The Simple Prompt I Use to Let My AI Assistant Handle Pricing Strategy
The Simple Prompt I Use to Let My AI Assistant Handle Pricing Strategy 💰🤖
By Dr. Julie Jones, Ph.D. in Artificial Intelligence
Pricing strategy is one of the most consequential decisions a business can make. A price that is too low erodes margins and signals low quality. A price that is too high suppresses demand and pushes customers to competitors. The classic tools—cost-plus, competitor benchmarking, value-based analysis—are all useful, but they require data, judgment, and a lot of time. What if you could compress that entire process into a single, well-structured prompt that your AI assistant can execute in seconds?
That is exactly what I want to walk you through today. Not a magical one-liner, but a reusable, structured prompt that turns an LLM into a disciplined pricing analyst. I have used variations of this prompt across SaaS, e-commerce, and professional services contexts, and the quality of the output is consistently strong. Let me show you why it works, what goes in it, and how to adapt it to your own business.
Why Most Pricing Decisions Go Wrong
Before we look at the prompt, it is worth understanding the failure modes we are trying to avoid. Research in behavioral economics and pricing literature points to a few recurring problems:
Anchoring on cost. Sellers start from what it costs them to make a product and add a "reasonable" margin. This ignores what the customer is actually willing to pay, which is usually much higher than the cost.
Anchoring on competitors. "They charge $50, so I'll charge $49" is a competitive strategy, not a value-based one. You are pricing by subtraction, not by the value you create.
One-size-fits-all pricing. Every customer is not equally price-sensitive. A flat price leaves money on the table for high-value customers and overcharges low-value ones.
Static pricing. Markets shift. Customer needs shift. A price set two years ago may be miscalibrated today.
A good pricing strategy addresses all four. And a good prompt for an AI assistant should ask about all four explicitly. That is the design principle behind the prompt I'll show you next.
Anatomy of the Prompt
Here is the full prompt. You can paste it into any conversational AI, fill in the bracketed fields with your business details, and run it.
You are a senior pricing strategist with 20 years of experience
across SaaS, e-commerce, and professional services. Your task
is to produce a pricing strategy analysis for my business.
BUSINESS CONTEXT:
- Product/Service: [describe what you sell]
- Target Customer: [who buys it, firmographics or
demographics]
- Current Price: [your current price or price range]
- Main Competitors: [name 2–3 competitors and their
approximate prices]
- Key Value Propositions: [list the 2–3 things customers
actually cite as reasons they buy]
- Volume: [rough monthly or annual units / contracts]
- Cost Structure: [fixed vs variable costs, if known]
- Goal: [revenue growth, margin improvement, market share,
or a specific number]
ANALYSIS REQUIREMENTS:
1. Value-Based Pricing: Estimate the customer's
willingness-to-pay using the value proposition, not
the cost. Show your reasoning.
2. Competitive Positioning: Place my price on a
1–10 scale relative to competitors and explain where
I should sit for the goal I stated.
3. Price Tiers: Design 2–3 tiers (or price points) that
match different customer segments. For each tier,
specify who it targets, what it includes, and the
target margin.
4. Psychological Pricing: Suggest any use of charm
pricing, anchor pricing, or decoy pricing, and
explain the mechanism.
5. Experiment Plan: Design one A/B or market test
that would validate the recommended price.
Specify the hypothesis, the metric, and the
minimum sample size.
6. Risk: Identify the 2 most likely ways this
pricing fails, and the signal that would tell me
to pivot.
OUTPUT FORMAT:
- Use headings for each section.
- Use a table for the tier comparison.
- Use a simple bar chart (text-based) to compare my
current price, the recommended price, and the
competitor range.
- Keep the total response under 1200 words.
- Be specific. Avoid generic advice. Every
recommendation should reference a fact from the
business context above.That is the whole thing. It is long, but every line has a job. Let me unpack the parts that matter most.
Why the Prompt Works
The prompt does five things that a casual question like "what should I charge?" does not:
It sets a role. "Senior pricing strategist with 20 years of experience" primes the model to reason like a domain expert, not a generalist. This is a well-documented technique in prompt engineering: role assignment shifts the distribution of likely outputs toward more specialized reasoning.
It feeds context, not just a question. The model has no memory of your business. By supplying the customer, the value proposition, competitors, volume, and goal, you give it the inputs it needs to do real analysis instead of generic advice. The prompt explicitly tells the model to reference these facts, which reduces the hallucination of irrelevant recommendations.
It decomposes the task. Instead of asking for "a pricing strategy," it asks for six specific deliverables: value-based estimate, competitive positioning, tiers, psychological pricing, an experiment, and risk analysis. This decomposition forces the model to work through each dimension rather than skimming the first one and stopping.
It constrains the output format. Asking for headings, a table, a text-based bar chart, and a word limit keeps the response structured and usable. You get something you can paste into a slide deck or a board memo, not a wall of prose.
It asks for falsifiability. The experiment plan and the risk section are the parts that turn the output from opinion into a testable hypothesis. A good pricing strategy is one you can validate in the market, and the prompt builds that in.
A Worked Example
Let me show you what the output looks like in practice. Suppose I'm a B2B SaaS company selling a workflow automation tool. I fill in the context as follows:
Product: workflow automation for mid-market finance teams
Customer: finance directors at companies with 200–2,000 employees
Current price: $2,400/month
Competitors: Tool A at $1,800, Tool B at $3,200
Value props: saves ~40 hours/week of manual reconciliation; audit-ready reports; no engineering time needed
Volume: 120 customers
Goal: grow ARR by 30% in 12 months without cutting price
The model returns an analysis that looks roughly like this:
Value-Based Estimate. If the tool saves 40 hours/week at a fully loaded cost of $65/hour, that is $26,000/week or about $1.36M/year in labor cost avoided. A customer who captures even 5% of that is getting $68,000/year in value. A price of $3,000/month ($36,000/year) captures roughly 5% of that value, which is a reasonable ask for a tool that requires no engineering.
Competitive Positioning.
Tool A |â– â– â– â– â– â– â– 1800
You |â– â– â– â– â– â– â– â– â– â– 2400
Tool B |â– â– â– â– â– â– â– â– â– â– â– â– â– â– 3200
Reco |â– â– â– â– â– â– â– â– â– â– â– â– â– 3000You are currently in the middle. Given your value proposition (no engineering time, audit-ready), you sit closer to Tool B in perceived value. Recommend moving to $3,000/month.
Tiers.
Tier | Target | Includes | Target Margin |
|---|---|---|---|
Core | 200–500 emp firms | 10 workflows, email support | 78% |
Pro | 500–1000 emp firms | 50 workflows, audit reports, 24/7 support | 82% |
Enterprise | 1000+ emp firms | Unlimited, SSO, dedicated CSM | 85% |
Psychological Pricing. Use $2,950 instead of $3,000 for the Pro tier. The charm price reduces perceived cost by a small amount but the anchor against Tool B at $3,200 makes $2,950 look like a bargain.
Experiment. Run a 4-week test with 20 new prospects. Show them the old price ($2,400) and the new price ($2,950) in alternating order. Measure conversion rate. Hypothesis: conversion drops by no more than 8% while average deal size rises 22%.
Risk. (1) If your customers are budget-constrained, the 23% price increase may suppress conversion beyond 8%. Signal: if conversion drops more than 12% in the first two weeks, roll back. (2) If Tool A launches a comparable audit feature, your differentiation weakens. Signal: monitor their changelog weekly.
That is a genuinely useful artifact. You can take the table into a sales training, paste the bar chart into a board deck, and hand the experiment plan to your growth team. The prompt turned a vague strategic question into a concrete, testable plan.
Adapting the Prompt to Your Business
The prompt is a template, not a script. A few adaptations that matter in practice:
If you are a consumer brand, replace the "target customer" field with a demographic or psychographic segment, and ask the model to consider price elasticity by segment. Add a line: "Consider price sensitivity differences between first-time buyers and repeat buyers."
If you are a marketplace, the pricing question is two-sided. Add: "Analyze pricing for both buyer-side and seller-side, and how the two interact."
If you are a professional services firm, the value-based estimate should be framed in hours saved or revenue generated for the client. Adjust the value proposition field to be specific: "saves 15 hours/week of compliance work" rather than "helps with compliance."
If you are early-stage with no competitors named, leave the competitors field as "none directly; closest analogs are [X, Y]." The model will still produce a useful positioning analysis.
If you want a different tone, add a line at the end: "Write in a confident, direct tone suitable for a founder's internal memo." Or "Write in a formal tone suitable for an investor update."
Common Mistakes to Avoid
Three mistakes I see often when people use AI for pricing:
Under-specifying the context. The model can only reason about what you tell it. If you say "I sell software," the output will be generic. If you say "I sell a workflow automation tool for mid-market finance teams that saves 40 hours/week of manual reconciliation," the output is specific. Spend the time to write a good context block.
Accepting the first answer. Run the prompt twice with slightly different context emphasis. Ask for the analysis from the perspective of a CFO in one run and from the perspective of a customer in the next. Compare the two. Disagreements between the two runs are often where the real insight lives.
Skipping the experiment. A pricing recommendation without a test is a hypothesis. The experiment section is the part that separates a thoughtful analysis from a confident opinion. If you skip it, you are guessing.
The Deeper Principle
What this prompt really does is externalize the structure of a pricing decision. A human strategist doesn't sit down and "think about pricing." They work through a checklist: value, competition, segments, psychology, validation, risk. The prompt encodes that checklist and hands it to a model that can execute it quickly and consistently.
That is the general lesson. For any domain where you have a structured reasoning process, you can write a prompt that encodes the steps and lets an AI assistant execute them. Pricing is just one example. The same technique works for financial modeling, product positioning, customer segmentation, and competitive analysis.
You don't need a pricing team to do pricing well. You need a structured question and a model that can answer it. The prompt above is that structured question.
Dr. Julie Williams is a professor of applied machine learning and the author of three books on practical AI for business. This article is for educational purposes and does not constitute financial advice.
