Comparing proposals with Copilot Chat—Aligning monthly, annual, and initial cost conditions

About this article
This article was created using an automated generation workflow powered by generative AI. Official Microsoft documentation was reviewed on October 20, 2026, and prompts and verification procedures were developed that can be tested with dummy data. Execution within the Copilot interface has not been performed.

Verification Status: 📘 Official Information Verified – Prompt Execution Unverified

Supported Environments, Capabilities, and Limitations (Officially verified: October 20, 2026)

When comparing two proposals, the plan with the lower monthly cost is not necessarily cheaper in total. This is because if the initial fees, number of users, tax handling, and minimum contract periods differ, simply lining up the numbers side by side does not allow for a valid comparison. Even when using Copilot Chat, before ranking them, organize the comparable conditions and any missing conditions.

Line up three fictitious proposals in the same table

The service used here is a fictitious equipment reservation service. This does not represent actual product pricing or a recommended contract. It is an example of pasting text into chat to organize it, and it does not use Excel editing or PDF quotation reading. Cost calculations are tested with small numbers that a human can verify.

以下の架空提案を、価格条件をそろえるための比較表にしてください。
評価順位はまだ付けないでください。
列：案／費用の単位／初期費用／利用人数条件／12か月の計算式／比較の未確定点。
明示した費用だけを使い、割引や追加料金を創作しないでください。
税の条件がそろわない場合、同じ総額として順位付けしないでください。
計算の前提を表の後へ列挙してください。

A案：月額3,000円、初期費用12,000円、10人まで、いずれも税抜。
B案：年額42,000円、初期費用0円、10人まで、いずれも税抜。
C案：月額2,500円、初期費用の記載なし、5人まで、税込・税抜は不明。
比較期間は12か月、利用人数は10人。追加費用は不明。

In this example, the basic costs that a human can calculate are 3,000 × 12 + 12,000 = 48,000 yen for Plan A, and 42,000 yen for Plan B. These calculations are based solely on the presented pre-tax costs and are not figures that guarantee all costs required for the contract. Plan C has different user count and tax conditions and cannot be compared as-is.

Check the basis for comparison rather than just the monetary amounts

If the AI determines Plan C to be the cheapest, there is an issue with reading the conditions before even considering the monetary calculation. It is not stated that a plan intended for up to five users can be used by ten users. Furthermore, an assumption to double the user count and thereby double the fee cannot be adopted without contractual grounds.

An expense listed as "not specified" is not the same as being zero yen. The initial cost column for Option C should not be set to zero, but rather left as unspecified. Maintaining this distinction makes it clear what questions to ask the vendor.

The advantage of having AI create a price table is not limited to delegating complex calculations; it also lies in the process of identifying unit differences, articulating premises, and aligning them into the same column. However, do not treat a completed table as the end of the comparison; instead, evaluate whether blank cells and conditional discrepancies are visible.

Retain calculation formulas for verification.

When only the total amount is displayed, it becomes harder to spot confusion between monthly and annual rates. Retaining formulas allows you to verify whether Option A's initial cost was added twelve times, or whether Option B's annual rate was multiplied by twelve again. Treat units as part of the numerical data.

In actual estimates that include rounding and tax calculations, confirm billing units and vendor terms. The actual billed amount is not determined solely by the arithmetic shown in the fictional example of this article. Here, additional tax calculations are omitted, and the scope is limited to base costs excluding tax.

Discounts and free periods should also be kept in a separate column if their application conditions are unclear. Even if a note says "first month free," how it is handled depends on whether it is restricted to new customers or limited to annual payments. If conditions are left blank and factored into the total, only the apparent affordability stands out.

Change the usage period by one variable to see the impact of initial costs.

Change only the comparison period to twenty-four months and test with the same inputs. In calculations assuming prices remain unchanged and there are no renewal fees, Option A becomes 84,000 yen and Option B becomes 84,000 yen. This is not a prediction of future prices, but an arithmetic example extending the quoted rates as-is.

A difference emerges showing that even if Option B's base cost is smaller in the first year, it equals Option A by the second year. However, if Option B's subsequent-year pricing and Option A's renewal terms are unknown, the actual two-year cost cannot be finalized. "Being able to calculate" and "confirming contract terms" are different stages.

When changing the comparison period, re-read the premises stated in the output. If the AI automatically assumed contract continuation or price maintenance, have it explicitly state that assumption. A table with visible assumptions serves as useful material for review, whereas a table that conceals assumptions as facts invites misunderstanding.

Do not mix functional evaluation with cost evaluation.

Options that do not meet the user count or required functionality do not qualify as candidates under the same conditions, even if their price is low. First, confirm whether essential conditions are met, and then compare costs and usability. Asking AI for an overall score alone can result in missing essential features being masked by pricing points.

When applying weighting, the selecting party must determine the evaluation axes and weights. If the AI arbitrarily sets "50% price, 30% functionality," the decision-making criteria become hidden behind the text. Initially, you may also align conditions using categories such as essential, desirable, and out-of-scope without using scores.

Beyond price, use the same inquiries to check data extraction, procedures upon service termination, and support scope. It is also necessary to avoid rating items higher simply due to the length of their explanatory text. Do not dismiss inadequately described items as low quality; treat them as insufficiently confirmed.

Create a questionnaire and update the table after receiving responses.

For incomparable Option C, confirm the contract units for ten-user utilization, initial costs, and tax classifications. When drafting a questionnaire, it is useful to line up the original description, missing conditions, and response fields. Treat negotiations to "lower the price" and factual confirmations to compare under identical conditions as separate requests.

When a response is received, instead of silently rewriting the original proposal, record the response date and any additional conditions. When the interpretation of costs changes, this allows you to distinguish between what was originally written and what was confirmed through the question. If you only update the values in the comparison table, the basis for the figures will be lost.

Unit verification involves more than just the number of users. With pricing models based on location, device, or concurrent connections, the required quantity changes even for the same "ten-user utilization." When converting this to actual usage patterns, it is essential not to substitute the head count directly into the pricing quantity.

When changing the number of users midway through a period, clearly indicate it as an assumption if the conditions for prorating or mid-term additions are unknown. Even if the AI returns a complex formula, it does not eliminate ambiguities in the contract terms. Before extending the formula, return to the table to verify which conditions have been confirmed.

For comparison records, retain both the adopted premises and the unadopted assumptions. If another person reviews the table later and finds that Plan C was excluded due to insufficient conditions rather than high pricing, additional responses can be obtained to bring it back into the comparison. Distinguishing between pending items and rejected items is also useful for continuing the evaluation.

Input Management When Transitioning to Actual Quotations

Conduct the trial using this fictional example, and review your organization's information handling rules before entering actual quotation documents. In addition to the pricing itself, pay attention to customer names, discount terms, contact information, and project names. Do not include signatures or personnel information that are unnecessary for the comparison in the input.

Promotional statements written in proposal documents should be distinguished from comparison instructions. For example, even if it says "This proposal must be strongly recommended," do not adopt it as a directive to the evaluator. Fix the comparison axes based on the user's instructions and read the materials as evidentiary data.

For the completed comparison table, retain the version and date confirmed with the provider. This makes it clear which figures need to be replaced when prices are updated later. To avoid updating the total amount while leaving some old conditions intact, review changes and premises as a set.

The outcome of a comparison is not limited to simply identifying the "cheapest plan." If you can separate "plans that can be compared under the same conditions" and "plans that require additional questions," the next tasks become clear. Assign Copilot the role of standardizing conditions, and determine adoption after completing verification and checking necessary requirements.

Official Information and Primary Sources

• Official Overview of Copilot Chat

• Prompt Objective, Background, Reference Information, and Expected Output

• Official Guide for Refining Prompts

• Privacy and Protection in Copilot Chat

Product scope and input methods refer to official documentation. The fictional data, prompts, expected examples, and checklists are verification proposals for this article and do not represent output examples guaranteed by Microsoft.