26 Manus Prompts for Deep Research & Reports (2026)

Goal: Research these [50 companies] in parallel and build one comparison table. Use Wide Research so each company is handled by its own sub-agent.
List: [paste 50 company names or URLs].
Constraints: Use only data published in the last [24 months]. Prefer official sites, filings, and reputable trade press; avoid unsourced aggregators and marketing blogs. Currency [USD], geography [global].
Deliverable: A spreadsheet, one row per company, columns: Company | HQ country | Founded | Employees | Est. revenue | Primary product | Target customer | Pricing model | Funding raised | Notable customers | Source URL.
Audit rules: Cite a source URL in the final column for every row. Leave any cell blank rather than guessing. Tag each revenue and employee figure with confidence High/Med/Low in a parenthetical. Add a sources appendix listing every URL used.

Goal: Compare these [40 products] in the [category] category. Run with Wide Research, one sub-agent per product.
List: [paste product names or URLs].
Constraints: Specs must come from the manufacturer's official product page or spec sheet only; ignore reseller listings. Snapshot pricing as of [today's date], region [US].
Deliverable: A spec matrix, one row per product, columns: Product | Brand | Price | [spec 1] | [spec 2] | [spec 3] | Warranty | Release year | Official page URL.
Audit rules: Every spec cell needs the official URL it came from in the last column. Blank beats guess. Flag any spec you could not confirm as "unverified". Sources appendix at the end.

Goal: Find and profile up to [100] vendors that offer [solution] for [target buyer]. Use Wide Research to scan and enrich in parallel; stop when you have 100 verified rows.
Constraints: Vendors must serve [geography] and be active as of [2026]. Exclude anyone with no working website. Sources: official sites and directories you can open, not paywalled databases.
Deliverable: A spreadsheet, columns: Vendor | Website | Country | Segment served | Core offering | Pricing signal | G2/Trustpilot rating | Contact page URL | Fit note (1 line).
Audit rules: One source URL per row minimum. Leave rating blank if none is published. Tag "Fit note" confidence High/Med/Low. Append a sources appendix and a count of rows dropped and why.

Goal: Build a competitive feature-and-pricing grid for [my product] against these [12 competitors]. Use Wide Research across the competitor list.
List: [paste competitor names/URLs].
Constraints: Use only each competitor's own pricing and product pages, captured as of [today's date]. Note the plan tier you priced. Geography [US], currency [USD].
Deliverable: A grid, one column per competitor plus mine, rows: Entry price | Top tier price | Free plan? | [feature A] | [feature B] | [feature C] | [feature D] | Support channels | Source URL.
Audit rules: Cite the exact pricing page URL per competitor. Mark any feature as Yes/No/Partial and blank if unclear. Confidence tag on every price. Sources appendix.

Goal: Enrich this list of [80 organizations] with firmographic detail. Run Wide Research, one sub-agent per organization.
List: [paste names or domains].
Constraints: Public sources only; last [18 months]. Do not infer employee counts from headcount estimators — use official or filed figures. Geography [global].
Deliverable: A spreadsheet, columns: Organization | Domain | Industry | Sub-sector | Employees | HQ city | Year founded | Recent news headline | News URL | Enriched source URL.
Audit rules: Two columns hold URLs; fill both or leave blank. Tag Employees and Industry with High/Med/Low. No guessing. Sources appendix and a summary line of coverage (rows fully vs partially filled).

Goal: Write an industry landscape report on [industry] in [geography] for [2026]. Use 1.6 Max for depth.
Constraints: Timeframe of evidence: last [24 months]. Sources: official statistics, filings, and reputable trade press; avoid vendor whitepapers as primary evidence. Keep it to [1,500 words] plus tables.
Deliverable: A report with sections: 1) Overview and scope, 2) Market structure and key segments, 3) Leading players (table: Player | Segment | Position | Source), 4) Value chain, 5) Demand drivers, 6) Risks and headwinds, 7) Outlook, 8) Sources appendix.
Audit rules: Cite a URL inline for every factual claim. Leave a figure out rather than estimate it. Tag each forward-looking statement Med or Low confidence. List every source in the appendix.

Goal: Produce a trend report on [topic/sector] covering the [8] most important shifts of the last [12 months].
Constraints: Only include trends with at least [2] independent supporting sources. Evidence window: last [12 months]. Avoid speculation with no citation.
Deliverable: A report where each trend has: Name | What changed | Why it matters | Evidence (2+ URLs) | Who is affected | Confidence (High/Med/Low). End with a "What to watch next" section and a sources appendix.
Audit rules: A trend with fewer than 2 sources is dropped, not padded. No claim without a URL. Blank beats guess on any number. Sources appendix mandatory.

Goal: Estimate the market size for [product/service] in [geography] for [2026], with a transparent method. Use 1.6 Max.
Constraints: Show both a top-down and a bottom-up estimate. Use only sourced inputs; every assumption must trace to a URL or be labelled an assumption. Currency [USD].
Deliverable: A report: 1) Definition and scope, 2) Top-down estimate (inputs table: Input | Value | Source URL), 3) Bottom-up estimate (same table format), 4) Reconciliation and final range, 5) Sensitivities, 6) Sources appendix.
Audit rules: Every input row cites a URL or is flagged "[assumption]". Give the final answer as a range with a confidence tag, not a single false-precise number. No unsourced figures. Sources appendix.

Goal: Scan the regulatory and policy landscape affecting [industry] in [geography] as of [2026].
Constraints: Use primary sources — regulator sites, official gazettes, and filed rules — via the Browser Agent where possible. Cover enacted rules and active proposals from the last [24 months]. Avoid law-firm marketing summaries as primary evidence.
Deliverable: A table: Rule/Proposal | Status (enacted/proposed) | Effective date | What it requires | Who it affects | Penalty/exposure | Official source URL. Plus a 200-word "So what for [my company]" summary.
Audit rules: Every row links the official document. Leave "Effective date" blank if not set. Tag "Who it affects" confidence. Sources appendix of official URLs.

Goal: Analyze buyer segments and demand signals for [product] in [geography] for [2026].
Constraints: Evidence from the last [18 months] only. Use search-trend data, public survey results, review sites, and community discussion you can link. No invented personas.
Deliverable: A report: 1) Segment map (table: Segment | Size signal | Key need | Buying trigger | Objection | Source URL), 2) Demand signals over time, 3) Willingness-to-pay signals, 4) Whitespace, 5) Sources appendix.
Audit rules: Each segment row cites a URL. Mark any signal you inferred as Low confidence. Leave "Size signal" blank if unsupported. Sources appendix.

Goal: Build a deep-dive profile of [company]. Use 1.6 Max and the Browser Agent to pull from official sources.
Constraints: Prefer the company's own site, filings, and press; cross-check any third-party claim. Evidence current as of [2026]. Flag anything older than [24 months].
Deliverable: A report: 1) Snapshot (founded, HQ, size, ownership), 2) Products and business model, 3) Customers and go-to-market, 4) Leadership team, 5) Funding/financial signals, 6) Recent moves (last 12 months), 7) Strengths and risks, 8) Sources appendix.
Audit rules: Every fact carries an inline source URL. Leave a field blank rather than guess. Tag financial and headcount figures High/Med/Low. Sources appendix listing all URLs.

Goal: Summarize the recent financial position and material events of [public company, ticker]. Use 1.6 Max.
Constraints: Use only official filings and investor-relations pages for the last [4 quarters]. Open the source documents with the Browser Agent. Do not restate analyst opinion as fact.
Deliverable: A report: 1) Latest results table (Revenue | YoY % | Net income | Margin | EPS | Source URL per quarter), 2) Balance-sheet highlights, 3) Guidance, 4) Material events / 8-K style item log (Date | Item | Summary | Filing URL), 5) Risk factors summary, 6) Sources appendix.
Audit rules: Every number cites the exact filing URL. Blank beats guess. Tag guidance as forward-looking, Med/Low confidence. Sources appendix of filing links.

Goal: Explain how [technology/approach] works and where it stands in [2026], for a technical-but-time-poor reader. Use 1.6 Max.
Constraints: Cite primary sources — docs, specs, papers, official blogs — not secondhand explainers. Note maturity and known limitations honestly. Evidence window: last [24 months] for state-of-the-art claims.
Deliverable: A report: 1) Plain-language overview, 2) How it works (mechanism), 3) Key implementations/players (table: Name | Approach | Maturity | Source), 4) Trade-offs and limits, 5) Adoption signals, 6) Open questions, 7) Sources appendix.
Audit rules: Every claim about performance or state-of-the-art cites a URL. Mark contested claims Med/Low. Leave a metric blank rather than approximate it. Sources appendix.

Goal: Tear down [product] — capabilities, positioning, and public roadmap signals as of [2026].
Constraints: Use the product's own docs, changelog, pricing, and public announcements; label any inferred roadmap item as inference. Last [12 months] for roadmap signals.
Deliverable: A report: 1) What it does, 2) Feature inventory (table: Feature | Tier | Notes | Doc URL), 3) Pricing and packaging, 4) Recent releases (changelog table with dates + URLs), 5) Roadmap signals (Signal | Source | Confidence), 6) Gaps vs [my product], 7) Sources appendix.
Audit rules: Every feature and release cites a URL. Roadmap items must be tagged Med/Low unless officially announced. Blank beats guess. Sources appendix.

Goal: Write a sourced explainer on [topic] for [audience], neutral and balanced.
Constraints: Present multiple viewpoints where the topic is contested, each with its own citation. Evidence from the last [24 months] for current-state claims; older sources allowed for background if labelled. No editorializing.
Deliverable: A report: 1) The short answer, 2) Background, 3) How it works / what's happening, 4) Different perspectives (each with source), 5) What's still uncertain, 6) Further reading, 7) Sources appendix.
Audit rules: Every substantive claim cites a URL. Contested points show at least two sides with sources. Tag uncertain areas Low confidence. Sources appendix.

Goal: Synthesize the recent literature on [research question]. Use 1.6 Max.
Constraints: Include only peer-reviewed papers or official reports from [year range]. Prefer primary studies over reviews. Minimum [10] sources. Note sample sizes and methods where available.
Deliverable: A synthesis: 1) Question and scope, 2) What the evidence agrees on, 3) Where it conflicts, 4) Evidence table (Study | Year | Method | Sample | Key finding | Link), 5) Gaps and open questions, 6) Sources appendix.
Audit rules: Every finding links its study. Leave method/sample blank if not reported — do not invent. Tag the overall strength of each conclusion High/Med/Low. Sources appendix with full citations.

Goal: Build an evidence table for the claim(s): [claim(s) to test].
Constraints: For each claim, gather supporting and contradicting evidence. Use only linkable sources; note the source type (study, filing, news, expert). Evidence window: [year range].
Deliverable: A table, one row per piece of evidence: Claim | Evidence summary | Supports/Contradicts | Source type | Source URL | Strength (High/Med/Low). Follow with a one-line verdict per claim and a confidence tag.
Audit rules: No row without a URL. Contradicting evidence must be included where it exists, not omitted. Verdict confidence must reflect the balance of evidence. Sources appendix.

Goal: Run a transparent, systematic-style screen of sources on [topic/question].
Constraints: State inclusion criteria [criteria] and exclusion criteria [criteria] up front. Search [named sources/databases]. Date range [year range]. Log every source you consider, kept or dropped.
Deliverable: 1) Criteria and search terms used, 2) Screening log (Source | Kept/Dropped | Reason | URL), 3) Included-sources summary table, 4) Synthesis of included evidence, 5) Sources appendix.
Audit rules: Every dropped source records a reason. No included source without a URL. Report counts (screened / included / excluded). Tag the synthesis confidence. Sources appendix.

Goal: Fact-check this document/set of claims: [paste claims or link]. Use the Browser Agent to check primary sources.
Constraints: Verify each claim against at least one independent, linkable source. Do not rely on the original document's own citations alone. Note when a claim is unverifiable.
Deliverable: A table: Claim | Verdict (Supported/Partly/Unsupported/Unverifiable) | Best evidence | Independent source URL | Confidence (High/Med/Low) | Note. Plus a summary of how many claims held up.
Audit rules: Every verdict cites an independent URL or is marked Unverifiable. No verdict without evidence. Leave notes blank rather than speculate. Sources appendix.

Goal: Produce a meta-summary of what [N] studies collectively say about [question].
Constraints: Only studies you can link and that report a measurable outcome. Note effect direction and, where reported, size. Flag studies with conflicts of interest or tiny samples. Range [year range].
Deliverable: 1) Bottom line (1 paragraph with overall confidence), 2) Outcome table (Study | Outcome measured | Direction | Effect size if reported | Sample | Caveat | URL), 3) Where studies disagree and why, 4) Sources appendix.
Audit rules: Blank beats guess on effect size and sample. Include contradicting studies. Overall confidence must reflect quality, not just count. Sources appendix.

Goal: Write a one-page executive brief on [topic/decision] for [audience].
Constraints: Strictly one page. Lead with the answer. Evidence from the last [12 months]; every claim sourced. No background padding.
Deliverable: A brief: 1) Bottom line (2–3 sentences), 2) Why it matters, 3) Three key facts (each with a source URL), 4) Options / recommendation, 5) Risks, 6) Sources appendix (can overflow to a second page).
Audit rules: Every fact on the page carries a source URL. Leave a number out rather than approximate. Tag the recommendation's confidence High/Med/Low. Sources appendix.

Goal: Write a decision memo recommending one of these options: [option A], [option B], [option C], for [the decision]. Use 1.6 Max.
Constraints: Score against weighted criteria: [cost 40%], [speed 30%], [risk 30%] (adjust to sum to 100%). Every score cites evidence. State assumptions explicitly.
Deliverable: A memo: 1) Recommendation up front, 2) Decision and context, 3) Criteria and weights, 4) Scoring table (Option | [criterion scores] | Weighted total | Evidence URLs), 5) Trade-offs, 6) Risks and mitigations, 7) Sources appendix.
Audit rules: Every score links evidence or is labelled "[assumption]". Do not fabricate figures. Tag the final recommendation's confidence. Sources appendix.

Goal: Run commercial due diligence on [target company] for [purpose: acquisition/partnership/vendor]. Use 1.6 Max and the Browser Agent.
Constraints: Verify claims against primary sources (filings, official site, court/regulator records where public). Flag red flags explicitly. Evidence current as of [2026]; flag anything stale.
Deliverable: A report: 1) Executive summary and go/no-go lean, 2) Business overview, 3) Financial signals (with URLs), 4) Customers and concentration, 5) Legal/regulatory checks, 6) Key risks and red flags (table: Risk | Evidence | Severity | URL), 7) Open questions to verify offline, 8) Sources appendix.
Audit rules: Every claim cites a URL. Unverifiable items go in "open questions", not the body as fact. Tag each risk severity and confidence. Sources appendix.

Goal: Write a board-ready situation memo on [issue/event] and its impact on [my company].
Constraints: Neutral tone, decision-oriented. Distinguish facts (sourced) from analysis (labelled). Evidence window: last [3 months] for the event, longer for context. No speculation presented as fact.
Deliverable: A memo: 1) Situation in brief, 2) What we know (facts, each with URL), 3) What it means for us (analysis), 4) Options with pros/cons, 5) Recommendation, 6) What to watch, 7) Sources appendix.
Audit rules: Every "what we know" line cites a URL. Analysis is clearly marked as analysis. Tag the recommendation confidence High/Med/Low. Sources appendix.

Goal: Compile a research digest on [theme] covering the most important developments this [period].
Constraints: Only items published in [period] with a linkable source. Group by sub-theme. Cap at [12] items so it stays scannable. No item without a source.
Deliverable: A digest: for each item — Headline | 2-line summary | Why it matters | Source URL | Confidence. Grouped under sub-theme headers, with a 3-bullet "top takeaways" at the top and a full sources appendix at the end.
Audit rules: Every item cites a URL. Drop rumor-only items or tag them Low confidence and label "unconfirmed". Sources appendix listing all links.

Goal: Set up a Scheduled Task that runs every [Monday at 8am] to monitor [topic/competitors/keywords] and deliver a fresh research digest.
Constraints: Each run covers only the last [7 days]. Sources: [named sites/feeds]. Skip anything already reported in a prior run. Every included item must be linkable.
Deliverable: A recurring report emailed/saved each week: 1) 3-bullet summary of the week, 2) New items table (Item | Summary | Why it matters | Source URL | Confidence), 3) "No change" note if nothing new, 4) Sources appendix. Save the recurring format as a Playbook.
Audit rules: Every item cites a URL. Never repeat a prior week's item. Tag unconfirmed items Low. Sources appendix each run.