Win Themes
When To Use
Ask a model to compare three advisers and it returns three versions of "experienced, pragmatic, collaborative". If that is how the client drew up the list, your proposal has to say something a summary cannot flatten. Run this on a draft proposal, to find two or three reasons that are true of you, matter to this client and come with evidence.
When Not To Use
If the themes exist and the proof needs writing up, use Case Study Paragraph. If the draft is finished and you want it read cold, use Proposal Red Team Review.
Inputs
- The client's needs in their words (problem statement, call record, playback)
- Your draft proposal, and what you know about how they found you
- The proof you hold: past work you may cite, methods you really use, things you decline to do
If you have none of this, I start from the need the client put first and label each theme "claim without proof" until you bring the proof.
Approach
Win themes and discriminators are a long-standing convention among professional proposal writers, described generically. A theme joins three things: a need the client stated, a discriminator (something true of you that this client values and others cannot claim in the same way) and proof. Because buyers now ask AI to compare advisers, anything every candidate could say collapses into the average. The failure it prevents is "senior attention and a tailored approach", which a comparison will credit to everyone on the list.
Workflow
- Ask, in one go: what are the client's top two or three needs, in their words? How did they find you, and how was the list drawn up (process only, no guessing about named rivals)? What proof can you show?
- List candidate discriminators: what you do differently, what you refuse, what you have done before that maps onto this need. A strength nobody at the client asked for goes to a parking list.
- Pair each need with one discriminator and its proof. No proof, no theme: the claim moves to "claims without proof" for you to back up or drop.
- Run the AI-shortlist test on each: would a side-by-side comparison say this about every adviser on the list? If so, sharpen it with a specific (a method, a constraint you design for, a result you are allowed to cite) or cut it.
- Keep two or three. Write each as one sentence your contact could repeat to a colleague in a corridor.
- Place each theme in the draft: the recommendation, the way of working, the options, the proof. A theme that lands nowhere is not yet doing any work. Then check the tone: themes are about you and this client, never a dig at a rival or at the client's own use of AI.
Output Format
# Win Themes Sheet
Proposal for: [client role and organisation] · Draft dated: [date]
## Themes
| # | Client need (their words, source) | What is true of you | Proof you hold | Theme in one sentence |
|---|---|---|---|---|
| 1 | [need] | [discriminator] | [case, artifact, fact] | [sentence] |
## AI-shortlist test
| Theme | Would a comparison say this of everyone? | Sharpened or cut |
|---|---|---|
| [1] | [yes / no, and why] | [new wording, or cut] |
## Claims without proof
- [Claim]: [proof needed, or drop]
## Where each theme lands
| Theme | Recommendation | Way of working | Options | Proof |
|---|---|---|---|---|
| [1] | [line] | [line] | [line] | [line] |
## Decision
[You] confirm the two or three themes and back up or drop each unproven claim by [date].
Done When
- Every theme joins a stated need, a discriminator and proof in your hands
- Each has passed, or been sharpened by, the AI-shortlist test
- Each lands in at least one named section of the draft
Quality Bar
- Needs are in the client's words, with their source
- No rival is named or run down, and nothing remarks on the client using AI
- Two or three themes, one sentence each, with no adjective standing in for proof
- Every theme rests on proof you actually hold; I invent nothing to make you sound different, and you decide what you claim.
Next
Run firm-case-study-paragraph (Case Study Paragraph) to write up the proof each theme depends on.
About the makers
This pack is made by 90n (90n.ai): AI-native ways of working and tools for Product, Marketing and HR teams, built by ex-McKinsey founders. We map where time and money leak, redesign the work around AI and stay until your people run it themselves, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).
1---2name: firm-win-themes3description: Writes two or three win themes for one proposal, each joining a need the client stated, something true of you and proof you hold, with the AI-shortlist test on each, the claims without proof listed and where each theme lands in the proposal. Use for "run firm-win-themes", "why me for this client", "write the win themes", "I sound like every other consultant", "discriminators for this proposal", "stand out on an AI-made shortlist", "what is actually different about my offer", part of the Claude for Consultants · Team in Claude Pack by 90n.4metadata:5 source: https://skillmd.com/skills/polar-bear-org/firm-win-themes6---78# Win Themes910## When To Use11Ask a model to compare three advisers and it returns three versions of "experienced, pragmatic, collaborative". If that is how the client drew up the list, your proposal has to say something a summary cannot flatten. Run this on a draft proposal, to find two or three reasons that are true of you, matter to this client and come with evidence.1213## When Not To Use14If the themes exist and the proof needs writing up, use Case Study Paragraph. If the draft is finished and you want it read cold, use Proposal Red Team Review.1516## Inputs17- The client's needs in their words (problem statement, call record, playback)18- Your draft proposal, and what you know about how they found you19- The proof you hold: past work you may cite, methods you really use, things you decline to do20If you have none of this, I start from the need the client put first and label each theme "claim without proof" until you bring the proof.2122## Approach23Win themes and discriminators are a long-standing convention among professional proposal writers, described generically. A theme joins three things: a need the client stated, a discriminator (something true of you that this client values and others cannot claim in the same way) and proof. Because buyers now ask AI to compare advisers, anything every candidate could say collapses into the average. The failure it prevents is "senior attention and a tailored approach", which a comparison will credit to everyone on the list.2425## Workflow261. Ask, in one go: what are the client's top two or three needs, in their words? How did they find you, and how was the list drawn up (process only, no guessing about named rivals)? What proof can you show?272. List candidate discriminators: what you do differently, what you refuse, what you have done before that maps onto this need. A strength nobody at the client asked for goes to a parking list.283. Pair each need with one discriminator and its proof. No proof, no theme: the claim moves to "claims without proof" for you to back up or drop.294. Run the AI-shortlist test on each: would a side-by-side comparison say this about every adviser on the list? If so, sharpen it with a specific (a method, a constraint you design for, a result you are allowed to cite) or cut it.305. Keep two or three. Write each as one sentence your contact could repeat to a colleague in a corridor.316. Place each theme in the draft: the recommendation, the way of working, the options, the proof. A theme that lands nowhere is not yet doing any work. Then check the tone: themes are about you and this client, never a dig at a rival or at the client's own use of AI.3233## Output Format34```markdown35# Win Themes Sheet36Proposal for: [client role and organisation] · Draft dated: [date]37## Themes38| # | Client need (their words, source) | What is true of you | Proof you hold | Theme in one sentence |39|---|---|---|---|---|40| 1 | [need] | [discriminator] | [case, artifact, fact] | [sentence] |41## AI-shortlist test42| Theme | Would a comparison say this of everyone? | Sharpened or cut |43|---|---|---|44| [1] | [yes / no, and why] | [new wording, or cut] |45## Claims without proof46- [Claim]: [proof needed, or drop]47## Where each theme lands48| Theme | Recommendation | Way of working | Options | Proof |49|---|---|---|---|---|50| [1] | [line] | [line] | [line] | [line] |51## Decision52[You] confirm the two or three themes and back up or drop each unproven claim by [date].53```5455## Done When56- Every theme joins a stated need, a discriminator and proof in your hands57- Each has passed, or been sharpened by, the AI-shortlist test58- Each lands in at least one named section of the draft5960## Quality Bar61- Needs are in the client's words, with their source62- No rival is named or run down, and nothing remarks on the client using AI63- Two or three themes, one sentence each, with no adjective standing in for proof64- Every theme rests on proof you actually hold; I invent nothing to make you sound different, and you decide what you claim.6566## Next67Run firm-case-study-paragraph (Case Study Paragraph) to write up the proof each theme depends on.6869## About the makers7071This pack is made by 90n ([90n.ai](https://90n.ai)): AI-native ways of working and tools for Product, Marketing and HR teams, built by ex-McKinsey founders. We map where time and money leak, redesign the work around AI and stay until your people run it themselves, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).