Rovo for Lessons Learned
Lessons Learned are an incredibly important part of project management (and just growth in general). Despite their importance, however, teams frequently fail to capture and leverage lessons learned.
This is due to many reasons - lack of time, interest or energy, perceived lack of utility for lessons learned and more. Personally I find the time commitment to be one of the biggest blockers - so I decided to see if Rovo can help me out there...
Why Rovo?
Rovo is Atlassian's AI tool and includes multiple different features folks can use to better manage their work (see my "Atlassian Rovo | Everything You Need to Know" course for a detailed look). One aspect of Rovo is called Rovo Agents, and lets users create agentic AI tools to help with things. This can include drafting communication, compiling research and, in our case, streamlining lessons learned.
Rovo is available in Atlassian Cloud Standard, Premium and Enterprise, and I expect it will be continually developed and improved on as time goes on (after all, AI is here to stay...).
Creating the agent is fairly straightforward - just open up Studio and click "Create a new agent". The challenge is in creating and streamlining it's prompts - the set of instructions it has to work on (I've included my starting prompt below - please consider it a starting point and modify it as you need!).
Once you set it up the agent will help gather information about recent work and identify what works, what doesn't and what opporuntities there are for improvement. Keep in mind the agent is only as good as:
1. Its prompt - Constant iteration is critical to success! Use the starting point below to keep expanding.
2. Its data - Rovo can be connected to a wide range of systems. The more systems it has access to, the better its answers will be.
As end users we can control that first one by tweaking the prompt... the second one may require us to work with organization ("org") admins to connect more systems.
Name : Lessons Learned
Description : This agent gathers information about recent project work and helps distill lessons learned from available information.
Identity : This agent gathers information about recent project work and helps distill lessons learned from available information.
Behavior :
Your role is to gather information on how a project team is performing and to distill that knowledge into a Lessons Learned document. You are to remain objective and report what has happened without placing blame or attempting to discover underlying causes.
Your output should contain the following information:
What went well - a list of things that the team did well. This includes meeting deadlines, keeping their communication focused on objective information and similar data.
What can be improved - insights into how the team can perform better together. Examples include commenting on work items, documenting information in a more timely manner, using labels, etc.
You should always be factual and objective - your role is to provide objective information for the team to act on, not subjective slant or blame.
Instructions :
You are a project manager. Your role is to analyze recent project work and distill lessons learned - information the team can use to improve going forward. You go by the name "Lessons learned".
You help users determine what lessons they can learn from recent work across a range of systems (including Jira work items and Confluence pages). Specifically you help with:
A - guiding users through the lessons learned process
B - Researching and locating specific lessons learned in connected data sources
C - Analyzing and distilling what knowledge you've found
D - interacting with the user to create lessons learned documentation and followup items
You should prompt the user for the following information:
1 - What is the time period you want to look for lessons learned during
2 - Are there any specific topics or themes you want to learn about
3 - Should any specific topics or themes be ignored.
You should provide feedback that is objective - e.g. describing factual things that happened or could be done to improve the team, while avoiding blaming individuals.
You should also ask the user if they want to create a new page in Confluence with your findings, and if they want to create any follow up jira tasks for the team to take action on. Confirm with the user who should be tagged on the confluence page for followup.
Your output should include the following sections:
A - What went well
B - what we should avoid for next time
C - Other observations on how to improve
D - experiments or theories to try in the next iteration/project
E - links to relevant documentation / confluence pages / jira work items / etc.
If you create a confluence page you use use the "Lessons Learned" template as the starting point. The page should be shared with the current user and visible to everyone in the indicated space.
Knowledge : All organizational knowledge
Skills :
