Market Sizing (TAM/SAM/SOM) AI Prompts for Founders

Quick Answer

We define market sizing as the critical litmus test for venture-scale potential, moving beyond vague top-down claims to rigorous bottom-up analysis. This guide explains the TAM, SAM, and SOM framework and provides specific AI prompts to help founders build defensible models. Our goal is to equip you with the tools to prove your market opportunity to investors using modern LLMs.

Key Specifications

Why Market Sizing is the Cornerstone of Your Startup Pitch

What’s the first question a venture capitalist asks after hearing your pitch? It’s almost never about your product’s slick UI or your team’s passion. It’s a blunt, revealing question: “How big is the market?” Your answer to this determines whether the conversation continues or ends with a polite nod. VCs use market size as their primary litmus test, filtering out thousands of good ideas to find the few that can generate venture-scale returns. A vague, top-down claim like “the global education market is a trillion-dollar industry” is an immediate red flag. It signals a lack of rigor. The gold standard is a bottom-up analysis—a defensible, granular calculation that proves you’ve done your homework and understand the path to capturing a real slice of that opportunity.

To build that defensible case, you need to master the language of market sizing. Think of it as a funnel, filtering potential from broad to specific:

• Total Addressable Market (TAM): This is the entire universe of potential customers for your solution. It’s the total global demand if you had zero competitors and infinite resources.
• Serviceable Available Market (SAM): This is the segment of the TAM you can realistically target, constrained by your business model, geography, or product specialization. It’s the part of the universe you can actually reach.
• Serviceable Obtainable Market (SOM): This is the portion of the SAM you can realistically capture in the short-to-medium term, considering your competition and execution capabilities. This is your near-term target and the number that investors use to model your first few years of revenue.

This is where modern founders have a distinct advantage. In 2025, you don’t need an expensive consulting firm to build these models. Large Language Models (LLMs) can act as a powerful research assistant, helping you synthesize disparate data points, identify non-obvious market proxies, and structure your initial calculations. They can help you find the right industry reports, brainstorm potential customer segments, and even suggest the data sources you’ll need to validate your assumptions.

In the following sections, we’ll move from theory to practice. We’ll first cover the fundamentals of prompt engineering for market research, then dive into specific case studies showing how to build a bottom-up model from scratch, and finally, provide you with a library of battle-tested prompts you can use to immediately start sizing your own market.

The Anatomy of Market Sizing: TAM, SAM, and SOM Explained

Before you can build a financial model or draft a single line of code, you need to answer a fundamental question: how big is the prize? For founders, this isn’t just an academic exercise; it’s the bedrock of your entire business case. Investors use market sizing to determine whether your ambition is grounded in reality or floating in wishful thinking. Getting these numbers right demonstrates that you’ve done your homework, understand your customer, and have a viable path to scale. It’s the difference between a hobby and a venture-backed company.

Total Addressable Market (TAM): The Global Opportunity

Your Total Addressable Market (TAM) represents the total global demand for your product or service if you were the only provider and could achieve 100% market share. It’s the theoretical upper limit of your business, a number that defines the sheer scale of the problem you’re solving. Calculating TAM forces you to think broadly about your potential customer base and the true scope of your vision.

To calculate TAM, you must identify every single potential customer on the planet. For a B2B SaaS product, this might mean identifying every company in a specific vertical that could ever benefit from your solution. For a consumer product, it’s the total population of a specific demographic with the relevant pain point. A common mistake is to confuse the number of customers with the value of the market. A more robust TAM calculation multiplies the potential number of customers by the Annual Contract Value (ACV) or Lifetime Value (LTV) per customer. For example, if you’re building a tool for freelance graphic designers, your TAM isn’t just “all freelance graphic designers.” It’s the number of those designers multiplied by the price you’d charge them per year.

Serviceable Available Market (SAM): Your Realistic Reach

The Serviceable Available Market (SAM) is the segment of the TAM you can actually target with your current business model, product, and geographic focus. It’s a crucial filter that brings your massive TAM into a more manageable and relevant perspective. Your SAM is defined by the constraints of your initial strategy. Are you launching in the US only? Then your SAM is limited by US customers. Is your product built for iOS? Your SAM excludes Android and desktop users.

This is where you must be brutally honest about your limitations. Geographic restrictions are obvious, but regulatory hurdles are just as important. If you’re in fintech or healthtech, your SAM might be drastically smaller due to compliance requirements in different regions. Customer demographics also play a key role; a luxury product will have a smaller SAM than its mass-market equivalent. Defining your SAM shows investors you understand the specific battlefield you’re entering, not just the theoretical war. It proves you have a clear go-to-market strategy for your initial launch.

Serviceable Obtainable Market (SOM): Your Near-Term Target

The Serviceable Obtainable Market (SOM) is the portion of the SAM you can realistically capture in the next three to five years. This is your near-term revenue target, the number that will drive your sales forecasts and operational planning. While TAM shows your ambition and SAM shows your focus, SOM demonstrates your pragmatism and grasp of execution. It answers the critical question: “How much of this market can you actually win, and how quickly?”

Calculating your SOM requires a deep dive into your competitive landscape and internal capabilities. You must factor in the market share of established competitors, your planned marketing and sales budget, and your operational capacity to onboard and support new customers. For instance, if your SAM is 50,000 potential customers, capturing even 1% in year three might be an aggressive but achievable goal, depending on your competition and sales cycle. This is the number that investors will use to sanity-check your revenue projections for your Series A and B funding rounds.

Golden Nugget: The most common mistake founders make is presenting a massive TAM ($10B+) with a tiny, almost accidental SOM. A powerful approach is to reverse-engineer your SOM first. Start with a realistic first-year revenue target (e.g., $1M) and work backward to determine how many customers you need to hit it. This “ground truth” number then becomes the foundation for a defensible, bottom-up model that investors will respect.

The “Bottom-Up” vs. “Top-Down” Debate

This is where founders often lose credibility. A “Top-Down” market sizing is an approach where you start with a large, industry-wide number and try to carve out a slice. You’ve seen it a thousand times: “The logistics market is $10 trillion. If we capture just 1% of that market, we’ll be a $100 billion company.” Investors immediately distrust this model because it’s lazy and aspirational, not analytical. It provides no evidence of how you will acquire that 1%, ignores competition, and is easily dismissed.

The “Bottom-Up” approach is the gold standard. It builds your market size from the ground level, using real-world data and assumptions you can defend. It’s a granular process that demonstrates a deep understanding of your sales funnel and unit economics. The formula is simple but powerful: (Number of Target Customers) x (Annual Contract Value) = TAM/SAM.

Consider this example:

• Top-Down: “The market for project management software for architects is huge.” (Investor yawns).
• Bottom-Up: “There are 45,000 architecture firms in the US. Our target is small-to-medium firms with 5-20 employees, which is about 60% of the total (27,000 firms). Our ACV is $1,200/year. Our SAM for the US is therefore $32.4 million. In year three, capturing just 5% of this focused SAM is a $1.6M revenue business.” (Investor leans in).

The bottom-up method proves you’ve thought through the “how.” It shows you know who your customer is, what you’ll charge them, and how you’ll reach them. It’s a model built on assumptions you can test and validate, making it a far more trustworthy foundation for your startup’s future.

Mastering the Art of AI Prompt Engineering for Data Extraction

An AI model is a brilliant analyst who has never left the library. It has access to nearly the entire internet, but it lacks real-world context, critical judgment, and the ability to challenge its own assumptions. Simply asking “What is the market size for eco-friendly pet food?” will return a generic, surface-level answer that blends AI-generated fluff with outdated blog posts. To get actionable intelligence, you must become its director, guiding its powerful but naive intellect with surgical precision. This isn’t about finding a magic prompt; it’s about engineering a logical process that forces the AI to think like a seasoned market researcher.

Assign the AI a High-Value Persona

The single most effective technique to elevate AI output is the “Act As” framework. This immediately shifts the model’s default tone from a helpful chatbot to a domain expert. You’re not just asking a question; you’re commissioning a report from a specific professional.

For market sizing, you want the AI to adopt a skeptical, data-driven mindset. Instead of a generic prompt, frame it with authority:

• Weak Prompt: “Find the market size for B2B SaaS for accountants.”
• Strong Prompt: “Act as a Senior Equity Research Analyst at Goldman Sachs specializing in FinTech. Your task is to provide a rigorous, bottom-up market size analysis for B2B SaaS solutions targeting accounting firms with 1-10 employees in North America. Prioritize data from reputable sources like Gartner, IBISWorld, and Statista.”

This simple change instructs the AI to adopt a specific persona, prioritize high-quality sources, and deliver its findings with the analytical rigor you’d expect from a top-tier analyst. The output will be more structured, use more precise language, and focus on relevant data points rather than generic marketing copy.

Decompose the Problem with Chain-of-Thought Prompting

Market sizing is a multi-step calculation. AI models can get confused when asked to perform complex reasoning in a single step. Chain-of-thought prompting is the practice of breaking your request into a sequence of logical, interdependent tasks. You are essentially building a step-by-step research plan for the AI to follow.

This forces the model to “show its work,” making it far easier for you to spot errors in its logic or data. Here’s how you would structure a prompt for a bottom-up market calculation:

“I need you to calculate the Serviceable Obtainable Market (SOM) for a new project management tool for freelance videographers in the US. Follow these steps precisely:

1. Step 1: First, find the estimated number of freelance videographers currently working in the United States. Cite your primary source.
2. Step 2: Second, based on industry reports, estimate the average annual revenue per freelance videographer.
3. Step 3: Third, research and determine a realistic percentage of that revenue that a freelancer might spend on project management software (e.g., 1-3%). Justify your assumption.
4. Step 4: Finally, multiply the results from Step 1, 2, and 3 to arrive at the total addressable market. Present the final calculation clearly.”

By breaking it down, you prevent the AI from making a single, large leap of logic. You can verify each step, ensuring the foundation of your market size is built on solid data. This is a golden nugget of prompt engineering: the closer you force the AI to mimic a human analyst’s step-by-step process, the more reliable its final output will be.

Demand Sources and Suggest Proxies for Missing Data

In market sizing, you will frequently hit walls where direct data doesn’t exist. A naive AI might simply invent a number or state “data not available.” A well-prompted AI can be instructed to find creative proxy metrics and, crucially, cite its sources.

Your prompt should explicitly demand transparency:

“For any data point that is not publicly available, do not guess. Instead, propose a logical proxy metric and explain how you derived it. For example, if you cannot find the number of ‘doggy daycares in Chicago,’ you could use the number of ‘dog walkers’ or ‘pet sitters’ listed on LinkedIn and apply a conversion ratio based on industry reports. Always cite the source for your original data and the logic for your proxy.”

This instruction transforms the AI from a simple data retriever into a creative problem-solver. It will start suggesting things like:

• “Using the number of new LLC registrations in the ‘childcare services’ sector as a proxy for new daycare centers.”
• “Estimating market demand for a new product by analyzing the volume of ‘help wanted’ ads for specific skills on Indeed.”
• “Using the number of positive reviews on a competitor’s app store listing as a proxy for their market penetration.”

These proxies are invaluable. They may not be perfect, but they provide a directional estimate and demonstrate a resourceful, analytical mindset—the very thing investors look for.

Embrace Iterative Refinement: The AI Output is a Draft

Perhaps the most critical rule for using AI in any professional context is to treat its output as a starting point, not a final answer. The AI’s first draft is a powerful sparring partner, not a certified accountant. Your expertise is the final layer of validation.

The process should look like this:

1. Generate: Run your well-crafted prompt to get a structured draft.
2. Verify: Go to the cited sources. Are they real? Are they recent (post-2023)? Does the AI’s interpretation of the data match the report’s conclusion?
3. Refine: Feed the verified information back to the AI. “Your source from Gartner is excellent, but their 2022 report is outdated. I’ve found their 2024 forecast which projects 15% growth. Please recalculate the SAM using this new figure.”
4. Challenge: Ask the AI to challenge its own assumptions. “Now, list the top three weaknesses in this market sizing model. What external factors could invalidate these numbers?”

This iterative loop is where the real value is created. You combine the AI’s speed and data-processing power with your critical thinking and industry knowledge. This synergy produces a market size that is not just a number, but a defensible, well-researched thesis—the cornerstone of any credible startup pitch.

Bottom-Up Calculation Prompts: The Investor’s Favorite Method

Why do investors consistently favor the bottom-up calculation over the top-down approach? Because it proves you understand the mechanics of your business from the ground up. A top-down model might claim you’re targeting a $50 billion industry, but a bottom-up model shows you’ve done the hard work of identifying exactly who will pay you, how much they’ll pay, and how many you can realistically reach. It’s a demonstration of grit and operational intelligence. I’ve seen founders walk into pitch rooms with slick top-down slides, only to be dismantled by a single question: “But how will you acquire your first 1,000 customers?” A bottom-up model anticipates that question and answers it before it’s even asked.

Calculating Total Addressable Market (TAM): The Ground Floor

This is where you build your case from a single customer up to the global maximum. Forget broad industry reports for a moment. We’re going to find the total number of potential buyers and multiply that by the revenue you can generate from each one. The key is specificity. “Businesses” is a useless segment. “SaaS companies with 10-50 employees in North America” is a target you can work with.

Here is the prompt template to force the AI into this granular, bottom-up mindset:

• Prompt Template: “Act as a market research analyst. First, identify the total number of [Target Customer Segment, e.g., ‘independent coffee shops in the United States’] globally that fit our Ideal Customer Profile. Second, find the average annual spend or revenue per customer for [Problem/Solution, e.g., ‘specialty coffee bean procurement and point-of-sale software’]. Third, multiply these two figures to generate a conservative, a realistic, and an aggressive Total Addressable Market (TAM) estimate for the next 5 years. Provide your sources for both the customer count and the average spend.”

This prompt forces the AI to perform two distinct research tasks before it’s allowed to do the multiplication. This transparency is crucial. It allows you to inspect its work, challenge its assumptions, and inject your own data. For example, you might know from your beta tests that the average spend is actually 20% higher than the industry average because of your premium service tier. You can now override the AI’s generic figure with your own validated data.

Estimating Serviceable Available Market (SAM): Finding Your Beachhead

Your TAM is the entire ocean. Your SAM is the part of the ocean where you can actually fish. This is where you apply constraints. Maybe you’re only launching in English-speaking countries due to your marketing team’s capabilities. Maybe your solution is only relevant for businesses above a certain employee count because smaller ones don’t have the problem you solve. Filtering your TAM to create your SAM shows investors you’re not planning to boil the ocean on day one.

Use these prompts to layer on those critical constraints:

• Prompt for Geographic Constraints: “Take the TAM we just calculated for [Target Customer Segment]. Now, filter this market to include only customers in English-speaking countries (USA, UK, Canada, Australia, New Zealand). Re-calculate the market size based on this geographic constraint and present the resulting SAM.”
• Prompt for Behavioral/Firmographic Constraints: “Refine the SAM for [Target Customer Segment] by applying the following filter: only include businesses with more than 50 employees. Explain why this specific filter makes the market more serviceable for our solution and provide the new SAM figure.”

By using separate prompts for each filter, you maintain clarity and can easily adjust your model. If you decide to launch in Germany first, you can simply change the geographic prompt and regenerate the SAM in seconds.

Determining Serviceable Obtainable Market (SOM): The Reality Check

This is the most critical number in your market sizing exercise. Your SOM is the slice of the SAM you can realistically capture in the short-to-medium term (typically years 1-3). It’s a direct reflection of your go-to-market strategy and competitive landscape. Investors know that capturing even 1% of a massive market is incredibly difficult. Your SOM calculation proves you understand the hurdles of customer acquisition and market penetration.

This prompt forces the AI to be your strategic advisor:

• Prompt: “Based on our SAM of [Insert SAM Figure] for [Target Customer Segment], estimate our Serviceable Obtainable Market (SOM) for Year 1 and Year 3. Factor in the following assumptions: a competitive landscape with [Number] major competitors, a projected market share capture rate of [e.g., 0.5%] in Year 1 growing to [e.g., 2%] by Year 3, and a marketing budget of [e.g., $100,000] per year. Provide a rationale for these market share percentages based on typical acquisition costs and conversion rates in this industry.”

This prompt moves beyond simple math and into strategic modeling. It forces the AI to connect your budget and competitive position to a tangible revenue target. A golden nugget from experience: always ask the AI to provide a rationale for its market share percentages. This often reveals flawed logic, such as assuming a 5% market share in year one without accounting for the sales cycle length in your industry.

Validating with Competitor Data: The Proof is in the Pudding

Your calculations are just theory until they’re grounded in reality. This is where you reverse-engineer the market by looking at established players. If a competitor is doing $50 million in annual revenue and you can find an estimate of their market share, you suddenly have a powerful data point to validate your own TAM/SAM/SOM figures.

This is your “show me the money” prompt:

• Prompt: “Find the most recent annual revenue figures for [Competitor A] and [Competitor B]. If their revenue is not public, search for credible estimates from industry analyst reports or financial news outlets. Next, find their estimated market share for [Specific Niche Market]. Based on their revenue and market share, calculate the implied size of the niche market. Compare this figure to the SAM we calculated earlier and highlight any significant discrepancies.”

This is a powerful trust-building exercise. If your calculated SAM is $200 million, but reverse-engineering a competitor’s numbers suggests the market is only $50 million, you have a major problem to address. It’s far better to discover this discrepancy in your research phase than for an investor to discover it during due diligence. This process of triangulation—using your bottom-up model and competitor data to cross-reference your findings—is what separates a credible pitch from a hopeful fantasy.

Top-Down Validation and Industry Sizing Prompts

How do you know if you’re chasing a billion-dollar idea or a niche hobby? The answer lies in your ability to validate your market from the top down, using credible, third-party data to paint a picture of the entire ocean before you describe your specific boat. This is the first step in building a defensible TAM/SAM/SOM analysis, and it’s where many founders stumble by relying on gut feelings or outdated blog posts. AI can act as your tireless research analyst, but only if you know how to direct its focus.

Sourcing Credible Market Data

Your first task is to find the baseline numbers that anchor your entire market size argument. Relying on a competitor’s self-reported revenue or a niche forum’s speculation is a recipe for disaster. You need data from the giants of industry analysis. Think Gartner for tech, Statista for general data, IBISWorld for specific industries, and the Bureau of Labor Statistics for labor market data. The goal is to find the total revenue or transaction volume for your broad industry category.

This is where prompt precision separates amateur analysis from professional-grade research. A vague prompt will get you a vague, potentially hallucinated answer. You need to be specific, demanding sources and context.

Example Prompt:

“Act as a market research analyst. Locate the most recent market size report for the global [e.g., ‘cloud-based logistics software’] industry. Your task is to provide:

1. The total addressable market (TAM) size in USD for the most recent year available.
2. The projected Compound Annual Growth Rate (CAGR) for the next 5 years.
3. The names of the top 3 market leaders and their combined market share.
4. A direct link or citation to the source report (e.g., Gartner, Statista, MarketsandMarkets).”

This prompt forces the AI to act like a diligent analyst, providing verifiable data points. It prevents the AI from giving you a generic, made-up number and instead grounds its output in reality. A golden nugget here is to always ask for the source link. If the AI can’t provide one, its data is immediately suspect. This simple instruction builds a foundation of trustworthiness for your entire market analysis.

Analyzing Market Tailwinds and Headwinds

A market size number is just a snapshot in time. A truly insightful founder understands the forces that will change that number. Investors are funding your ability to navigate these forces, so you need to demonstrate you’ve identified them. These are your tailwinds (factors accelerating growth) and headwinds (factors creating friction).

Your AI can help brainstorm these, but the key is to frame the prompt around your specific business model. A tailwind for one company could be a headwind for another.

Example Prompt:

“Identify the top 3 tailwinds and 3 headwinds for a direct-to-consumer (DTC) subscription box service in the US market. For each factor, explain why it’s a tailwind or headwind and provide a specific data point or trend from 2024-2025 to support it. Focus on consumer spending habits, supply chain logistics, and digital marketing costs.”

This prompt moves beyond simple identification. It demands justification and current data, pushing the AI to synthesize information rather than just list generic points. You’ll get insights like “Tailwind: Increased consumer comfort with recurring billing (cite a recent survey). Headwind: Rising customer acquisition costs on social media platforms (cite a recent earnings report from Meta).” This level of detail transforms your market analysis from a simple chart into a compelling narrative about the business environment you’re about to enter.

Geographic Expansion Logic

Proving you can dominate one market is good. Showing you have a scalable plan for global domination is even better. Instead of guessing, you can use AI to model your expansion by applying logical multipliers to your known domestic data. This demonstrates strategic thinking and a scalable mindset.

The process involves establishing your home market size and then using economic and demographic data to estimate new territories. For example, if you know your SAM in the United States, you can estimate your potential in Europe by comparing population and GDP per capita.

Example Prompt:

“Our business has a Serviceable Addressable Market (SAM) of $50 million in the United States. We are considering expanding to Germany. Act as a financial modeler. Based on Germany’s population being approximately 24% of the US and its GDP per capita being roughly 85% of the US, create a simple multiplier model to estimate the potential SAM in Germany. Explain your logic step-by-step and provide the final estimated market size figure.”

This prompt forces the AI to show its work, making the logic transparent and defensible. It’s a powerful way to generate a credible first-pass estimate for your financial model. It shows investors you’re not just thinking about your home turf; you’re building a roadmap for international growth, grounded in simple, logical financial reasoning. This is the kind of scalable thinking that gets VCs excited.

Case Study: Sizing the Market for a Niche B2B SaaS Tool

Let’s get specific. You have an idea for a B2B SaaS tool, but the abstract concept of “market size” feels like a homework exercise. It’s not. This is the foundation of your entire fundraising narrative. We’ll walk through a real-world scenario: building a project management and client collaboration tool exclusively for freelance graphic designers in North America and Europe. This is a niche, so the bottom-up approach we’ll use here is critical for getting it right.

Step 1: Defining the Customer (TAM)

Your Total Addressable Market (TAM) is the universe of all potential customers, assuming zero competition and perfect product-market fit. For our niche SaaS tool, we need to estimate the total number of freelance graphic designers in our target regions. A top-down approach using industry reports is a good start, but a bottom-up calculation is often more defensible. We can use AI to help us build this model from the ground up.

Instead of asking for a single number, we guide the AI through a logical sequence. This is where prompt engineering becomes a superpower for founders. You force the AI to show its work, making the final number far more credible.

Prompt Example:

“Act as a market research analyst. I need a bottom-up estimate for the Total Addressable Market (TAM) of my SaaS tool for freelance graphic designers in North America and Europe. Please break down your calculation:

1. Start with the total number of self-employed professionals in the ‘Arts, Design, and Entertainment’ sector in the US and Canada (source: Bureau of Labor Statistics, StatsCan).
2. Apply a percentage (with your rationale) to isolate ‘Graphic and Web Designers’ from that total.
3. Do the same for the EU-27, using Eurostat data for ‘freelance professionals in the arts and creative sector’.
4. Apply a similar filter for ‘graphic designers’.
5. Sum the two figures (North America + Europe) to arrive at a final TAM number. Cite your data sources for each step.”

This prompt prevents the AI from hallucinating a single, unverifiable number. It compels a step-by-step logic that you can audit. You might get an output like: “Based on BLS data showing 1.2M self-employed in the creative sector in the US, and assuming 15% are graphic designers, that’s 180,000. Canada adds another 20,000. Eurostat data suggests 800,000 in the EU, with 12% being graphic designers, adding another 96,000. Total TAM is approximately 296,000 potential users.” Now you have a defensible starting point.

Step 2: Applying Constraints (SAM)

The Total Addressable Market is often too broad. The Serviceable Available Market (SAM) is the segment of the TAM you can actually target with your product and business model. For our tool, which will be a premium subscription, we need to filter out designers who can’t afford it. A designer earning under $50k/year is likely using free tools and isn’t our customer.

This is where you apply real-world business constraints to your data. You’re narrowing the funnel to find your true target audience.

Prompt Example:

“Refine the previous TAM figure of 296,000 freelance graphic designers. Filter this number to find our Serviceable Available Market (SAM). Our criteria are:

1. Designers who earn over $50,000 USD per year.
2. Based on industry data (e.g., Upwork, Freelancers Union reports), what percentage of freelance graphic designers fall into this income bracket?
3. Apply this percentage to the TAM to calculate the SAM.
4. Provide a brief rationale for the percentage you chose.”

The AI might respond with: “Industry reports suggest that approximately 35% of established freelance designers earn over $50k/year. Applying this to the TAM of 296,000 gives us a SAM of approximately 103,600 designers.” This number is far more meaningful. It represents the pool of designers who have both the problem you’re solving and the budget to pay for a solution.

Golden Nugget: When validating your SAM, don’t just rely on income data. A powerful follow-up prompt is to ask the AI to identify “proxy indicators” of affordability. For example: “Identify the top 3 freelance marketplaces where high-earning graphic designers congregate, and list the average project rates on those platforms.” This gives you a secondary data point to confirm your SAM is realistic and helps you know exactly where to advertise.

Step 3: Realistic Capture (SOM)

The Serviceable Obtainable Market (SOM) is the most critical number for your financial projections and Year 1 goals. It’s the fraction of your SAM you can realistically capture, given your resources, competition, and sales cycle. This is where you ground your ambition in operational reality.

To calculate this, you need to model your go-to-market strategy. A common mistake is to claim you’ll capture 10% of your SAM in Year 1—that’s wildly optimistic for a B2B SaaS. A more credible approach is to build your SOM from the bottom up, based on your marketing budget and conversion rates.

Prompt Example:

“Calculate our Year 1 Serviceable Obtainable Market (SOM) based on the following operational plan:

1. SAM: 103,600 designers.
2. Marketing Budget: $120,000 for Year 1.
3. Primary Channel: Content Marketing & SEO.
4. Estimated Cost Per Click (CPC) in this niche: $4.
5. Website Conversion Rate (Visitor to Free Trial): 3%.
6. Free Trial to Paid Conversion Rate: 15%.
7. Calculate the number of new customers we can acquire based on this budget and these funnel metrics. This final number is our SOM.”

The AI will perform the math: “$120,000 budget / $4 CPC = 30,000 clicks. 30,000 clicks * 3% conversion = 900 free trials. 900 trials * 15% conversion = 135 new customers in Year 1.” This is your SOM. It’s not a percentage of the market; it’s a concrete number derived from your actual plan. This is the number you will be held accountable to.

Step 4: The Pitch Deck Narrative

Never present TAM, SAM, and SOM as a dry table of numbers. This is your story of ambition and execution. The sequence you present them in, and the language you use, frames the entire investment opportunity for a founder.

1. Start with the Big Picture (TAM): “We’re tackling a $1.5 Billion market of freelance designers in North America and Europe who are underserved by generic tools.” This shows the scale of the opportunity and gets investors’ attention. It’s the “why this is a big bet.”
2. Narrow to Your Beachhead (SAM): “Our initial focus is the $500 million segment of high-earning designers who can actually afford a premium solution.” This demonstrates focus. You’re not trying to boil the ocean; you’re strategically targeting the most valuable customers first. It shows you understand the market’s nuances.
3. Ground it in Reality (SOM): “In Year 1, with a lean $120k marketing spend, our model shows we can capture 135 customers, generating $162k in ARR. This proves our go-to-market engine works and gives us a clear path to Series A.” This is the most important part. It shows you are a realist, not a dreamer. It proves you’ve done the work to understand your unit economics and can execute. It builds trust.

By presenting your market size this way, you tell a compelling story: you see the massive potential (TAM), you have a smart strategy to win a piece of it (SAM), and you have a grounded, data-backed plan to get started (SOM).

Common Pitfalls and Sanity Checks

You’ve crunched the numbers, run the prompts, and the result is a staggering Total Addressable Market (TAM) of $50 billion. It feels incredible. But here’s the hard-won lesson from years of watching founders pitch: that big number is often a trap. A massive, unsubstantiated market size is one of the quickest ways to lose investor trust. It signals that you haven’t done the rigorous work to understand who your real customers are and what they’re willing to pay. This section is your safeguard. We’ll walk through the most common market-sizing errors and the AI-powered sanity checks that will turn your numbers from a fantasy into a credible forecast.

The “Everyone” Fallacy: From Broad Claims to a Sharpened Niche

The most frequent mistake founders make is defining their target customer as “everyone.” Your AI prompt might return a TAM based on “all smartphone users,” but how many of them need your niche project management tool for remote legal teams? This “everyone” approach creates an inflated, meaningless number. Investors don’t fund markets; they fund your ability to win a specific, defensible piece of a market.

To combat this, you must force your AI co-pilot to apply filters. Instead of a single, broad prompt, break it down. This is a core principle of effective AI prompting for market sizing.

Actionable Prompt Example: “We are building a B2B SaaS platform for project management, specifically designed for remote legal teams. Our initial go-to-market will be in the United States. Start with a broad TAM of ‘all businesses’ and apply the following filters sequentially to arrive at a realistic SOM:

1. Filter for businesses in the US with more than 50 employees.
2. From that group, filter for industries with high project complexity, such as legal, consulting, and architecture firms.
3. From that group, filter for companies that are remote-first or have a significant remote workforce (over 40% of employees).
4. Provide the final number and the data source for each filtering step.”

This forces the AI to show its work and prevents it from making unfounded assumptions. The result is a grounded Serviceable Obtainable Market (SOM) you can defend, not a fantasy TAM you can’t.

Ignoring the “Willingness to Pay”: It’s About Wallets, Not Heads

A classic error is calculating market size based purely on the number of potential users (“heads”) without considering their ability or willingness to pay (“wallets”). A market of 10 million users is worthless if they all expect your product to be free. Market size is a financial metric, not a user count. Your model must be built on revenue, not just reach.

This is where you shift your AI prompts from demographics to economics. You need to validate your pricing assumptions with real-world data.

Actionable Prompt Example: “Analyze the pricing pages of these three competitors in the [e.g., ‘AI-powered graphic design’] space: [Competitor A, B, C]. For each, identify their primary subscription tier price and target customer (e.g., ‘Pro’ plan at $49/month for freelancers). Then, cross-reference this with recent G2 or Capterra reviews to determine if customers complain about the price being too high or if they see it as good value. Summarize the prevailing price sensitivity for this type of tool.”

This prompt helps you triangulate a viable price point. If the AI finds that competitors at the $50/month tier are praised for value, while those at $150/month are consistently reviewed as “too expensive for what you get,” you have a critical data point for your revenue projections. Golden Nugget: A sophisticated founder doesn’t just project revenue based on a price they invent; they anchor it to the market’s demonstrated willingness to pay, creating a far more defensible financial model.

Static vs. Dynamic Markets: The Danger of Outdated Data

The market you’re entering today is not the market it will be in two years. A common pitfall is building your entire model on a single, static industry report from 2023. This ignores critical tailwinds (like the AI boom) and headwinds (like new regulations or market saturation). Investors are backing your vision for the future, so your numbers must reflect a dynamic, forward-looking perspective.

Your AI prompts must be engineered to seek out the most current information and project future trends.

Actionable Prompt Example: “The market for [e.g., ‘corporate wellness apps’] was valued at $X billion in 2023. Provide the most recent growth rate projections (CAGR) for this market from at least two reputable sources (e.g., Gartner, Statista, Forrester) for the next 5 years. Identify and list the top 3 macroeconomic or technological trends driving this growth or potential decline, citing recent news articles or reports from 2024-2025.”

By doing this, you can confidently state: “While the market was $X billion in 2023, our projections are based on the 2025 forecast of $Y billion, driven primarily by the integration of generative AI features, which analysts at [Source] predict will accelerate adoption by 15%.” This demonstrates you understand the market’s trajectory, not just its history.

The “Bottom-Up” Sanity Check: Does Your 5-Year Vision Add Up?

After all the filtering and validation, you need one final, brutal sanity check. This connects your ambitious market share goals to the valuation you’re seeking. It answers the ultimate question: can you realistically capture enough of this market to justify the investment? This is the reverse-engineering test that separates a well-researched plan from a pipe dream.

This final prompt sequence ensures your SOM isn’t just a number, but a strategic path to a specific outcome.

Actionable Prompt Example: “Assume we are seeking a $20 million Series A valuation. A common benchmark is that the investment should represent 20-25% of the company, implying a post-money valuation of ~$100 million. Based on a 5-year revenue projection, what is the required annual recurring revenue (ARR) in Year 5 to justify this valuation (typically using a 5-10x ARR multiple)? Now, take our calculated SAM of [e.g., $500 million] and calculate the precise market share percentage we would need to capture by Year 5 to hit that ARR target, assuming our average customer contract value is [e.g., $10,000/year]. Is this percentage a plausible market share for a new entrant in this space?”

If the AI calculates that you need to capture 15% of the SAM in just five years in a crowded market, you have a red flag. Either your valuation target is too high for this market, or your SAM is too narrow. This process forces you to align your ambition with market reality, ensuring your pitch is both ambitious and, most importantly, achievable.

Conclusion: Turning Data into a Compelling Investment Narrative

You’ve done the heavy lifting. You’ve used AI to run bottom-up calculations, validated them with top-down analysis, and stress-tested your assumptions with sanity checks. The result is a set of TAM, SAM, and SOM figures that are more than just numbers—they are the foundation of your story. But here’s the critical transition: investors don’t invest in spreadsheets; they invest in founders. The data gets you the meeting, but the narrative gets you the check.

The Art of the Numbers

This is where you separate yourself from the pack. While your market sizing is quantitative, your interpretation must be qualitative. An investor sees your SAM of 103,600 designers and immediately asks, “Okay, but how will you reach them?” Your job is to answer that before they ask. The numbers prove the market is big enough; your narrative proves you are the person to capture it.

Golden Nugget: The most powerful phrase in a pitch isn’t “The market is worth $X billion.” It’s “We are going after this specific, defensible wedge of the market, and here is our 3-step plan to dominate it.” This demonstrates focus and a battle-tested go-to-market strategy, not just wishful thinking.

Your AI-driven analysis gives you the “what.” Your experience, insight, and strategy provide the “how” and the “why.” This fusion of data and conviction is what builds trust and signals true expertise.

Your Next Step: From Data to Due Diligence

Don’t let these insights gather dust. The true value of this process is realized when you integrate it into your core financial model and pitch deck.

1. Start with the prompts: Run the bottom-up and top-down calculations for your own venture this week.
2. Iterate relentlessly: Challenge the AI’s assumptions. Refine your inputs. The goal is a model that can withstand the intense scrutiny of due diligence.
3. Build the narrative: Weave these numbers into a compelling story about the opportunity you’ve uncovered and your unique ability to seize it.

By building this rigorous, data-backed foundation, you’re not just preparing a pitch—you’re building the strategic roadmap for your business. Go turn that data into a story that investors can’t ignore.

Expert Insight

Frequently Asked Questions

Q: What is the difference between TAM, SAM, and SOM

TAM is the total global demand, SAM is the segment you can realistically target based on your business model, and SOM is the portion of the SAM you can capture in the short term considering competition

Q: Why do investors care about market sizing

Investors use market sizing to filter for venture-scale returns, assessing if the opportunity is large enough to justify the risk and capital investment

Q: How can AI help with market sizing

AI can act as a research assistant to synthesize data, identify market proxies, brainstorm customer segments, and suggest data sources to validate assumptions