for various team resources (Email team, Web team, Creative team, etc.), Fusion generates a main project from a preconfigured template ("Project A").
2. Team-Specific Projects: Fusion then automatically creates related team-specific projects (e.g., "Email: Project A", "Creative: Project A") using templates tailored to each team.
3. Automated Task Fields: Within these templates, each task automatically populates the "Project#" field with the unique number from "Project A". The "Task Type" field is pre-set with values such as "team-centric" or "milestone".

I also included a table stating this in another way.

Our goal is to automate the creation of both the main and team-specific projects, ensuring every task references the "Project A" number for consistent tracking and reporting across all related projects.

I need to document our proposed automation for leadership before implementation, and I’m new to Fusion.  So I'd appreciate any insights or confirmation on whether Fusion can handle this level of automation, or any challenges you think we might encounter. I just want to avoid presenting something to leadership that Fusion is not capably of doing. If it can do it I am certain my team we can figure it out. 

Thanks for your help!

[Lawson02]

Seems pretty feasible to me. As long as the request contains everything to complete an execution and there are guardrails preventing bad input, everything proposed looks good to me. I haven't done too many large projects yet as I am still pretty new to Fusion but I am the only dev for my company. For the largest scenario I have created so far, I put a large emphasis on testing within the preview environment and dynamic case handling. One concern I would have for a scenario this large is how to handle it breaking halfway through the execution.

[JohnMi]

@Lawson02 Thank you. I was hoping it could handle most tasks that are logical and unambiguous.

[Rick_MacDuffie · Accepted solution]

Sounds a lot like the request to project pipeline scenario I have had running for 4+ years.  Yes.  Feasible.  A couple of tips I learned along the way.

1. Create a Data store to map the project template IDs to use in the workflow and look them up; assuming Project A may not always have all the same sub-projects.  When Project A needs Email, create an empty project, look that template up and apply that template.
2. Use a Set multiple variable module ahead of any branching logic to initialize data and check to make sure you have everything you need in the format you need it.  This will help address the risks you run with a large scenario failing mid-execution.
3. My workflow has created 8K+ projects and the 3 most common errors I encounter are
   1. Service time out errors which are easily handled with built-in retry logic
   2. IF any tasks get reschedule in the project there can be occasions where the Project's Planned Completion / Planned Start Date get misaligned and a warning is generated.  This is hard to troubleshoot since it is possible that a person with permission to the project starts editing it before your scenario completes its execution.
   3. In my scenario, permissions are all established in the proejct templates.  This means that the project owner that is getting set has permission to edit tasks before the execution is complete and that can cause an issue where the workflow can't edit a task that the PM has changed outside the scope of the execution.  It is a rare thing that could be addressed by not granting permission to the resulting projects until the entire execution is complete.  I just haven't come up with a great solution to solve for that.

[JohnMi]

This is very helpful @Rick_MacDuffie . Thank you. Do you know of a good way to estimate data store space or is it just run it and see?