22 Grok DeepSearch Prompts for Research (2026)

Use DeepSearch. Act as a market analyst. Build a market-size briefing for [MARKET / INDUSTRY].
Pull live sources and cite each with a working link. For every figure include the source, publisher, and publication date.
Return a table with columns: metric | value | year | source (primary/secondary) | link.
Cover: current market size (revenue or units), 3-year CAGR, top 3 growth drivers, top 3 headwinds, and 2 credible forecasts.
Distinguish primary sources (company filings, regulators, official statistics) from secondary (analyst notes, press).
Flag any figure where sources disagree by more than 20% and show both numbers. End with a "confidence & gaps" note listing what you could not verify.

Use DeepSearch. Scan the last 12 months for real trends shaping [MARKET / TOPIC].
Retrieve and cite live sources; prefer material published in the last 90 days and label anything older.
For each trend give: one-line claim, 2 supporting sources (primary/secondary, with links), and 1 counter-signal or skeptic source if it exists.
Return as bullets grouped by strength: strong evidence / emerging / weak or hyped.
Explicitly flag trends that are mostly narrative with thin sourcing. Do not present opinion as fact — separate reporting from commentary.

Use DeepSearch. Research who buys in [MARKET] and why, using live cited sources.
Return a structured profile: primary buyer segments, jobs-to-be-done, willingness to pay signals, and the top purchase objections.
Back every claim with a link and mark it primary (surveys, filings, first-party data) or secondary (articles, blogs).
Include 3 direct customer quotes or complaints found on X or forums, each with a link and date.
Flag where the evidence is anecdotal versus statistically representative, and note the sample size when a survey is cited.

Use DeepSearch. Map the current regulatory landscape for [INDUSTRY] in [REGION / COUNTRY].
Cite primary legal and government sources first (statutes, regulator pages, official filings); use news only to explain them.
Return a table: rule / regulation | status (in force, proposed, contested) | effective date | who it affects | source link.
Include 2 pending changes and their expected timeline. Flag any rule where reporting is unclear or sources conflict, and say so plainly.
Note the "as of" date for the whole answer since laws change.

Use DeepSearch. Benchmark pricing and unit economics across the top players in [MARKET].
Pull live pricing pages and cite each; for cost or margin figures, cite filings or credible analyst sources and mark primary/secondary.
Return a comparison table: company | plan/tier | list price | billing basis | notable terms | source link (dated).
Add a short note on typical discounting and any prices that changed in the last 6 months.
Flag any figure that is estimated rather than published, and never invent a number — leave a cell blank and note "not found" instead.

Use DeepSearch. Produce a competitor teardown of [COMPANY].
Retrieve and cite live sources. Return sections: what they sell, target customer, pricing, recent product moves (last 12 months), strengths, weaknesses, and public traction signals.
Mark each fact primary (company site, filings, official posts) or secondary (press, analysts) with a dated link.
Include 3 things people criticize about them, each with a linked source.
Flag rumor versus confirmed. If a claim is only on X or a single blog, label it low-confidence and say so.

Use DeepSearch. Compare [COMPANY A], [COMPANY B], and [COMPANY C] on features and positioning.
Pull live product pages, docs, and recent announcements; cite each cell with a dated link.
Return a matrix: capability | A | B | C, using "yes / no / partial / unclear" and a source link per confirmed cell.
Below the matrix, summarize each player's positioning in one sentence with a supporting quote and link.
Where a feature claim is marketing language with no proof, mark it "claimed, unverified." Do not fill gaps from memory.

Use DeepSearch. Assess momentum for [COMPANY] using live signals.
Retrieve and cite: recent funding rounds (amount, date, investors), headcount or hiring trend, notable exec moves, and product launch cadence.
Prefer primary sources (filings, official announcements, the company's own job pages) over secondary reporting; label each.
Return a timeline table: date | signal | source (primary/secondary) | link.
Flag anything where the only source is a rumor, a single tweet, or an unconfirmed report, and note it clearly.

Use DeepSearch. Analyze customer sentiment about [COMPANY / PRODUCT] over the last 6 months.
Pull live sources: review sites, forums, and X posts. Cite each with a link and date.
Return: top 5 praises and top 5 complaints, each with 2 linked examples and an estimate of how common it is (rough, and say it's rough).
Separate verified issues (acknowledged by the company or widely reported) from isolated anecdotes.
Note any coordinated or suspicious posting patterns you see, and flag sentiment that may be from bots or a single vocal group.

Use DeepSearch. Map how [COMPANY] goes to market.
Cite live sources for each finding. Cover: primary sales motion (self-serve, sales-led, partner), key channels, partnerships, and marketing angles.
Return bullets grouped by channel, each with a dated link and a primary/secondary tag.
Include 2 recent campaigns or launches and how they were received, with sources.
Where a channel is inferred rather than stated, say "inferred" and explain the reasoning. Flag anything you could not confirm.

Use DeepSearch. Fact-check this claim: "[CLAIM]".
Retrieve live sources and cite them. Return a verdict: true / mostly true / mixed / mostly false / false / unverifiable.
Support the verdict with at least 3 independent sources, ranked primary before secondary, each with a dated link.
Show the strongest evidence for AND against. If sources contradict, present both and explain which is more credible and why.
Do not guess — if the evidence is thin, say "unverifiable" and list what would be needed to confirm it.

Use DeepSearch. Trace this statistic to its origin: "[STAT / NUMBER]".
Follow the citation chain back to the primary source (the original study, dataset, or official report), citing each hop with a link.
Return: original source (with date and methodology), what it actually measured, and how the number has been restated or distorted downstream.
Flag if the stat is outdated, taken out of context, or cites a source that does not actually say it.
If you cannot reach a primary source, say so and stop — do not present a secondary source as the origin.

Use DeepSearch. Check X and the live web for this trending claim about [TOPIC]: "[CLAIM]".
Find the earliest sourceable origin of the claim and cite it with a link and timestamp.
Return: who is spreading it, whether credible outlets or primary sources confirm it, and a verdict.
Separate confirmed facts from speculation and from clearly partisan framing.
Flag if the claim appears to be driven by a coordinated push, a misread source, or a satire/parody account. Note when evidence is still developing.

Use DeepSearch. Verify whether [PERSON] actually said: "[QUOTE]".
Search live sources for the original context — transcript, video, or primary reporting — and cite it with a dated link.
Return: verdict (authentic / misquoted / fabricated / unverifiable), the accurate wording if it differs, and the full context.
Prefer primary sources (recordings, official transcripts) over secondary retellings, and label each.
If you can only find the quote in secondary sources that cite each other, flag it as unconfirmed and say so.

Use DeepSearch. Map where sources agree and disagree on [CONTESTED QUESTION].
Retrieve a range of live, credible sources across viewpoints and cite each with a dated link.
Return: points of consensus (with sources), points of genuine disagreement (with sources on each side), and the reason for the disagreement (data, methodology, or interpretation).
Mark each source primary or secondary and note obvious bias where relevant.
Do not manufacture a false balance — if the evidence clearly leans one way, say so and show it.

Use DeepSearch. Write a state-of-the-field review of [TOPIC / RESEARCH AREA].
Retrieve live sources, prioritizing primary literature (peer-reviewed papers, preprints, official reports) over secondary summaries; cite each with a link and year.
Return sections: what is well established, what is actively debated, and what is still open.
For each key finding, cite at least one primary source and note the strength of evidence.
Flag claims that rest on a single study or on non-peer-reviewed work, and separate established results from early or preliminary findings.

Use DeepSearch. Identify the most important papers on [TOPIC].
Return a table: title | authors | year | one-line contribution | why it matters | link.
Cite live sources; prefer the paper's official page or DOI over a blog about it, and verify the title, authors, and year against a primary listing.
Include a mix of foundational and recent (last 2 years) work, and label which is which.
Do not invent citations — if you cannot confirm a paper exists at a real source, drop it. Flag any reference you could only partially verify.

Use DeepSearch. Map the main scholarly debates within [TOPIC].
For each debate, retrieve and cite the leading positions and the strongest primary source behind each; label primary/secondary with dated links.
Return: debate | position A + source | position B + source | current lean of evidence | key open question.
Explain why the disagreement persists (limited data, competing methods, or values).
Be explicit where the evidence is genuinely unsettled, and do not force a winner if the field has not reached one.

Use DeepSearch. Assess the quality of evidence behind the leading claims about [TOPIC].
Retrieve the underlying studies and cite each with a link. For each major claim, note study type, sample size, and known limitations.
Return a table: claim | best supporting study | study type | evidence strength (strong/moderate/weak) | link.
Flag claims that rely on small samples, correlational data presented as causal, or non-replicated results.
Separate what the data can support from what commentators have inferred beyond it.

Use DeepSearch. Build a due-diligence dossier on [COMPANY].
Retrieve and cite live sources. Cover: legal entity and leadership, funding and ownership, products and customers, litigation or regulatory actions, and public controversies.
Prefer primary sources (filings, court records, official registers) over press; label each and include a dated link.
Return a structured report with a clear "verified facts" section and a separate "unconfirmed / needs checking" section.
Flag every gap. Do not speculate about undisclosed matters — note the absence of information rather than filling it in.

Use DeepSearch. Research the professional background and public reputation of [PERSON] in the context of [ROLE / COMPANY].
Retrieve live sources and cite each with a dated link. Cover: verified career history, notable public statements, affiliations, and any controversies or disputes.
Prefer primary sources (official bios, filings, first-person posts, recordings) over secondary; label each.
Return "confirmed" and "reported but unverified" sections. Distinguish the person from others with the same name and note if identity is uncertain.
Stick to professionally relevant, public information and flag anything you cannot corroborate.

Use DeepSearch. Run a red-flag scan on [COMPANY / PERSON] for [PURPOSE, e.g. partnership, investment, hire].
Retrieve live sources and cite each. Look for: lawsuits, regulatory actions, financial distress signals, safety or ethics issues, and reputational risks.
Return a risk table: flag | severity (high/med/low) | evidence source (primary/secondary) | date | link | how confirmed it is.
Separate confirmed issues from allegations and rumors, and label each. Weigh source credibility explicitly.
Where you find nothing, say "no material red flags found in searched sources" rather than implying a clean record.