Getting started on a project is intimidating, and the temptation to jump-start the process with an AI prompt is strong. But shelve that impulse. Previous Teamwork Lab research has shown that using AI at the start of brainstorming comes with a hidden cost. Your team trades a sense of ownership and originality for speed.

The Teamwork Lab hypothesized that talking out loud about a project would help team members overcome initial resistance and capture original thinking. The Lab conducted two separate studies with Atlassians doing real work, asking them to start a new workflow that prioritized human thinking out loud first and AI polish second.

research at a glance

What we did:

The Teamwork Lab ran two separate studies on brainstorming and goal-setting.

  • Brainstorming: In a study conducted from March to April 2026, 55 participants used Loom to think out loud on a real work problem that they were stuck on. They used Rovo to turn those ideas into a structured first draft.
  • Goal-setting: In a separate study conducted from January to February 2026, 42 Atlassians recorded goal-setting conversations with their managers in Loom. They used Rovo to turn the transcripts into SMART goals that were aligned with team OKRs and job-level expectations.

What we learned:

  • In the first study, 84% felt less resistance to starting the task and 78% got to a usable draft faster than with their usual approach.
  • In the second, 91% said the process improved goal quality over previous cycles.
  • To get quality AI outputs, human thought is key. It’s as easy as thinking out loud.

What is think-aloud ai brainstorming?

Under pressure for time, many people jump straight to brainstorming with AI. But that can result in generic drafts that are a pain to review and improve. Think-aloud brainstorming preserves the messy human thinking that results in creative ideas that move the work forward and help capture important context.

What does think-aloud AI brainstorming look like?

The Teamwork Lab asked 55 Atlassians to try a new workflow to overcome the blank page problem when starting a new project.

The participants:

  1. Recorded a Loom (screen recording tool) to think-aloud.
  2. Used Rovo (AI tool) to turn it into a clean, structured draft.
  3. Reviewed the draft in Confluence (team workspace for sharing knowledge).

Afterward, they completed a survey on the experience.

What we discovered:

  • The ‘think-aloud with AI’ method improved task initiation (84% lower resistance) and drafting speed (78% faster usable drafts).
  • The workflow also reinforced human creativity and mental well-being. 82% said it reminded them of the importance of thinking deeply about a problem before using AI, and 75% said the result felt like their own ideas – distinct from generic AI output.
  • A strong majority (84%) said they are likely to use this workflow again for future projects.

What does think-aloud goal-setting look like?

Writing high-quality SMART goals (specific, measurable, achievable, relevant, and time‑bound) that are aligned with team strategy is also a brainstorming-intensive task. We asked 42 Atlassians who had completed at least one prior performance review to use Loom to record a goal-setting conversation with their manager. Participants then used Rovo to draft SMART goals and OKRs (objectives and key results) aligned with their career goals and company objectives.

We discovered that:

  • 91% of participants said the Loom + Rovo process led to higher quality goals.
  • 69% of participants reported greater clarity on what was expected of them from their manager, and 79% reported a greater understanding of how their goals contributed to broader team goals.
  • 83% of participants said using this process led them to think about new ways they can use AI, and 88% of participants said they would use the process again for future goal-setting.

Though small, these studies show that chaining AI tools like Loom and Rovo can turn complex, ambiguous workflows (like goal‑setting) into clear, review-ready outcomes.

“For many, getting your ideas on paper is the hardest part. This workflow helps remove that barrier by allowing you to just talk, letting your ideas flow, and then bringing in AI to organize your thoughts into something shareable.”

Sami Nesnidol
Lead behavioral scientist teamwork lab

How to think-aloud brainstorm with AI

As outlined in our Atlassian playbook, think-aloud brainstorming is a five-step process.

Step 1: Pick the work you’re stuck on

Pick something that’s messy and high-judgment, like a decision with a lot of stakeholders or a long-delayed project kickoff. Then open a blank Confluence page to use as a scratchpad.

Step 2: Record 10 to 15 minutes of thinking aloud

Explain the situation as if to a trusted colleague who knows nothing about it. Cover what you’re trying to achieve and why, who will use the result, what’s unclear or risky, and the options you’ve considered. Don’t worry about making it perfect.

If you’re having trouble articulating what the problem is, try these prompts:

  • “If I had to decide today, I’d…”
  • “What’s keeping me up at night about this is…”
  • “If this went amazingly well in six months, what would be true?”

Step 3: Use AI to structure your draft

Ask your AI assistant to review the video and organize your thoughts. What comes back will be a rough draft. You can use AI to clarify your thoughts, for example: “Based on this transcript, write a clear problem statement, the key goals, and three to five non-goals,” or “Point out places where my reasoning is fuzzy or contradictory.”

Step 4: Edit with a critical eye

Read the draft and ask what’s missing, whether the AI has made any mistakes, and what sounds generic. Every AI-generated draft needs human review before passing it on to other stakeholders. (Check out our research on AI workslop for more information.)

Step 5: Decide what the draft should drive

Turn it into a project brief, a decision summary in Slack, or work items in Jira or Trello. Every run should end with a clear next step.

The bottom line

When you talk before you prompt, you’re more likely to actually start the project, and you’re more likely to use the results. It also prevents mental exhaustion, preserving a positive relationship with creative work. exhaustion and preserves a positive relationship with creative work.

Human-first brainstorming also has a host of other benefits for your team besides overcoming initial resistance and producing usable results faster. If your team is having trouble overcoming the blank page problem, try out our play to get unstuck and move forward on your next project or task.


Methodology

The Teamwork Lab ran two studies with Atlassians. In both studies, participants talked through their thinking in a Loom recording, then used Rovo to turn it into a draft. After each study, participants completed a survey.

Brainstorming study. From March to April 2026, 55 Atlassians who didn’t already use this workflow picked a real piece of work they were stuck on, such as a strategy, a project poster, or a feature plan. Each person recorded a Loom thinking out loud about it, then asked Rovo to turn the recording into a Confluence outline or draft they could refine.

Goal-setting study: From January to February 2026, 42 Atlassians took part during an optional goal-setting cycle. Each person recorded a goal-setting conversation with their manager in Loom, then used two template prompts to give Rovo the transcript, their job level, and their growth profile. Rovo drafted SMART goals aligned with team OKRs, and participants refined the goals themselves.