8 AI agent prompts for Research
Copy-paste, beginner-friendly prompts to help you get cited market, vendor and customer research in minutes. Each one is based on a real-world example and works with ChatGPT, Claude or an AI agent. 5 are free below, and 3 more are in AgentPlaybook Pro.
Free Research prompts
Market size & trend report
Get a cited overview of a market before you commit.
Do a research report on [market] in [region]: size estimates (with sources), growth trends, top 5 players, customer pain points, and 3 underserved niches. Cite every number.
Compare software tools
Pick the right tool with a cited comparison.
Compare the top five [category] tools for a [team size] team. Table: pricing, key features, ease of use, integrations, best for. Then recommend one for my situation: [describe].
Interview insights synthesizer
Turn customer call transcripts into themes and quotes.
Here are customer interview transcripts: [paste]. Extract the top 5 pain points with supporting quotes, how often each appears, and what feature ideas they suggest. Challenge any assumption I have: [your assumption].
Stress-test a technical document before you share it
Have an agent check a paper, report or deck for unsupported claims, math or logic errors, and gaps in evidence before reviewers see it, like Google's Deep Review, which gave pre-submission feedback on 16,000+ manuscripts at STOC, ICML, NeurIPS and ICLR before launching publicly for docs and presentations.
Review this document before I share it: [paste or attach the paper, report, or deck]. Audience: [reviewers / investors / my manager / clients]. Go section by section and list: (1) every key claim and whether the document actually supports it (supported / weak / unsupported), (2) any math, numbers, or logic that doesn't add up, with where it is, (3) missing evidence, comparisons, or caveats a sharp reader would ask about, (4) anything confusing or contradictory. Rank issues by how much they'd hurt credibility, and suggest a concrete fix for the top 5. Don't rewrite the document; just give me the review.
Run a swarm of experiment agents on one hard question
Split a research or optimization question into many experiments, let agents run them in parallel for hours, and have them design follow-up experiments from the results. Third Moment Research runs 10,000 agents at once on 36 to 48 hour investigations with OpenAI's Codex, and one run cut a backtest from about three hours to a few minutes.
You are the lead researcher coordinating a batch of experiment agents. Question: [e.g. why is our backtest slow / which pricing page converts best]. Data and code you may use: [describe]. Budget: [number of parallel runs, time limit]. Step 1: list 5-10 hypotheses, each with a single experiment that could confirm or kill it and the metric to measure. Step 2: write a self-contained brief for each experiment agent (inputs, exact steps, what to report). Step 3: after I paste back results, rank the hypotheses, explain what the evidence shows, and propose the next round of follow-up experiments. Keep a running log of what was tried and ruled out so no agent repeats work.
More in AgentPlaybook Pro
- Vendor shortlistCompare suppliers against your requirements.
- Daily business briefingStart each day with a short summary of what matters.
- Split a big analysis across many agents, then combine their findingsWhen a dataset is too big for one careful read, give each slice (one product, region, customer segment or experiment) to its own agent, have each write a short evidence-linked report, then have one more agent combine them into ranked findings. OpenAI, PerturbAI and Scripps Research did this on Oct 21, 2026 with nearly 2,000 parallel agents, one per gene perturbation in a whole-brain CRISPR atlas; the run took 32 minutes, recovered 48 of 55 testable findings from the original study, and surfaced new leads that experts ranked as promising.
32 copy-paste agent prompts (PDF) plus 5 new prompts every week.
Get it free