AI Prompts for Competitive Analysis

I sell [PRODUCT] to [BUYER ROLE] at [COMPANY TYPE]. Here is our homepage copy, our last [N] closed-lost reasons, and the search terms we rank for: [PASTE]. Build a competitor set in three tiers, and put each name in exactly one: 1. Direct - sells the same thing to the same buyer 2. Adjacent - solves the same problem a different way 3. Status quo - what they do instead of buying anything For each name, cite which of my inputs put it there. Do not add companies that appear in none of my inputs; list those separately under "worth checking" and say why you thought of them.

Here is the material I have gathered on the [CATEGORY] market: [PASTE]. Tell me what this material supports, and separate it into three lists: - Stated: claims a source in my material makes directly. Quote it and name the source. - Supported: conclusions I can defend from two or more of my sources. Name both. - Unsupported: things I might assume that nothing here establishes. The third list is the one I care about most. Do not fill gaps with what you know about this market from elsewhere.

Using only the pages I have pasted below for [COMPETITOR], extract: - the primary value proposition, in their exact words - the buyer they name, and where they name them - the three benefits they lead with, in the order the page presents them - the objections the copy pre-empts without being asked Quote the exact wording for every item and say which page it came from. Where you cannot find one of these, write "not stated on the pages provided" rather than inferring it. PAGES: [PASTE]

Compare the feature lists I have pasted for [OUR PRODUCT] and [COMPETITORS]. Produce a table with one row per feature and these columns: feature, who states it, exact wording used, and a confidence column with one of EXPLICIT / VAGUE / ABSENT. Rules: - EXPLICIT only when a page names the capability outright. - VAGUE when the page gestures at it ("powerful reporting") without saying what it does. - ABSENT when no provided page mentions it at all. ABSENT means "not in my sources", not "they do not have it" - label it that way in the output. Do not add features that appear in none of the sources. SOURCES: [PASTE]

From the pricing pages I have pasted below, build a comparison of every plan for [COMPETITORS]. One row per plan, with these columns: vendor, plan name, price, billing period the price assumes, unit the price is charged per, minimum commitment, and what moves a buyer to the next tier up. Rules: - If a page shows both annual and monthly prices, give both in separate columns. Never collapse them into one number. - Record any one-time fee, onboarding charge, credit allocation or seat minimum in its own row, not in a footnote. - Where a plan says "contact sales" or "custom", write exactly that. Do not estimate. - End with a list headed "Not stated on these pages" covering anything a buyer would need that the pages do not say. PRICING PAGES (saved [DATE]): [PASTE]

Build a SWOT for [COMPETITOR] using only the sources I have pasted below. Every entry must end with the source it came from, in square brackets. An entry with no source does not go in. Two constraints that matter more than completeness: - Leave a quadrant sparse if my sources do not support it. Four thin quadrants of guesses are worse than two solid ones. - Strengths and weaknesses are about them. Opportunities and threats are about the market acting on them. Do not use the last two quadrants to restate the first two. SOURCES: [PASTE]

From the case studies, customer logos and pricing I have pasted for [COMPETITOR], infer who they are actually built for: company size, industry, buyer role, and the use case they optimise for. For each conclusion, give the evidence behind it and rate it STRONG (several sources agree), WEAK (one source), or INFERRED (my reasoning rather than their statement). Then answer separately: who do these sources suggest they are NOT built for, and what in the material says so? SOURCES: [PASTE]

Here is a technology profile for [COMPETITOR]'s site, their integrations page, and their status page: [PASTE]. Group what this shows into: analytics and attribution, marketing and lifecycle, sales and support, infrastructure, and payments. Then answer the question the list does not: what do these choices imply about how they sell and who they sell to? Separate what the data states from what you are inferring, and name the two or three entries that carry the most weight in your reading. Ignore anything that is present on almost every website, and say which entries you discarded for that reason.

Here is the page text and navigation structure for [COMPETITOR]'s site: [PASTE]. Work through it in this order: 1. What journey does the navigation assume a visitor is on? 2. Which pages exist purely to rank, and which exist to convert? Say what tells you. 3. What is gated, and what does the gate ask for? 4. What claim is repeated most often across pages, in their words? 5. What would a buyer still not know after reading all of this? Answer from the provided text only. If the structure suggests a page you cannot see, name it as missing rather than assuming its content.

Here is an export of [N] reviews of [COMPETITOR], each with its date and star rating: [PASTE]. Code them into recurring themes of praise and of criticism. For each theme give: - the number of reviews mentioning it and the share of the total - two representative quotes, verbatim, with their dates - the date range the theme spans Then answer two things separately: - which themes appear only in reviews older than [DATE], and may describe something already fixed - which criticisms are about the product, and which are about pricing, support or onboarding Do not paraphrase quotes. Do not merge two themes because they feel similar; if you are unsure, keep them apart and say so.

Here are [N] win/loss interviews, each labelled WON or LOST and tagged with the competitor in the deal: [PASTE]. Identify why deals were won and lost against [COMPETITOR], ranked by how often each reason appears. For every reason, split the evidence into two columns: - SAID: the buyer stated this. Quote them. - INFERRED: I am reading it from what they said. Say what you are reading it from. Then flag any reason that appears in both won and lost deals, since those are usually the ones being mislabelled internally.

Here are [N] sales call transcripts with speaker labels: [PASTE]. Find every moment a competitor is named. For each one give: the competitor, who raised them (buyer or seller), the exact words used, what was being discussed immediately before, and how the rep responded. Then summarise: - which competitor is raised most often by buyers rather than by us - the three most common claims buyers repeat about competitors - the moments where a rep had no answer, quoted directly Work only from these transcripts. Do not assess whether the buyers' claims are true.

Here are cancellation notes and exit survey answers for [N] churned accounts: [PASTE]. Separate them into: left for a named competitor, left for an in-house or manual alternative, stopped doing the job entirely, and reason not stated. Report the size of the fourth group prominently; it is usually the largest and is routinely dropped from summaries. For those that named a competitor, give the reason in the customer's words and the account's tenure. Do not infer a competitive loss from a cancellation that names no alternative.

Draft a battlecard for competing against [COMPETITOR], using only the material pasted below. Sections, in this order: 1. How they position themselves, in their words 2. Where they are genuinely strong. Do not soften this. 3. Where our customers and their reviewers say they fall short, with the source 4. Three objections a buyer raises, with a factual response to each 5. When we should walk away from the deal Rules: mark any claim you could not source with [UNSOURCED] so I can cut it. Keep every response factual and non-disparaging - a rep will say these out loud to a buyer who may already like them. Nothing in section 4 may assert what the competitor will do in future. MATERIAL: [PASTE]

Buyers say "[OBJECTION, IN THEIR WORDS]" when comparing us with [COMPETITOR]. Using the material pasted below, draft three responses: 1. One that concedes where they are genuinely better, then reframes 2. One that reframes on a strength we can evidence from my material 3. One that asks a question to surface what is actually behind the objection For each, name the evidence it rests on. Then write the one-sentence version a rep can say on a call without reading from a card. Do not write anything disparaging, and do not claim a capability my material does not support. MATERIAL: [PASTE]

Here is our messaging and that of [COMPETITORS], all taken from live pages: [PASTE]. Produce three lists: - SHARED: claims we all make. Quote one example of each from two different vendors. - CONTESTED: claims where we say opposite things about the same thing. - OURS ALONE: claims only we make, and that nobody else in this set makes. Then tell me honestly which items in the third list are genuinely differentiating and which are category boilerplate that the others simply have not bothered to write down.

I have a call in [TIME] with [ACCOUNT], a [SIZE] [INDUSTRY] company. [COMPETITOR] is in the deal. Here is the account context and our current battlecard: [PASTE]. Give me, in under 200 words total: - the two things about this competitor that matter for THIS account, not in general - the single question most likely to surface whether they are a real contender - the one claim the buyer is most likely to repeat from the competitor, and my answer - what I should not say Prioritise by what is relevant to this account's size and industry. Say which parts of the battlecard you judged irrelevant here and why.

[ACCOUNT] currently uses [COMPETITOR]. Here is what I know about their setup, contract and renewal timing: [PASTE]. Build a displacement plan covering: - the switching costs a buyer would actually face, separated into data, process, integrations and retraining - which of those we can remove, which we can reduce, and which we cannot touch - the timing that makes a switch realistic given the renewal date - the two internal stakeholders most likely to resist, and what each one is protecting - what would make this account not worth pursuing Be explicit about which of these I told you and which you are inferring from the account type.

Here are the changelog entries for [COMPETITOR] between [START] and [END]: [PASTE]. Group them into themes and, for each theme, give the number of entries, the date range, and whether the entries are launches, improvements or maintenance. Then answer: where is their engineering effort concentrated, and what did they NOT ship in this period that the themes would predict? Two rules. Judge investment by the substance of entries, not by their count, since a vendor can post ten small fixes and one platform change in the same week. And do not infer a roadmap: say what the period shows, and say plainly where the pattern is too short to read.

Here are two copies of [COMPETITOR]'s pricing page, the first saved [DATE A] and the second saved [DATE B]: [PASTE]. List every difference, grouped as: - price changes, with old and new figures - packaging changes: a capability moving between tiers, a limit changing, a plan renamed or removed - wording changes that alter what a buyer is promised - changes to trial, minimum seats or billing terms Packaging changes matter most and are the easiest to miss, so check tier by tier rather than reading the page top to bottom. If a feature appears in both copies but under a different plan, that is a packaging change, not a no-change. Say explicitly if nothing changed.

Here are [COMPETITOR]'s open roles, saved [DATE]: [PASTE]. Summarise by function, seniority and location. Then read the postings themselves rather than the titles: quote any line that names a technology, a market, a customer segment or a named process, and say what each one implies. Answer two questions: - what does this hiring pattern suggest they are building or entering? - what would the same pattern look like if they were simply replacing attrition? Tell me which of the two the evidence actually supports, or say it cannot distinguish them.

Here is [COMPANY]'s earnings call transcript for [QUARTER]: [PASTE]. Give me: - the numbers management led with, and any metric they reported last quarter and did not report this one - every mention of competition, quoted, with who asked - the analyst questions management answered indirectly, with the question and the answer side by side - language that changed from the previous quarter, if I have given you both Separate prepared remarks from the Q&A throughout. Prepared remarks are written; the Q&A is not, and they are not equally informative.

Here is [COMPANY]'s [FILING TYPE] for [PERIOD]: [PASTE]. Extract, quoting the filing and giving the section for each: - how they describe their market and who they name as competition - the risk factors that are specific to this company rather than boilerplate - any segment, customer concentration or geographic disclosure - language that changed from the prior filing, if I have provided it Risk factors are mostly standard text. Flag only the ones that name something specific, and say what makes each one specific.

[COMPETITOR] announced a [AMOUNT] [ROUND] round. Here is the announcement, the investor list and any coverage I have: [PASTE]. Tell me: - what they said the money is for, in their words - what the investor list suggests about the direction, and how confident that reading is - what this changes about how they will compete in the next four quarters, and what it does not change - which claims here are the company's own framing rather than independently reported Be specific about the last one. A funding announcement is a marketing document, and the amount raised says nothing about revenue.

Here is everything I collected on [COMPETITORS] between [START] and [END], each item with its date and origin: [PASTE]. Write a digest of at most 300 words for [AUDIENCE]. Order by what changes a decision, not by date. For each item give: what happened, what it means for us, and what if anything we should do. Items where the honest answer to the third is "nothing" still belong in the digest - say "no action" rather than manufacturing one. End with a line naming what you would have expected to see in this period and did not.

Here is ad copy and creative for [COMPETITOR], saved [DATE], with the run dates I could see: [PASTE]. Group the ads by the promise they make and by the audience they address. For each group give the number of variants, how long they have been running, and the exact hook used. Then tell me: which message are they investing behind, judged by variant count and duration rather than by how good the ad is? And which of their ads address an objection rather than a benefit, since those tell you what their buyers push back on. Do not infer budget. The number of variants is not spend.

Turn the analysis below into a one-page brief for [AUDIENCE], who has [TIME] to read it and must decide [DECISION]. Structure: the conclusion first, then what changed, then why it matters to us, then the two or three decisions it implies with a named owner for each. Hard limits: 300 words, no heading longer than six words, and no sentence that would survive unchanged in a brief about a different competitor. Leave out anything that does not bear on the decision named above, and list what you left out at the end so I can check your judgement. ANALYSIS: [PASTE]

Assemble a competitive intelligence report on [COMPETITORS] for [AUDIENCE] from the sources below. Sections: what changed this period, positioning, product, pricing, go-to-market, what customers say, and what we recommend. Every factual claim carries its source and the date that source was captured. Any section where my material is thin gets a line saying what is missing rather than being padded. At the end, list the three things a reader would most want to know that these sources cannot answer. SOURCES: [PASTE]

Here is my competitor set with the evidence I hold on each: [PASTE]. Place them on these two axes: [AXIS 1] and [AXIS 2]. For every placement, state the evidence and how confident it is. Where two companies would land in the same place, say what would separate them and what evidence I would need to do it. Then challenge the axes themselves: do they separate this set usefully, or do they put most companies in one corner? If the second, propose two axes that would do better and say why.

Work only from the material I provide in this conversation. Rules, in priority order: 1. If a fact is not in my material, say "not in the provided sources". Do not supply it from memory. 2. Every factual claim ends with the source it came from, in square brackets. 3. Where two of my sources disagree, show both and say they disagree. Do not resolve it. 4. Quote exact wording for anything a reader might dispute: prices, plan names, capability claims, numbers. 5. Distinguish what a source states from what you conclude. Label conclusions as such. 6. If my material is too thin to answer, say so and name what is missing, rather than answering anyway. Confirm you have read these rules before I paste the material.

You support the competitive intelligence function at [COMPANY], which sells [PRODUCT] to [BUYER]. How you work: - You work from the material provided in the conversation. You do not answer competitor questions from memory, because competitor facts change and your memory has a date on it. - Every factual claim carries its source. Claims without one are labelled [UNSOURCED]. - You separate what a source states from what you infer, always, even when the inference is obvious. - You never assert what a competitor will do next. You describe what they have done and what it would imply if continued. - You are never disparaging about a competitor. Everything you draft may be read aloud to a buyer who likes them. - When asked for something your material cannot support, you say so and name what would be needed. Default output: the answer first, the evidence second, and what is still unknown third.

You are the competitive assistant for [COMPANY]. Your knowledge files contain our battlecards, positioning and competitor profiles, each with the date it was last reviewed. On every answer: 1. Name which knowledge file you used and the date it was last reviewed. 2. If that date is more than [N] days old, say so before answering. 3. If the question is about a competitor with no file, say so. Do not answer from general knowledge. 4. For anything a rep will say to a buyer, give the factual version and flag anything you could not source. You are talking to sales reps between calls. Lead with the answer, keep it under 150 words unless asked for more, and never open by restating the question.

You run on a schedule and report to [AUDIENCE]. Each run, read the inputs I have connected and compare them with the state recorded at the end of your last run. Report only what changed. For each change: what it was, which input showed it, when, and whether it needs a decision. A run with no changes reports "no changes" and stops - do not manufacture an item to fill the report. Rank by whether it changes a decision, never by recency. Two things you must not do: do not assert a change you cannot point to in an input, and do not summarise the same change twice across runs. At the end of every run, record the state you compared against, so the next run has a baseline.