How to Use AI to Turn Vendor Quotes Into a Side-by-Side Recommendation Memo

Problem statement and who this is for

Vendor decisions often get delayed for a simple reason: the information arrives in different formats and no one has time to normalize it. One quote is a PDF, one is an email, one is a pricing sheet, and someone still has to explain the tradeoffs.

This workflow is for office managers, executive assistants, clinic or school admins, coordinators, and operations staff who gather quotes and need to present a clean recommendation.

Prerequisites

• Two or more vendor quotes or proposals
• A small set of comparison criteria, such as price, turnaround time, contract terms, support, or scope
• One AI tool such as ChatGPT, Claude, or Gemini
• A decision-maker who needs a brief, not a document dump

Numbered workflow steps

1) Choose comparison criteria before you paste anything

Do not let the model decide what matters.

Use 4 to 6 criteria such as:

• total cost
• what is included
• timeline or turnaround time
• support or service level
• contract flexibility
• risks or missing details

2) Normalize the quote content

Pull the relevant parts from each quote into one working input. If needed, paste excerpts rather than entire files.

3) Ask for a comparison table plus a recommendation memo

Use this prompt block:

{
  "task": "Turn vendor quotes into a side-by-side recommendation memo",
  "input": {
    "comparison_criteria": [
      "total cost",
      "what is included",
      "timeline",
      "support level",
      "contract flexibility",
      "missing details or risks"
    ],
    "vendor_materials": "PASTE THE QUOTE CONTENT HERE"
  },
  "instructions": [
    "Create a side-by-side comparison using only the information provided.",
    "Do not invent pricing, features, or contract terms.",
    "If information is missing, label it as missing.",
    "Write a short recommendation memo that explains the practical tradeoffs.",
    "If no clear winner exists, say that plainly."
  ],
  "output_format": {
    "comparison_table": "Plain text table or bullets",
    "recommendation_memo": "One short memo",
    "follow_up_questions_for_vendors": ["Bullets"]
  }
}

4) Review the missing details first

This is where bad recommendations happen.

If one quote is missing setup fees, response times, or contract terms, do not let the model quietly compare incomplete information as if it were complete.

5) Generate a decision-maker version and an internal working version

The decision-maker version should be short. The internal version can include the follow-up questions and open gaps.

{
  "task": "Create decision-maker and working versions of a vendor recommendation",
  "input": {
    "verified_comparison": "PASTE THE VERIFIED COMPARISON HERE"
  },
  "instructions": [
    "Create a concise version for the decision-maker and a slightly fuller internal version.",
    "Do not add new facts.",
    "Keep both versions practical and easy to scan."
  ],
  "output_format": {
    "decision_maker_version": "Plain text",
    "internal_working_version": "Plain text"
  }
}

6) Keep the memo attached to the original quote pack

The memo should not replace the source material. It should make the source material usable.

Tool-specific instructions

ChatGPT

Useful for quickly organizing quote content into criteria and drafting a short recommendation memo.

Claude

Useful when quote language is dense and you want careful comparison without overclaiming.

Gemini

Useful if the final memo will move into Docs or Gmail in Google Workspace. Availability depends on plan and settings.

Quality checks

• The comparison criteria were chosen by you.
• Missing details are labeled clearly.
• The recommendation explains tradeoffs, not just a winner.
• No features, fees, or terms were invented.
• The memo is short enough for a busy decision-maker.

Common failure modes and fixes

Failure mode: The model picks the cheapest option by default

Fix: give explicit comparison criteria that include scope, support, and risk.

Failure mode: Incomplete quotes look comparable

Fix: add a missing details review before finalizing the memo.

Failure mode: The memo is too vague to support a decision

Fix: require practical tradeoffs and a clear requested next step.

Failure mode: The output is too long

Fix: separate a short decision-maker memo from an internal working version.

Failure mode: Follow-up questions are missed

Fix: always generate a vendor clarification list as part of the output.

Sources Checked

• OpenAI Help Center, File Uploads FAQ, accessed 2026-03-07: https://help.openai.com/en/articles/8555545-file-uploads-with-chatgpt-and-gpts
• Anthropic Help Center, What kinds of documents can I upload to Claude?, accessed 2026-03-07: https://support.claude.com/en/articles/8241126-what-kinds-of-documents-can-i-upload-to-claude.ai
• Google Workspace Admin Help, Gemini AI features now included in Google Workspace subscriptions, accessed 2026-03-07: https://support.google.com/a/answer/15756885

Quarterly Refresh Flag

Review by 2026-06-05 to confirm upload support and current Workspace AI drafting availability.