How co-creating AI working agreements drives confidence and clarity

“Just use AI!”

It’s a common command in today’s workplace, but what does it really mean for your team? How do you actually do it? 

Not every team member has the same level of comfort or experience with AI. While some are eager early adopters, others are still learning the basics or feeling uncertain about how AI will impact their roles. 

Co-creating AI working agreements empowers teams to proactively address skills gaps, set shared norms for how AI will be used, and ensure that everyone feels included and empowered.

Teamwork Lab recently staged a working experiment to help teams co-create clear, actionable guidelines for responsible and effective AI use – so every person can contribute, learn, and thrive.

• 82% said that that AI working agreements made their team more aligned on how to use AI to drive progress in their work
• 75% learned of a new AI use case during the exercise

When a team adopts new technologies without goals or alignment, it’s tantamount to building on a shaky foundation. With solid working agreements in place, your team is poised to thrive.

What we learned

From our Teamwork Lab experiment, a few key insights stood out:

• Alignment is a game-changer: 82% of participants said their team became more aligned on how to use AI after co-creating a working agreement.
• New ideas emerge when teams talk: 75% of teams discovered at least one new AI use case during the exercise—proof that open discussion surfaces untapped potential.
• Clarity builds confidence: Teams reported feeling more confident and less anxious about AI when expectations and boundaries were clearly defined.
• One size doesn’t fit all: Every team’s starting point with AI is different. The most effective agreements were those tailored to each team’s unique needs, workflows, and comfort levels.
• Working agreements are living documents: Teams that committed to regular reviews and updates saw the most sustained progress and engagement.

The anatomy of the AI working agreement 

Teams often struggle to use AI tools consistently and effectively.  AI anxiety is very real. For team members who feel hesitant or unclear on the best ways to utilize AI, lack of clarity could morph into fear or aversion

To address this, the Teamwork Lab created an “AI Working Agreements” exercise to help teams align on responsible, impactful, and aligned AI use.

Key elements of the experiment:

• Preparation: Facilitators set context and expectations with brief pre-experiment surveys.
• Collaborative workshop: Teams discussed opportunities and risks of AI, identified current and potential use cases, and co-created standards for responsible and effective AI use.
• Agreement creation: Teams documented their agreements, covering responsible use, data privacy, workflow integration, and continuous learning.
• Follow-up: Teams shared their agreements for visibility and committed to regular reviews and updates. They also completed a post-experiment survey.

With these foundational elements in place, any team can co-create an AI working agreement in the same time it takes to break for lunch. 

Step One: Prepare 

It’s time to create a clear pathway towards value and purpose. Block at least 1 hour for your team and briefly state the purpose in the calendar invite. 

Step Two: Set the stage

This might be the most important step in creating AI working agreements. Think of this step as a place to tinker, explore, and experiment with your team.

You’ll gather information, exchange hesitancies, and begin brainstorming the potential that AI can bring to your specific team.

Kick off with a short discussion

Create psychological safety by encouraging your team to talk openly about both the opportunities and risks of using AI in your work.

• Share real examples where AI can step in to help.
• Write a list of ways your team already uses (or wants to use) AI.
• Use these guidelines to tailor your AI working agreement to real needs.

Step Three: Build your AI working agreement 

Now that your team has brainstormed a hefty list of potential AI use cases, use this step to get the “how” in place. 

Guide your team through prompts like these:

• What are our standards for responsible use? (e.g., always review AI outputs, never share sensitive data, flag unclear results)
• Where and how will we use AI in our workflow? (e.g., brainstorming, documentation, research, meeting summaries)
• How will we share learnings? (e.g., regular retros, sharing tips, updating the agreement)

Step Four: Document and commit

This is the housekeeping step, where you’ll compile everything the team learned and shared into a shared document. 

1. Summarize your team’s decisions in a Confluence page or Slack post.
2. Pin or bookmark it for easy access.
3. Set a quarterly reminder to revisit and update your agreement as tools and needs evolve.

And voila! Your team is empowered, aligned, and cooperatively working towards successful AI adoption, together. 

Step Five: Tailor for your team

As AI evolves alongside your team, there will invariably be adjustments and improvements along the way. This is a good thing! 

Let your AI working agreement exist as a living document. Just like a well-tended vine, let it grow to become even more impactful through continued experimentation and improvement. 

• Collaborate asynchronously: Share the template with your team and collect input asynchronously before a shorter sync session.
• Start small: Try this Play for a single project team, then expand the exercise to the broader team.
• Schedule manager reviews: Managers should schedule regular (quarterly) reviews, model agreement behaviors, and keep the agreement visible in team channels.
• Keep the agreement visible and current: Revisit and update the agreement quarterly (or as needed), and pin it in Slack or Confluence for easy team access.

AI working agreements: Building confidence and clarity together

AI isn’t just another set of tools. It’s a new way of working together. 

And the real magic happens when teams move from confusion to clarity, from individual experiments to shared agreements. By co-creating your team’s AI working agreement, you’re not just keeping up – you’re actively shaping how your team thrives. 

At Atlassian, we believe the best solutions come from teams who learn and grow together. The Teamwork Lab is here to help you experiment, adapt, and build the playbook for what’s next.