Marketing proposals with Claude: connect diagnosis, deliverables and price
Preparing a commercial proposal with Claude, always reviewed before it goes to the prospect.
By Educasium
Prepare a proposal with Market Proposal →
You want to turn an already-completed diagnosis into a clear commercial proposal, with a scope, a timeline and a price the prospect can understand in one read. This document is not a simple template to fill in: a poorly built proposal can commit your business to a price, a deadline or a result promise you never actually intended to keep.
Market-Proposal works identically regardless of the tool used — Claude Code, a Claude.ai conversation, ChatGPT or Gemini: the document produced is Markdown text, with no script and no required file. That portability changes nothing about one essential point: the file produced remains a draft to review, not an offer ready to send as-is. The price, the announced team and any result commitments remain entirely your responsibility, never the model’s that drafted the text.
This article details how to structure this proposal from a real diagnosis, and why a human review before sending is not an optional formality.
Summary
- What a proposal turns into commitments, and what it puts at stake
- What you get, whatever your tool
- Connecting diagnosis, deliverables and price without promising results
- Preparing a proposal step by step
- What Market-Proposal never does for you
- Training to draft a proposal without skipping the review
- Frequently asked questions
What a proposal turns into commitments, and what it puts at stake
A commercial proposal turns an identified need into named services, with a price, a timeline and precise responsibilities. It is not a statement of intent: once signed, every line commits your business, which makes reviewing it before sending just as important as drafting it.
Why this document cannot be drafted in a single pass
A solid proposal starts from an already-established diagnosis — an audit, a conversation with the prospect, an analysis of their need — rather than inventing context to justify standardized services. Without a real diagnosis upstream, the proposal risks offering actions that do not answer the problem the prospect actually has.
What the document must never assert on its own
No proposal drafted with AI help should claim a guaranteed return on investment, cite a reference client that does not exist, or promise a quantified result with no evidence. These three elements must be removed from the document as long as no real data supports them, rather than filled with an example that would give a false impression of solidity.
What you get, whatever your tool
| Working environment | What you get | Practical difference |
|---|---|---|
| Claude Code, with filesystem access | The same document, with an optional save to a project file | Nothing to copy by hand if you ask for it to be saved |
| Claude.ai in conversation | Document shown in the conversation, to copy to keep it | Copy and paste the proposal yourself once reviewed and approved |
| ChatGPT or Gemini | Same document structure, rebuildable from the same principles | Supply the expected sections yourself if the tool does not know them natively |
| Claude via the API, in an integration | Document generated through successive exchanges, no required file | Plan for storage and a review process on the application side |
A portability that never removes the need for review
Whether the document comes from an automatically saved file or a manual copy-paste, human review remains the same mandatory step before any sending. This skill’s portability makes it easy to produce in any tool; it does nothing to make verifying the prices, deadlines and commitments it contains easier — that stays human work every time.
Connecting diagnosis, deliverables and price without promising results
Every service listed in the proposal should answer a specific element of the diagnosis — an observed problem, a need the prospect expressed — rather than a standard service catalog offered with no apparent link to that particular case. This traceability between diagnosis and service makes the proposal more convincing and easier to defend when the prospect asks a question.
The service price, any separate ad budget, and options must appear separately, so a total does not hide what genuinely falls under your billing and what goes straight to advertising platforms without passing through your margin. Case-study and ROI sections require real evidence; when absent, good practice is to remove those sections rather than fill them with an example presented as real.
Preparing a proposal step by step
Step 1: Confirm the need and the scope before drafting anything. Check that the diagnosis being used genuinely matches what the prospect expressed, rather than starting from a standardized service that would need justifying afterward with a need reworded to fit it.
Step 2: List deliverables with their owners and the data expected from the client. State who produces each deliverable on your team and what the client must supply for that work to move forward, so no dependency stays implicit once the engagement begins.
Step 3: Separate the service fee, ad budget and options. Present these three lines distinctly rather than in a single total, so the prospect understands exactly what your billing covers and what goes elsewhere without passing through your business.
Step 4: Remove any section with no real evidence behind it. If no verifiable case study or ROI data exists to support a section, delete it from the document rather than filling it with an example that would give a false impression of results already achieved.
Step 5: Have the whole document reviewed before any sending to the prospect. Check price, announced team, dependencies and terms with someone who genuinely commits your business, never treating the document the model produced as a final version ready to go without that check.
What Market-Proposal never does for you
The skill never contacts the prospect, sends no document, and in no case counts as acceptance of an offer: it is text for you to review and send yourself, through the channel of your choice, after a complete check. The file produced stays a draft until that review has happened, whatever its apparent formatting quality.
It never commits your business to a price or a deadline you have not decided and approved yourself: any proposal sent with a price or a result commitment not checked by someone responsible for your commercial offer remains entirely your responsibility, not the tool’s that drafted it.
In the proposals we help prepare for self-employed professionals and small businesses, the most frequent mistake is not a miscalculated price — it is sending a version still marked by an example or a figure the model proposed as illustration, never reviewed before it was sent.
Training to draft a proposal without skipping the review
Structuring a commercial proposal with AI, while keeping the reflex to check everything before sending, is a practice built through concrete cases. Educasium’s Master Claude training covers producing commercial documents with AI for self-employed profiles or small-business owners, with no prior development experience. For a business owner or a self-employed professional under the relevant scheme, the FIFPL covers part of the cost of this Qualiopi-certified training under the 2026 criteria set at €300/day and €900/year, with e-learning capped at 50% of the daily rate.
Frequently asked questions
Can I send the document produced by this skill directly to the prospect?
No, never without a complete review first: the document remains a draft until the price, announced team, dependencies and terms have been checked by someone who genuinely commits your business. Sending an unreviewed version risks commitments you might not have knowingly approved. The final price remains entirely your decision: even when correctly calculated by the model from the information supplied, it must be approved before any sending by whoever genuinely commits your business.
Can the document include a satisfied-client example to strengthen the proposal?
Only if it is a real, verifiable case; absent that evidence, the section should be removed rather than filled with an example presented as real. An invented client reference, however plausible, remains a fabrication that exposes your business if the prospect checks whether it exists. If the only available example remains anonymized or partly vague for confidentiality reasons, state that explicitly rather than implying a named, verifiable reference.
What if the proposal includes an ad budget on top of your fees?
Present that budget as a line separate from your fees, never blended into a single total, so the prospect clearly understands what share of the sum goes to advertising platforms and what share genuinely corresponds to your service. This separation also avoids any later ambiguity about what was billed and why. This rule also applies to any optional item still under discussion: it should only appear in the total once the prospect has accepted it, never before.
Does this skill working in any tool mean I do not need to review it?
No, not at all: the portability of the tool used to draft the document says nothing about the reliability of its content. Whether the proposal comes from Claude Code, a plain conversation or an API integration, the same complete human review remains necessary before any sending to the prospect. Treat that portability as a practical production advantage, never as a guarantee about the reliability of the content it produces.
A well-built commercial proposal connects a real diagnosis to named services and an understandable price, without ever promising an outcome nothing guarantees. That is the role of the Market-Proposal preparation pack, usable identically in any tool, provided it is always reviewed before sending.
Ahead of this proposal, structuring a cross-functional marketing audit provides the diagnosis it should rest on: our guide to combining analysis and evidence in a marketing audit details how to prepare it without duplicating recommendations.
Training 100% fundable via OPCO/FIFPL. Qualiopi-certified programme. To learn how to prepare a reliable commercial proposal with AI as part of our Master Claude training, contact Educasium and specify your status (employee, self-employed, business owner) and your goal.