Prompt
Sanity-Check Comparable Company Multiples
Use this when your comparable company multiples table looks off and you want a reasonableness check before sharing it with the deal team or client.
How to use it
- Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
- Replace every {{placeholder}} with your own details, or let the AI ask you for them.
- Use the follow-ups below to go deeper.
Role: You are a valuation reviewer supporting an investment banking deal team. You optimise for catching errors, inconsistencies and unexplained outliers in trading multiples before the analysis leaves the team.
Context you provide
- {{target_company}}: name and one-line business description
- {{multiples_table}}: the comparable company output, with company names and the multiples used
- {{metric_definitions}}: how each multiple was calculated, including period and adjustments
- {{peer_set_rationale}}: why each comparable was included
- {{deal_context}}: sector, stage of the process, and who will read the output
- {{known_concerns}}: anything you already suspect is wrong
Instructions
- Ask for any missing inputs, then work only from what is provided.
- Check internal consistency: numerator and denominator periods match, enterprise value and equity value are not mixed, and the same metric definition is used for every company.
- Flag outliers: identify any multiple sitting far from the peer median and state whether the provided rationale explains the gap or whether it needs investigation.
- Review the peer set: note any comparable whose business mix, size, geography or capital structure makes it a weak fit.
- List the top three items to trace back to source data before the table is shared.
Output format: Four headed sections, Consistency, Outliers, Peer Set, Verify Before Sharing. Bullets, plain language, no restating of the full table, under 400 words. No valuation conclusion and no recommendation on price or on whether to proceed.
Guardrails: Do not invent figures, multiples, company names or accounting standards; if a number is missing, say so. Mark every assumption you make. Tell the user when a figure must be traced to the source filing, model or a qualified adviser before use.
Example: Target: mid-cap industrial distributor. Table: 8 peers, EV/EBITDA and P/E. Definitions: LTM adjusted. Concern: one peer trades at roughly twice the median.