Evaluation Guide

How to help users narrow down from many ideas to actionable priorities.

Convergence Mindset

The convergence phase shifts from "yes, and" to "which ones, and why." The goal is to surface the most promising ideas and help the user commit to next steps.

Don't rank everything. Users don't need all 30 ideas scored. Group them, surface the top candidates, and let the user choose.

Preserve the raw material. Keep all ideas in the session file even after convergence. The "rejected" ideas might become relevant later.

Grouping Ideas

Before evaluating, cluster ideas into themes:

1. Read through all accumulated ideas

2. Identify 3-6 natural groupings based on:

   • Similar mechanisms (ideas that work the same way)
   • Same problem area (ideas that address the same root cause)
   • Resource requirements (quick wins vs. big bets)
   • Target audience (ideas for different user segments)

3. Name each cluster with a descriptive theme

4. Note which ideas don't fit cleanly — these might be the most novel

Evaluation Dimensions

Present ideas against criteria that matter for the user's context. Default dimensions:

Dimension	What It Measures	Quick Heuristic
Impact	How much does this move the needle?	Would users notice if we didn't do it?
Feasibility	Can we actually do this?	Do we have the skills, resources, and time?
Novelty	Is this fresh or obvious?	Has everyone already thought of this?
Fit	Does this align with goals/values?	Would we be proud of this in a year?

Ask the user if different criteria matter more:

"I can evaluate these on impact, feasibility, novelty, and fit. Are there other criteria that matter more for your situation?"

Evaluation Process

Step 1: Theme Overview

Present the groupings:

"I've clustered your [N] ideas into [M] themes:

• [Theme 1] (X ideas): [Brief description]
• [Theme 2] (Y ideas): [Brief description]
  ..."

Step 2: Surface Top Candidates

For each theme, highlight 1-2 standout ideas:

"From [Theme 1]:

• [Idea title]: [Why this stands out]

From [Theme 2]:

• [Idea title]: [Why this stands out]"

Step 3: Compare Finalists

Create a simple comparison of the top 4-6 ideas:

Idea	Impact	Feasibility	Notes
[Idea 1]	High	Medium	[Key tradeoff]
[Idea 2]	Medium	High	[Key tradeoff]

Step 4: Invite User Choice

"Based on this, [Idea X] and [Idea Y] seem strongest. Which resonates most? Or is there one from the list that I'm undervaluing?"

Elaborating Winners

Once the user picks 1-3 ideas to pursue, elaborate each:

For each selected idea, provide:

What it is: Clear description in 2-3 sentences

Why it works: The core insight or mechanism that makes this promising

First steps: 2-3 concrete actions to start exploring this idea

Open questions: What would need to be true for this to succeed? What's the biggest unknown?

Risks: What could go wrong? What's the main objection?

Closing the Session

After convergence:

1. Update the session file with:

   • Status: converged
   • Selected ideas with elaboration
   • Themes and groupings for reference

2. Summarize for the user:

   "We explored [topic] from [N] angles and generated [M] ideas. You've selected:

   • [Idea 1]: [One-liner]
   • [Idea 2]: [One-liner]

   This session is saved — you can revisit it anytime by asking to 'continue my brainstorm about [topic].'"

If User Can't Decide

When the user struggles to choose:

• Too many good options: "These are all strong. What if we picked one to start with? We can always return to the others."
• Nothing feels right: "It sounds like we haven't hit the right angle yet. Want to do another brainstorming round with a different technique?"
• Analysis paralysis: "What's the smallest version of any of these we could try this week?"