When to Use This Prompt

Use this prompt when you need competitive intelligence on a market or niche before making a strategic decision, you want to understand customer language and unmet needs from real sources, you need TAM/SAM/SOM estimates with cited sources, you want a research brief you can share with stakeholders or investors, or you need to spot where your internal data conflicts with market data.

AI Market Research — Master Prompt

A structured, four-phase prompt that turns any AI with web search into a professional market research analyst. It retrieves hard data, audits its own sources, interviews you for internal context, then synthesizes everything into a stakeholder-ready strategic brief.


Requirements

RequirementDetails
AI modelMust have web search enabled (Claude with web search, ChatGPT with browsing, Perplexity, etc.)
Your inputDomain knowledge for Phase 3 interview. The more specific you are, the better the synthesis.
Time15-30 minutes for the full workflow. Phase 1 (retrieval) takes the longest.

The Complete Prompt

Copy everything below and paste it into a new AI session with web search enabled:

You are conducting professional market research. You have web search available.
Your job is to complete a full research-to-strategy workflow autonomously,
asking me for input only when you genuinely need it.

RESEARCH TARGET: [describe your market, niche, company, or decision here]
DECISION I NEED TO MAKE: [what are you trying to decide or understand]

---

PHASE 1 — RETRIEVAL (act like a search engine, not a reasoning engine)

Run targeted web searches to gather hard data. For each search:
- Be surgical: 3-6 word queries beat verbose ones
- Target specific source types: analyst reports, G2/Capterra reviews,
  Reddit threads, SEC filings, industry publications, earnings calls
- Run multiple searches from different angles — do NOT rely on one query
- For any statistic or claim, note the source and date
- Explicitly seek out CONFLICTING data, not just confirming data

Gather data across these areas (skip any not relevant):
1. Top 8-10 competitors: positioning, pricing, recent moves, stated weaknesses
2. Customer language: exact words buyers use about frustrations, fears, desires
3. Market size: TAM/SAM/SOM estimates with source dates and methodology
4. Trends: technologies gaining traction, behaviour shifts, expert predictions

Do NOT analyse yet. Just retrieve and document with sources.

---

PHASE 2 — QUALITY CHECK (audit your own retrieval)

Before moving to synthesis, self-audit what you found:
- Flag any claim supported by only a single source
- Flag any statistic older than 18 months
- Flag any source that is a company blog, press release, or AI content farm
- Identify the 3-5 claims you are LEAST confident in
- Rate overall research confidence: High / Medium / Low
- List what you could not find that would strengthen the research

Show the audit summary before proceeding.

---

PHASE 3 — INTERNAL DATA COLLECTION

Ask these questions one by one (do not ask all at once):
1. What do your customers say when they explain WHY they needed a solution?
2. What is the most common objection you hear, and what sits behind it?
3. What do your best customers wish existed that doesn't yet?
4. Where do deals stall or fall apart — and what's the real reason?
5. What have you tried that didn't work, and what did you learn?

After each answer, acknowledge what's useful and ask the next.

---

PHASE 4 — SYNTHESIS (now reason, not retrieve)

Combine external research + internal data into a strategic brief:

1. HIGH-CONFIDENCE INSIGHTS — claims supported by BOTH external and internal data
2. CONFLICTS AND TENSIONS — where external and internal contradict. Explain the gap.
3. HIDDEN ADVANTAGES — what internal data reveals that research cannot capture
4. THE 3 STRATEGIC IMPLICATIONS — highest-priority actions with reasoning and risk
5. REMAINING UNKNOWNS — what we still don't know, and fastest way to find out

Flag confidence level on each major conclusion.
If uncertain about something, say so explicitly.

How the Phases Work

### Phase 1: Retrieval The AI runs dozens of targeted web searches across competitors, review sites (G2, Capterra), Reddit threads, analyst reports, SEC filings, and industry publications. It gathers data on competitors (positioning, pricing, recent moves), customer language (exact words from real users), market size (TAM/SAM/SOM with sources), and trends (last 6-12 months). It does NOT analyze yet — pure data gathering. ### Phase 2: Quality Check The AI audits its own research before moving forward. It flags single-source claims, statistics older than 18 months, company blogs or press releases used as evidence, and AI content farms. It identifies the 3-5 least confident claims and rates overall research confidence. It shows you this audit so you can request deeper searches where needed. ### Phase 3: Internal Data Collection The AI interviews you with 5 targeted questions about your business reality: customer language, common objections, unmet needs, deal-breakers, and failed experiments. Questions are asked one at a time so each answer can inform the next question. This phase is where the real differentiation happens — it combines your internal knowledge with external data. ### Phase 4: Synthesis The AI combines external research and internal data into a strategic brief with five sections: high-confidence insights (backed by both sources), conflicts and tensions (where data contradicts), hidden advantages (internal-only insights), three strategic implications (with reasoning and risk assessment), and remaining unknowns (with fastest path to resolution). Every conclusion includes a confidence rating.

Usage Tips

Fill in the placeholders. Replace [describe your market] and [what are you trying to decide] with specific, concrete descriptions. “AI writing assistant market for B2B SaaS companies in North America” beats “AI writing tools.” The more precise your target, the more useful the research.
Use a model with strong web search. This prompt is designed for models with real-time web access. Perplexity Pro, Claude with web search, or ChatGPT with browsing all work well. Without web search, Phase 1 falls back to training data and loses most of its value.
Be brutally honest in Phase 3. The synthesis quality depends on the contrast between external data and your internal reality. If you give polished marketing answers instead of honest ones, you get generic recommendations back. The messy truth produces the best strategic insights.
Limitations. The AI’s web search has access limitations — some analyst reports are paywalled, some review sites block automated access, and very niche markets may have limited public data. The Quality Check phase (Phase 2) is designed to surface these gaps so you know where to supplement with your own research.