Preparing Offers and Proposals — Artificial Intelligence in Sales

Gains:

Ability to produce a customer-specific value-based proposal draft from discovery notes
Ability to print pricing, scope and ROI sections clearly and verifiable
Ability to structure and customize the offer to prioritize objections

The offer is the document that the customer reads most carefully during the sales process, just before saying "yes". However, most proposals are copy-paste documents where the seller explains himself, full of feature lists, and barely addresses the customer's problem. A good offer is not a product brochure, but a mirror in which the customer can see his own situation and the solution specific to that situation. In this unit, you will learn how to build a value-based, verifiable and objection-prioritizing offer with artificial intelligence support, based on your discovery notes; You will learn how to manage the risk of fake figures of artificial intelligence, especially in the ROI and pricing sections.

The Skeleton of a Good Offer

A strong proposition follows a certain logical flow: it starts from the customer's world, moves to the solution and the justification for the investment. Standard sections and their purposes:

Section

Purpose

Common mistake

Executive summary

"why, what, how much" on one page

Explain yourself, not the customer

Problem / current situation

Reflecting pain from the customer's language

Common expressions

Starting from Discovery Notes

The raw material for a value-based proposal is the notes from the discovery call: the problem the customer describes in their own words, the metrics they measure, internal influencers, budget signals, and the decision process. AI is powerful at turning these notes into a structured outline.

Role: You are an expert in B2B sales writing value-based proposals. DRAFT a proposal from the following discovery notes. Discovery notes:- Customer: {{ company, industry, size }}- Main problem (customer's own statement): {{ ... }}- Impact of this problem: {{ e.g. Lost Leave the Price and ROI sections BLANK; I will fill them.

Beware: The prompt leaves price and ROI blank on purpose. The reason for this is the risk of fake numbers, which we will see shortly.

Writing ROI and Verifiable Value

The ROI section is the most convincing but most dangerous part of the proposal. The guarantee language here (“you will recover your investment in 3 months”) is both unethical and risky; One failed customer destroys all your credibility. The correct approach is to state the assumptions clearly and give range.

Weak ROI sentence:

With this solution, you reduce your costs by 40% and earn your investment back within 2 months, guaranteed.

Powerful ROI phrase:

Based on the ~120 hours of manual reporting burden per month you shared in discovery, saving around 36-60 hours per month seems possible if 30-50% of this process is automated (the typical range we see with our customers). This estimate is based on the cost per hour assumption you gave and the actual result will vary based on team adaptation.

Difference: the second version is based on the data given by the customer, gives a range, clearly states its assumption and makes no guarantees. Enforce this discipline when printing ROI to AI:

Write an ROI section based on the following VERIFIED data. Data you can use: {{ customer-provided metrics }}Our actual impact range: {{ e.g. 30-50% based on historical data }}Strict RULES:- Only use the above data; new number FITTING.- Using warranty language; Use expressions such as "typical range", "estimated". - Write the assumption underlying each calculation. - Where you are not sure, write "this assumption should be verified with the customer".

Beware: AI likes to fill the gap, and if it doesn't have data it can produce realistic-looking but completely made-up numbers (e.g. "industry average is 47%)." Tie EVERY number in the quote to a source: either it's customer-provided data, it's your actual historical data, or it's clearly marked "assumption."

Risk of Fake Figures in Pricing

Never let the AI write the pricing section freely. AI doesn't know your price list; It may invent a unit price, discount rate or package name that it has not seen from you. Proper workflow: you (or your pricing/engineering team) give the prices, AI just presents them in a neat, penciled, and justified format.

Prepare the following price items into a clear and professional pricing table for the customer. CHANGE numbers, ADD new items, FIT the discount. Items:- {{ Installation: 25,000 TL (one-time) }}- {{ Monthly license: 8,000 TL / 10 users }}- {{ Training: included }}Add the concrete benefit the customer receives next to each item in a sentence. Show the total and payment terms on a separate line.

Validating the AI Sketch with Real Data

AI is a tool for speed, not a source of truth. Healthy workflow:

Collect and structure discovery notes.
Have AI generate draft (with or without price and ROI).
Get technical/engineering verification: Can the promised scope and schedule actually be delivered? Confirm with the implementation team.
Add prices from official source: Price list, approved discount authorization.
Track each number: For each number in the quote, ask "where does this come from?" Answer the question.
Do an objection check: Have known objections been met already?
Simplify and send: Write the executive summary last, put it at the beginning.

Structure Prioritizing Objections

A good offer quietly addresses objections during the sales call within the document. The “this might be expensive” concern is met by ROI and tiered pricing; Concerns of “transition will be difficult” are met by clear timetable and included training; The "what if it doesn't work" concern is met by realistic success criteria and a pilot phase proposal. Ask the AI openly: “What three objections would a CFO raise when reading this proposal, and what part of the proposal addresses them?” This reveals blind spots in your draft.

Tip: Break pricing into items and stages if possible, rather than one giant number. When the customer sees what is for what, the perception of "expensive" turns into the perception of "justified investment". Additionally, offering a small pilot phase reduces the risk of the big decision.

Mini Case: Quality Software for a Manufacturing Company

Mert sells quality management software to manufacturing companies. In his discovery with "Dora Metal", the customer says that he experiences an average of 3 returns per month due to the late discovery of faulty parts and that each return has a certain cost. Mert gives this data to artificial intelligence and creates a value-based draft. In the ROI section, the AI writes “we reduce returns by 90%” in the first round; Mert rejects this and tightens the prompt and revises it as "based on the 3 returns/month given by the customer, early detection is expected to prevent some of this rate; the exact rate depends on process compliance." He adds the prices from his own price list and has the implementation team confirm the 6-week installation plan. Result: because the customer sees his own figures in the proposal, he feels confident saying "you calculated this from my data".

Common Mistakes

Self-explanatory executive summary: Company history in the first paragraph. The customer wants to see "we" (his problem) not "me".
Making AI make up prices: It invents a number it does not see; Always add from official source.
Guarantee language: Unverifiable promises such as "absolutely", "guarantee", "X% return" undermine trust and legal security.
Statistics of unknown origin: "Companies in the sector save 60%" — from where? Do not use unsourced numbers.
Unclear scope: If included/excluded is not written, the project will become a discussion later.
Poor next step: “Call if you have questions” is passive; Suggest one clear action (e.g. “30-minute confirmation call Tuesday, July 14”).

In summary

The offer is not a product brochure, but a document reflecting the customer's situation and his specific solution.
Discovery notes are the raw material of the value-based offering; AI quickly turns them into a structured draft.
Base ROI on customer-provided data, range, write assumptions clearly, avoid guarantee language.
Never let AI match pricing; You add the figures from the official source, AI just edits the presentation.
Link each issue to a source and confirm scope/schedule with technical team.
A good proposal addresses known objections early and silently within the document.

Application task

Write a real or realistic discovery scenario (customer's problem, at least two concrete metrics, budget signal). Using these notes, have the AI produce a draft proposal excluding price and ROI. Then transform the ROI section into a version with assumptions marked, based solely on the metrics the client provided and your realistic range of impact. Add the pricing table yourself and simply ask the AI to edit its presentation. Finally, I asked the AI “what three objections would a CFO have to this proposal?” ' and evaluate whether your draft meets these objections.