Do Market Research with AI

1. 1

   Define Your Research Questions

   Market research without a clear question is an expensive way to confirm what you already think. Start by defining exactly what you need to know — and what you'll do with the answers.

   ChatGPT Claude

   I need to do market research and want to make sure I'm asking the right questions before I start gathering data.
   
   Business context:
   - Company/project: [description]
   - Stage: [e.g., 'pre-product idea validation' / 'early startup testing go-to-market' / 'established company entering new segment' / 'product expansion decision']
   - The decision I need to make: [what specific business decision will this research inform? e.g., 'whether to build a mobile app' / 'which customer segment to focus on first' / 'how to price a new product tier']
   - What I already believe (my hypotheses): [list your current assumptions — these are what you're trying to validate or invalidate]
   - What I already know (data I have): [any existing data, customer feedback, sales data, web analytics]
   - Timeline: [when do you need this research complete?]
   
   Help me:
   
   1. **Sharpen my research questions**: Turn my decision into 3-5 specific, answerable research questions. E.g., instead of 'is there a market for this?' → 'What is the current spend of [target buyer] on solving [problem], and how satisfied are they with existing solutions?'
   
   2. **Prioritize the questions**: Rank them by importance to the decision. What's the one question that if answered would change my decision most dramatically?
   
   3. **Identify the right research methods**: For each question, what's the most efficient way to get a reliable answer? Options include: secondary research (AI synthesis of public data), customer interviews, surveys, competitor analysis, or A/B tests. Match method to question.
   
   4. **Flag my biases**: Based on my hypotheses, where am I most likely to see what I want to see in the data rather than what's actually there? What should I watch out for?
   
   5. **Research scope decision**: Given my timeline and resources, what can I realistically answer well in [X days]? Help me scope this down to what's most essential.

   Copy

   Tip: The question 'is there a market for this?' is unanswerable. It's not specific enough to research and not specific enough to act on. Every market research question should be specific enough that you can imagine what an answer would look like — a number, a percentage, a clear yes/no, a ranked list. If you can't visualize what the answer looks like, rewrite the question.

2. 2

   Conduct Competitive Landscape Analysis

   Map who's already solving your target customer's problem, how they're doing it, what their weaknesses are, and where you could position differently.

   Perplexity ChatGPT

   Help me build a competitive landscape analysis for [my product/market category].
   
   My product/offer: [brief description]
   Target market: [who you're selling to]
   The problem you solve: [in one sentence]
   
   Conduct a competitive analysis:
   
   1. **Competitor identification**: List all meaningful competitors in these categories:
      a) Direct competitors (solving the same problem for the same buyer in a similar way)
      b) Indirect competitors (solving the same problem differently, or solving a related problem)
      c) Non-consumption alternatives (what does the target buyer do TODAY if no competitor existed? Including doing nothing, DIY, or using a general tool)
   
   2. **Competitor profile** (for each major direct competitor — up to 5):
      - Company name, founding year, funding stage/amount if known
      - Core product description in one sentence
      - Primary target customer
      - Pricing model and approximate pricing
      - Key strengths (what they do well)
      - Known weaknesses (negative reviews, common complaints, unmet needs their customers express)
      - Positioning statement (how they describe themselves)
      - Estimated market share or traction (users, revenue, growth signals)
   
   3. **Competitive positioning map**: Create a 2x2 framework comparing competitors on the 2 dimensions most relevant to my market. Suggest the best 2 dimensions to use for my specific market.
   
   4. **White space analysis**: Based on the competitive map, where is there underserved demand? What customer needs are all competitors partially addressing but none addressing well?
   
   5. **Differentiation opportunity**: Given my strengths and the competitive gaps identified, what's the most defensible positioning angle for me to take? Why would a customer choose me over the established options?

   Copy

   Tip: Read the 2 and 3-star reviews of your competitors' products on G2, Capterra, Amazon, or the App Store. These reviews are gold — they tell you exactly what customers like enough to stick with but are frustrated enough about to complain publicly. That gap between what competitors provide and what customers wish they provided is your product opportunity.

3. 3

   Analyze Your Target Customer

   Build a data-grounded picture of who your target customer is, what they care about, how they make decisions, and what would make them switch. Go beyond demographics into psychology and behavior.

   Perplexity ChatGPT Claude

   Help me build a detailed target customer profile based on research and synthesis.
   
   My product/service: [description]
   Target customer hypothesis: [your current assumption about who they are]
   
   Help me build a comprehensive customer profile:
   
   1. **Demographic and firmographic baseline**: For [target customer type], what do we know about their demographics or company characteristics? (age range, income, geography, job titles, company size, industry — whatever's relevant to my offer)
   
   2. **The Job-To-Be-Done**: What is the underlying 'job' they're hiring a product like mine to do? Use the JTBD framework: when [situation], I want to [motivation], so I can [desired outcome]. E.g., 'When I'm trying to close a month-end report on Friday afternoon, I want to pull accurate data quickly without involving the data team, so I can stop working at a reasonable hour.'
   
   3. **Buying journey**: How does someone in this profile typically:
      - Discover they have the problem I solve?
      - Start looking for solutions?
      - Evaluate options?
      - Make a purchase decision?
      - Who else is involved in the decision (for B2B)?
   
   4. **Language they use**: What words and phrases do people like this use to describe the problem I solve? Where do they talk about it online (subreddits, forums, LinkedIn groups, Slack communities, review sites)? This is research gold — the exact language they use is what I should use in my marketing.
   
   5. **Objections and switching costs**: What would make this buyer NOT buy? What are they afraid of? What would they be giving up by choosing me over their current solution?
   
   6. **Customer segments**: Are there meaningfully different sub-groups within my target market who would need different messaging, products, or go-to-market approaches? What are they?

   Copy

   Tip: The JTBD framework is more useful than traditional demographics because it explains why people buy, not just who they are. Two customers with completely different demographics can have the exact same JTBD — and the same marketing will reach both. Conversely, two customers who look identical demographically might have completely different jobs, and the same marketing will reach neither. Focus on the situation and motivation, not the person.

4. 4

   Size the Market and Estimate Opportunity

   Build a defensible market size estimate using the TAM/SAM/SOM framework — not to convince investors with a big number, but to understand if the business opportunity is worth pursuing.

   Perplexity ChatGPT

   Help me build a market sizing estimate for my opportunity.
   
   My product: [description]
   Target customer: [description]
   Price point: [what you charge or plan to charge]
   Geography: [which markets you're targeting initially]
   
   Build a market size estimate using bottoms-up methodology:
   
   1. **TAM (Total Addressable Market)**: The maximum possible market if every potential buyer purchased. Build this two ways:
      a) Top-down: Find an existing market research estimate for the broadest relevant category, then note where that number comes from and how reliable it is.
      b) Bottom-up: [Number of potential buyers in geography] × [annual spend per buyer] = TAM. Walk me through finding each number using publicly available data.
   
   2. **SAM (Serviceable Addressable Market)**: The portion of TAM you could realistically serve given your product's actual scope, geography, and buyer qualification criteria. What assumptions narrow TAM to SAM?
   
   3. **SOM (Serviceable Obtainable Market)**: Your realistic 3-year target — the piece of SAM you could actually capture given your resources, sales motion, and competitive position. Justify the percentage you're targeting based on comparable company benchmarks.
   
   4. **Sensitivity analysis**: What are the 2-3 assumptions that most affect the market size estimate? Show me a range (conservative / base / optimistic) based on varying these assumptions.
   
   5. **Market sizing red flags**: What are the common mistakes people make in market sizing that make investors and board members skeptical? What should I avoid claiming?
   
   6. **Go/no-go threshold**: For a business of my type and stage, how large does SOM need to be to justify pursuing this opportunity? What's the minimum viable market size for this to be worth building?

   Copy

   Tip: Bottoms-up market sizing is always more credible than top-down. 'The global SaaS market is $200B and we're targeting 1%' is meaningless — that 1% number is invented. 'There are 180,000 companies in the US with 50-500 employees in our target industries (source: LinkedIn), 30% of them have the problem we solve based on our customer research, and they currently spend $500-2,000/year on workarounds — implying a SAM of $27M-$108M' is credible because every number can be sourced and challenged constructively.

5. 5

   Synthesize Insights and Build a Research Summary

   Turn raw research into a clear, decision-ready summary that can be presented to stakeholders or used as the foundation for a go-to-market strategy.

   ChatGPT Claude

   Synthesize all my market research into a decision-ready summary.
   
   Research inputs:
   - Research questions (from Step 1): [paste]
   - Competitive landscape findings (from Step 2): [paste key findings]
   - Target customer profile (from Step 3): [paste key findings]
   - Market sizing (from Step 4): [paste key findings]
   - Any other data or customer interviews I've conducted: [paste or summarize]
   
   Build a research synthesis document:
   
   1. **Executive summary** (one page, ~300 words): The most important 5-6 findings from across all research areas. Written for a busy executive or investor — each point should be a clear insight, not a data dump. Format: Insight statement + 1 sentence of supporting evidence.
   
   2. **Key insights vs. key assumptions**: Separate what the research actually found (insights — evidence-based) from what I'm still assuming (assumptions — need further validation). This honesty prevents acting on beliefs as if they were facts.
   
   3. **Answers to original research questions**: For each question from Step 1, what's the best answer I can give based on the research? Rate each answer: High confidence (multiple sources confirm) / Medium confidence (directional but limited data) / Low confidence (best hypothesis, needs validation).
   
   4. **Strategic implications**: Based on these findings, what are the 3-5 most important implications for my product, marketing, or go-to-market strategy? Don't just summarize the data — tell me what to DO differently based on it.
   
   5. **Research gaps and next steps**: What are the top 3 questions I still can't answer well that would most change my strategy? What's the minimum research needed to answer each, and how long would it take?
   
   6. **One-paragraph POV**: Based on everything, write a clear, opinionated one-paragraph synthesis of the market opportunity — the kind of thing I'd say to a co-founder or investor to describe what I've learned and what I believe.

   Copy

   Tip: The most common mistake in research synthesis is summarizing data without interpreting it. 'Competitor A has 4.2 stars and Competitor B has 3.8 stars' is data. 'Customers of Competitor A consistently praise reliability but complain about onboarding complexity — suggesting a meaningful opportunity for a solution that prioritizes fast time-to-value even at a slight reliability trade-off' is an insight. Insights tell you what to do. Data just tells you what is.

Recommended Tools for This Scenario

Perplexity

Freemium

AI-powered search engine that answers questions with cited sources

• Real-time web search with inline source citations
• Pro Search multi-step deep research automation
• Multiple model options (Sonar, GPT-4o, Claude)

Get Started →

ChatGPT

Freemium

The AI assistant that started the generative AI revolution

• GPT-4o multimodal model with text, vision, and audio
• DALL-E 3 image generation
• Code Interpreter for data analysis and visualization

Get Started →

Claude

Freemium

Anthropic's AI assistant built for thoughtful analysis and safe, nuanced conversations

• 200K token context window for massive document processing
• Artifacts — interactive side-panel for code, docs, and visualizations
• Projects with persistent context and custom instructions

Get Started →