Multi-Scenario Org Design and Headcount Modeling

Open your current roster export. Replace employee names with role titles or short ID codes (Manager A, IC 1, IC 2), and keep reporting lines, salary band midpoints, and job level intact.

1. Add a role-label or ID column
2. Remove the name column entirely
3. Save as a CSV with columns for ID, title, level, reports-to, and salary band

What you should see: A CSV where every row is a role, not a person, with reporting lines still traceable through the ID codes. Troubleshooting: If the export mixes salary bands with actual individual pay, use the band midpoint instead of the real figure. Actual pay isn't needed to model structure and cost at this stage.

Start a new ChatGPT conversation, click the paperclip icon, and upload the roster CSV.

1. Click the paperclip icon in the chat bar and select the roster file
2. In your message, state the business goal in one or two sentences (stand up a new function, absorb a merged team, flatten a layer of management)
3. State any hard constraints (no net new headcount, a specific budget ceiling)

What you should see: ChatGPT confirms it has read the file and usually echoes back the current headcount and total cost as a sanity check before proposing anything.

Ask for a defined number of options, each with a short name, so the comparison stays organized rather than turning into one long narrative.

1. Ask for exactly two or three named scenarios (for example, "Option A: Flatten," "Option B: Split into pods")
2. Ask each option to include new reporting lines, headcount change, and cost change versus current state
3. Ask for span-of-control (direct reports per manager) to be calculated for each option

What you should see: A response with a labeled section per scenario, each including a simple structure description, a cost delta, and a span-of-control number.

Follow up asking ChatGPT to condense the scenarios into a single comparison table rather than three separate write-ups, since that's what a leader will actually want to see in the meeting.

1. Ask for a table with one row per scenario and columns for headcount, cost delta, span-of-control, and key tradeoff
2. Ask for a one-sentence recommendation, clearly labeled as a starting point for discussion, not a final answer

What you should see: A compact table plus a short recommendation paragraph you could screenshot directly into a slide.

Before this goes in front of a leader, spot-check the cost delta ChatGPT calculated against your own source numbers.

1. Pick the scenario you're leaning toward
2. Manually add up the salary band midpoints for the roles in that scenario using your own spreadsheet
3. Compare that total against the cost figure ChatGPT gave you, and ask it to show its calculation if the numbers don't match

What you should see: A cost figure you've confirmed independently, not just one you're taking on faith. This matters because a wrong number in a leadership pitch is the kind of mistake that costs credibility.

Ask for a short, neutral summary of what changes for each existing role under the chosen scenario, since that's usually the leader's next question after seeing the numbers.

1. Ask for a two-sentence summary of impact per existing ID or role
2. Keep this output role-based, not tied to specific names, until any personnel changes are finalized through your normal process

What you should see: A short paragraph per affected role explaining the change in plain terms, ready to be turned into real names only after the redesign is approved.