25 ChatGPT Prompts for Analyzing a Competitor’s Website and Marketing Strategy
AI can do much more than write blog posts. With the right instructions, it can help you examine a website, uncover weaknesses in a marketing strategy, understand competitors,…
At a glance
- 30 prompts
- 24 min read
- copied 1 time
AI can do much more than write blog posts.
With the right instructions, it can help you examine a website, uncover weaknesses in a marketing strategy, understand competitors, identify customer questions, evaluate conversion barriers, organize research, and turn scattered information into an actionable plan.
The challenge is getting useful analysis instead of generic advice.
A weak prompt might produce a list of predictable recommendations. A well-designed prompt gives the AI a specific job, enough evidence to work with, clear boundaries, and a defined format for the final answer.
This guide contains 25 practical AI prompts for website analysis, marketing research, competitive analysis, content planning, positioning, and conversion optimization.
Each prompt is designed for a different task rather than repeating the same generic instruction with a different topic.
You can use the prompts with ChatGPT or another capable AI system. Replace the variables with your own information, provide the relevant source material, and review important recommendations before acting on them.
What Makes an AI Prompt Useful for Marketing Analysis?
A useful prompt does not need to be extremely long. It needs to remove ambiguity.
Before asking AI to analyze something, make five things clear:
1. Give the AI a specific job
Instead of:
“Analyze my website.”
Try:
“Evaluate whether the homepage clearly communicates what the company offers to a first-time visitor and identify the three biggest sources of confusion.”
The second instruction gives the model something measurable to investigate.
2. Supply the evidence
AI cannot reliably evaluate information it has never been given.
Depending on the task, your source material might include:
- Website copy
- Competitor pages
- Product descriptions
- Customer reviews
- Search queries
- Sales notes
- Analytics observations
- Existing content
- Pricing information
- Brand guidelines
- Customer interviews
- Research documents
3. Separate facts from interpretation
Tell the AI which information is confirmed and which conclusions are assumptions.
For example:
“Treat the supplied website copy as factual evidence. If you infer something about customer intent, label it as an interpretation.”
This makes the resulting analysis easier to audit.
4. Define what you want back
If you need a table, ask for a table.
If you need prioritized recommendations, specify the prioritization criteria.
If you need an editorial brief, define the fields it should contain.
5. Tell AI what not to do
For business analysis, one of the most useful instructions is:
“If the evidence does not support a conclusion, say that the information is insufficient instead of guessing.”
That single constraint can significantly reduce unsupported recommendations.
A Reusable Prompt Framework
You can use this basic structure when creating your own marketing-analysis prompts:
Fill in the blanks below, or click a highlighted word in the prompt.
Role:
You are helping me with [specific area].
Objective:
I need to [desired outcome].
Business context:
[Explain the company, website, product, audience, market, or situation.]
Evidence:
[Paste the information the AI should analyze.]
Questions to investigate:
[List the specific questions.]
Boundaries:
Use the supplied evidence as the primary source.
Do not invent facts, statistics, customer information, or business claims.
Clearly identify assumptions or areas where evidence is missing.
Output:
Return the analysis as [table/report/checklist/brief/etc.].
Prioritize the most important findings and explain why they matter.
The following prompts take that framework further by adapting the instructions to specific marketing problems.
1. Discover Weaknesses in Website Positioning
A website can look professional while still failing to communicate a compelling reason to choose the business.
Use this prompt when you want AI to examine your positioning from a visitor’s perspective.
What to provide
Give the AI:
- Homepage copy
- Product or service description
- Target audience
- Main competitors
- Current positioning statement
- Any known customer concerns
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Act as a positioning strategist reviewing a website for clarity and differentiation.
Here is the business context:
Business:
[DESCRIBE THE BUSINESS]
Primary audience:
[DESCRIBE THE TARGET CUSTOMER]
Main offer:
[DESCRIBE THE PRODUCT OR SERVICE]
Current positioning:
[PASTE CURRENT POSITIONING]
Website content:
[PASTE RELEVANT WEBSITE COPY]
Evaluate the positioning from the perspective of someone encountering this business for the first time.
Identify:
1. What the business appears to offer.
2. Who the offer appears to be designed for.
3. What problem the business seems to solve.
4. What makes the offer appear different.
5. Which claims are clear and which are vague.
6. Where the messaging could apply equally to many competitors.
7. Important information a visitor may still need.
8. Three opportunities to make the positioning more distinctive.
Separate direct observations from your interpretations.
Do not invent customer research, market statistics, competitive advantages, or company claims that are not present in the supplied material.
Finish with a concise recommended positioning direction.
Example use
A marketing team can provide its homepage and several competitor descriptions and ask AI to identify where its message is interchangeable with the competition.
Check the result
Look for recommendations that are actually supported by the supplied website and business information. Do not automatically accept statements about what “customers want” unless customer evidence was provided.
2. Build a Customer Journey Map
A customer journey analysis helps reveal what someone may need before moving from awareness to action.
Instead of asking AI to create a generic journey, give it the information you already have.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Build a customer journey map using the information below.
Business:
[BUSINESS DESCRIPTION]
Audience:
[TARGET CUSTOMER]
Offer:
[PRODUCT OR SERVICE]
Available customer information:
[REVIEWS / INTERVIEWS / SALES NOTES / WEBSITE DATA]
Map the journey across these stages:
- Problem recognition
- Initial discovery
- Research and comparison
- Evaluation
- Decision
- Purchase or conversion
- Post-purchase experience
- Repeat purchase or advocacy
For every stage, identify:
• likely customer objective
• information they may need
• questions they may ask
• possible hesitation
• relevant website content
• useful conversion action
• evidence currently available
• evidence that is missing
Do not assume a customer behavior is factual unless the supplied material supports it.
Clearly mark hypotheses as hypotheses.
At the end, identify the three stages where better information or messaging could have the greatest potential impact.
Example
A SaaS company could provide sales-call notes, website pages, and customer questions to identify where prospects become uncertain during evaluation.
3. Examine Competitor Messaging
Competitor analysis becomes more useful when AI is asked to distinguish what competitors actually say from what their messaging might imply.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Analyze the messaging used by these competitors.
Competitor A:
[CONTENT]
Competitor B:
[CONTENT]
Competitor C:
[CONTENT]
Our business:
[CONTENT]
Compare the businesses across:
- Primary promise
- Target audience
- Problem emphasized
- Benefits emphasized
- Features emphasized
- Tone
- Proof or trust elements
- Calls to action
- Differentiation claims
- Repeated industry language
Create a comparison table first.
Then identify:
1. Messaging patterns shared by multiple competitors.
2. Claims that appear overused.
3. Areas where our messaging is similar to competitors.
4. Ideas competitors communicate that we do not.
5. Areas that appear underdeveloped across the market.
6. Potential differentiation opportunities.
Only describe something as a competitor claim when it appears in the supplied material.
Do not create competitor facts or infer market leadership without evidence.
Why this works
It forces the model to examine the actual language instead of producing a generic “competitor analysis.”
4. Locate Conversion Friction
A visitor may understand a website but still hesitate to take action.
Use this prompt to investigate potential obstacles.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Act as a conversion-review specialist.
Analyze the following customer journey:
[PASTE JOURNEY OR WEBSITE FLOW]
Primary conversion:
[DEFINE DESIRED ACTION]
Audience:
[DESCRIBE AUDIENCE]
Identify every point where a visitor could reasonably hesitate, become confused, lose confidence, or abandon the process.
For each potential friction point, provide:
- Location in the journey
- What the visitor may be trying to determine
- What information is currently available
- What appears unclear
- Why the uncertainty could matter
- Recommended improvement
- Evidence supporting the recommendation
- Priority: High / Medium / Low
Do not claim that a particular issue causes abandonment unless performance data proves it.
Distinguish observed problems from conversion hypotheses.
Finish with the five improvements I should investigate first.
Useful input
Include checkout screens, pricing pages, forms, landing-page copy, FAQs, or screenshots converted into text.
5. Evaluate an Existing Content Strategy
AI can turn a large collection of existing content into a strategic overview.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Review this content inventory and determine how well it supports the business strategy.
Business goal:
[GOAL]
Target audience:
[AUDIENCE]
Products or services:
[OFFER]
Existing content:
[PASTE CONTENT LIST]
Evaluate the inventory for:
- Topic coverage
- Audience relevance
- Search intent coverage
- Funnel-stage coverage
- Commercial relevance
- Educational value
- Repetition
- Missing topics
- Weak or outdated themes
- Opportunities for content clusters
Group the findings into:
A. Strong existing coverage
B. Areas with excessive overlap
C. Important gaps
D. Opportunities worth testing
Do not recommend content merely because a keyword exists.
Explain how each major recommendation connects to the stated business goal.
End with a prioritized 90-day content plan.
6. Compare Competitor Offers
Price alone rarely explains how two offers differ.
Use this prompt to compare the complete offer.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Compare the following competing offers.
Competitor 1:
[OFFER DETAILS]
Competitor 2:
[OFFER DETAILS]
Competitor 3:
[OFFER DETAILS]
Our offer:
[OFFER DETAILS]
Analyze:
- Target customer
- Core offer
- Included features
- Service scope
- Pricing information provided
- Guarantees
- Support
- Onboarding
- Proof
- Limitations
- Upgrade paths
- Calls to action
Create a factual comparison table.
Then answer:
1. Where are the offers genuinely similar?
2. Where are there meaningful differences?
3. Which differences are likely to be easy for customers to understand?
4. Which differences require stronger explanation?
5. Where could our offer be differentiated?
Never estimate competitor pricing or features that are not included in the source material.
7. Improve Calls to Action
A CTA should tell the visitor what to do and why doing it makes sense.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Review these calls to action as a conversion copy editor.
Page purpose:
[PAGE PURPOSE]
Audience:
[AUDIENCE]
Current CTAs:
[PASTE CTAs]
For each CTA, evaluate:
- Clarity
- Specificity
- Relevance to the page
- Action required
- Implied commitment
- Consistency with visitor intent
- Whether the surrounding copy gives enough reason to act
Suggest alternatives only where there is a clear weakness.
For every proposed replacement, explain what changed and what uncertainty it is intended to reduce.
Do not use exaggerated urgency, unsupported claims, or artificial scarcity.
Return the result as:
Current CTA | Issue | Recommended CTA | Reason
8. Identify Trust Signals
Trust is particularly important when a visitor is being asked to spend money, submit information, or make a significant decision.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Audit the following website content for trust signals.
Website content:
[PASTE CONTENT]
Business type:
[BUSINESS TYPE]
Primary conversion:
[DESIRED ACTION]
Identify existing evidence of:
- Expertise
- Experience
- Customer proof
- Reviews
- Case studies
- Certifications
- Guarantees
- Transparent business information
- Policies
- Contact information
- Security or privacy information
- Product/service evidence
For every item, classify it as:
Present
Partially present
Not evident from the supplied material
Do not assume a business possesses credentials, certifications, reviews, awards, or customer results that are not provided.
Then recommend the most useful trust elements to investigate or add, prioritizing those that address real decision-making concerns.
9. Review Homepage Information Architecture
A homepage has limited attention to capture.
The following prompt focuses on whether the order of information makes sense.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Review this homepage structure from the perspective of a first-time visitor.
Homepage content:
[PASTE CONTENT]
Business:
[BUSINESS]
Audience:
[AUDIENCE]
Primary conversion:
[CONVERSION]
Determine whether the page answers these questions in a logical sequence:
1. What is this?
2. Who is it for?
3. What problem does it address?
4. Why should the visitor care?
5. Why should they trust the business?
6. What can they do next?
Identify:
- Missing information
- Repeated information
- Sections that appear out of sequence
- Sections that may need stronger evidence
- Weak transitions
- Unclear CTAs
Then propose an improved homepage section order.
Do not rewrite the entire homepage unless necessary. Focus first on information architecture and visitor comprehension.
10. Find Unanswered Buyer Questions
One of the easiest ways to discover useful content opportunities is to identify questions a prospective customer still has.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Act as a buyer-question researcher.
Business:
[BUSINESS]
Product/service:
[OFFER]
Target customer:
[AUDIENCE]
Existing website information:
[PASTE CONTENT]
Customer questions or sales notes:
[PASTE DATA IF AVAILABLE]
Identify questions a prospective buyer may still need answered before choosing the offer.
Group the questions into:
- Basic understanding
- Suitability
- Features or service details
- Cost and value
- Comparison
- Risk
- Implementation
- Support
- After-purchase concerns
For every question, state whether it is:
Supported by supplied evidence
Suggested by supplied evidence
A reasonable hypothesis requiring validation
Prioritize questions that could influence a buying decision.
Do not manufacture customer questions and present them as researched facts.
11. Create an Evidence-Based SWOT Analysis
A SWOT analysis is only useful when the four categories are grounded in evidence.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Create an evidence-based SWOT-style assessment from the information below.
Business:
[BUSINESS]
Market/context:
[CONTEXT]
Evidence:
[PASTE MATERIAL]
Separate:
Strengths — internal advantages supported by evidence
Weaknesses — internal limitations supported by evidence
Opportunities — external possibilities suggested by the evidence
Threats — external risks supported or reasonably indicated by the evidence
For every item, include:
- Finding
- Supporting evidence
- Confidence: High / Medium / Low
- Why it matters
Do not turn assumptions into facts.
If the supplied material is insufficient for a category, say so.
After the SWOT, identify the three strategic questions management should investigate next.
12. Compare Brand Positioning Across Competitors
This is different from simply comparing competitor wording. It focuses on the space each brand is attempting to occupy.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Compare the positioning of these brands using only the supplied material.
Brand A:
[MATERIAL]
Brand B:
[MATERIAL]
Brand C:
[MATERIAL]
Our brand:
[MATERIAL]
For each brand, identify:
- Intended audience
- Primary problem
- Main promise
- Supporting benefits
- Distinctive language
- Proof used
- Emotional angle
- Functional angle
- Apparent category position
Then create a positioning comparison.
Look specifically for:
1. Areas where brands sound interchangeable.
2. Positions that appear strongly occupied.
3. Ideas that receive little attention.
4. Potential whitespace for our brand.
5. Risks of copying a competitor too closely.
Do not claim that an opportunity is commercially viable without additional market validation.
13. Find Real Differentiation Opportunities
Not every difference is a meaningful differentiator.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Help me determine which potential differentiators could actually matter to customers.
Business:
[BUSINESS]
Audience:
[AUDIENCE]
Current offer:
[OFFER]
Competitor information:
[COMPETITOR DATA]
Potential differentiators:
[LIST DIFFERENCES]
For each proposed differentiator, evaluate:
- Is it genuinely different?
- Is it easy for a customer to understand?
- Is it relevant to the target audience?
- Is it supported by evidence?
- Could competitors easily reproduce it?
- Does it affect the buying decision?
- Does the website currently communicate it?
Score each dimension from 1–5, but explain the reasoning.
Separate factual observations from strategic judgment.
End with:
1. Strong differentiators
2. Weak differentiators
3. Differentiators requiring validation
14. Audit Landing-Page Clarity
A landing page should make the visitor’s next decision easier.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Act as a landing-page clarity reviewer.
Landing-page copy:
[PASTE PAGE]
Campaign/source:
[TRAFFIC SOURCE]
Target audience:
[AUDIENCE]
Desired action:
[CONVERSION]
Evaluate the page in this order:
1. Message match
2. Offer clarity
3. Audience clarity
4. Problem clarity
5. Benefit clarity
6. Evidence and credibility
7. Objections
8. CTA clarity
9. Unnecessary distractions
For every significant weakness, provide:
Current wording or section:
[REFERENCE]
Why it may create confusion:
[EXPLANATION]
Suggested direction:
[RECOMMENDATION]
Do not rewrite sections simply for stylistic preference.
Prioritize changes that could improve comprehension or reduce uncertainty.
15. Select Suitable Marketing Channels
Not every business needs to be everywhere.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Recommend marketing channels based on the business situation below.
Business:
[BUSINESS]
Offer:
[PRODUCT/SERVICE]
Audience:
[AUDIENCE]
Customer buying process:
[BUYING PROCESS]
Available resources:
[BUDGET / TEAM / TIME]
Existing channels:
[CURRENT CHANNELS]
Evaluate these channel categories:
- Search
- Organic social
- Paid social
- Email
- Partnerships
- Communities
- Referral
- Events
- Content marketing
- Direct outreach
- Other relevant channels
For each suitable channel, explain:
• Why it may fit
• What customer stage it can support
• Required resources
• Content or assets needed
• Main limitation
• What should be tested first
• What evidence would indicate success
Do not claim a channel will definitely produce a particular result.
Recommend a small initial channel mix rather than attempting to use every available platform.
16. Turn Existing Content Into Reusable Formats
One useful piece of research can often become several different assets.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Take the following source material and identify ways it could be repurposed without losing its original meaning.
Source:
[PASTE CONTENT]
Audience:
[AUDIENCE]
Business objective:
[OBJECTIVE]
Identify suitable transformations such as:
- Short social posts
- Email content
- FAQ entries
- Video scripts
- Carousel concepts
- Checklists
- Guides
- Sales enablement material
- Webinar topics
- Case-study structures
- Short-form educational content
For each recommendation, provide:
1. Source section
2. New format
3. Intended audience
4. Core takeaway
5. Additional information required
Do not create new factual claims during repurposing.
Preserve important qualifications and context from the original material.
17. Build a Competitor Research Checklist
A repeatable checklist prevents competitive research from becoming random browsing.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Create a competitor-research checklist for this business.
Business:
[BUSINESS]
Industry:
[INDUSTRY]
Primary competitors:
[COMPETITORS]
Research objective:
[OBJECTIVE]
Organize the checklist into:
### Company
- Positioning
- Audience
- Business model
### Offer
- Products/services
- Features
- Packages
- Pricing information
### Website
- Navigation
- Key landing pages
- Conversion paths
- Trust elements
### Content
- Core topics
- Formats
- Publishing patterns
- Educational resources
### Search visibility
- Important topics
- Search-intent coverage
- Content gaps
### Customer experience
- Reviews
- FAQs
- Support information
- Objection handling
### Differentiation
- Promises
- Proof
- Unique mechanisms
- Messaging patterns
Mark each item as:
To research
Researching
Verified
Needs validation
Finish with a compact version that a researcher can use as a recurring checklist.
18. Convert Research Into a Strategic Brief
Research is only valuable when someone can act on it.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Turn the research below into a decision-ready strategic brief.
Research:
[PASTE RESEARCH]
Business objective:
[OBJECTIVE]
Audience:
[AUDIENCE]
Create the brief using these sections:
1. Executive summary
2. What the evidence tells us
3. Important customer observations
4. Competitive observations
5. Key opportunities
6. Risks and uncertainties
7. Recommended priorities
8. Actions to take
9. Questions requiring further research
For every recommendation, connect it to the evidence that supports it.
Do not introduce new statistics, market facts, or assumptions.
Clearly label conclusions that require validation.
Make the final section practical enough for a team to turn into an action plan.
19. Review Pricing-Page Messaging
Pricing pages need to answer more than “How much does it cost?”
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Review this pricing-page content from the perspective of someone deciding whether to buy.
Pricing page:
[PASTE CONTENT]
Offer:
[PRODUCT/SERVICE]
Audience:
[AUDIENCE]
Evaluate:
- Price visibility
- Package differentiation
- Feature clarity
- Value explanation
- Intended customer for each plan
- Upgrade logic
- Limitations
- Commitment required
- Objection handling
- CTA clarity
- Missing decision information
Identify questions a buyer could still have after reading the page.
Then suggest improvements to the page structure.
Do not recommend fake discounts, artificial urgency, unsupported savings claims, or invented pricing advantages.
Where the source does not explain why a plan is better, say that the value proposition needs clarification rather than inventing one.
20. Analyze Competitor FAQs for Content Opportunities
FAQs can reveal how businesses explain their offers—and where customers may still need more information.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Analyze these competitor FAQs.
Competitor A:
[FAQ CONTENT]
Competitor B:
[FAQ CONTENT]
Competitor C:
[FAQ CONTENT]
Our website:
[OUR FAQ CONTENT]
Classify questions into themes such as:
- Product/service basics
- Eligibility or suitability
- Pricing
- Process
- Technical questions
- Delivery
- Support
- Returns/cancellation
- Risk
- Results
- Comparison
Identify:
1. Questions answered by several competitors.
2. Questions answered by only one competitor.
3. Important questions our site does not address.
4. Areas where competitor answers appear vague.
5. Questions that could become useful content topics.
Do not assume that a competitor's answer is correct.
Treat competitor FAQ wording as source material for analysis, not as authoritative truth.
21. Detect Content Cannibalization Risks
Publishing many articles around the same subject can create unnecessary overlap.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Review this list of pages for potential content overlap.
Website:
[WEBSITE]
Page inventory:
[URL + TITLE + PRIMARY TOPIC + SEARCH INTENT]
Look for pages that appear to target substantially similar:
- Topics
- Search intents
- Questions
- Audiences
- Commercial goals
For each possible overlap, provide:
Page A:
Page B:
Shared subject:
Intent similarity:
Potential problem:
Evidence:
Recommended action:
Possible actions include:
Keep separate
Merge
Redirect one page
Differentiate intent
Change the primary topic
Strengthen topical boundaries
Review manually
Do not recommend merging pages simply because they mention the same keyword.
Explain whether the underlying search intent and purpose are actually similar.
22. Create a Competitive Content-Gap Report
A content gap should be more than “the competitor has an article we don’t.”
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Create a competitive content-gap analysis.
Our content inventory:
[OUR CONTENT]
Competitor content:
[COMPETITOR CONTENT]
Target audience:
[AUDIENCE]
Business objective:
[OBJECTIVE]
Compare the inventories by:
- Topic
- Search intent
- Funnel stage
- Audience need
- Content format
- Depth
- Commercial relevance
- Originality
Create four categories:
1. Topics competitors cover that we do not.
2. Topics we cover more effectively.
3. Topics both sides cover but could be improved.
4. Audience questions neither side appears to answer.
For every recommended gap, include:
Topic:
Audience need:
Intent:
Business relevance:
Existing evidence:
Suggested content format:
Priority:
Do not assume that every competitor topic deserves to be copied.
Prioritize opportunities based on audience usefulness and business relevance.
23. Extract and Document Brand Voice
Brand voice should be based on actual language rather than a collection of vague adjectives.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Analyze the following brand content and create a practical voice guide.
Brand material:
[PASTE EXAMPLES]
Identify recurring patterns in:
- Sentence length
- Vocabulary
- Formality
- Directness
- Emotional tone
- Use of technical terminology
- Use of contractions
- First/second/third person
- Calls to action
- Confidence level
- Use of humor
- How benefits are explained
- How problems are discussed
Then create:
### Voice principles
[List the main principles]
### Do
[Examples]
### Avoid
[Examples]
### Before and after
Show three short examples of generic copy rewritten to match the observed voice.
Base the guide on the supplied material.
Do not invent a brand personality that is not supported by the examples.
24. Audit an Existing Competitive-Analysis Report
Sometimes the problem is not the research itself but the conclusions drawn from it.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Audit this competitive-analysis report as a critical reviewer.
Original research:
[PASTE SOURCE DATA]
Competitive-analysis report:
[PASTE REPORT]
Check every major conclusion for:
- Evidence
- Relevance
- Logical consistency
- Unsupported assumptions
- Missing context
- Contradictions
- Overgeneralization
- Confusion between correlation and causation
- Recommendations that do not follow from the evidence
Create a table:
Finding | Evidence | Confidence | Potential issue | Recommended correction
Then identify:
1. Conclusions that are strongly supported.
2. Conclusions requiring additional evidence.
3. Conclusions that should be removed or rewritten.
4. Important questions the report failed to investigate.
Do not rewrite the report until the evidence problems have been identified.
25. Run a Complete Website-Analysis Workflow
For a full website review, separate discovery, diagnosis, recommendations, and quality control instead of asking AI for everything simultaneously.
Prompt
Fill in the blanks below, or click a highlighted word in the prompt.
Act as a senior website and digital-marketing analyst.
I need a structured review of the following website.
Business:
[BUSINESS]
Website content:
[PASTE CONTENT OR RELEVANT PAGE DATA]
Audience:
[AUDIENCE]
Business objective:
[OBJECTIVE]
Available competitor information:
[COMPETITOR DATA]
Analyze the material in the following sequence.
PHASE 1 — BUSINESS UNDERSTANDING
Determine what the business appears to offer, who it serves, and what action the website wants visitors to take.
PHASE 2 — MESSAGE REVIEW
Evaluate:
- Positioning
- Value proposition
- Audience clarity
- Differentiation
- Brand messaging
PHASE 3 — CUSTOMER EXPERIENCE
Review:
- Navigation
- Information hierarchy
- Customer questions
- Objection handling
- Trust
- Conversion path
PHASE 4 — CONTENT
Evaluate:
- Topic coverage
- Search intent
- Content gaps
- Repetition
- Content structure
- Internal opportunities
PHASE 5 — COMPETITIVE CONTEXT
Compare the supplied competitor information and identify meaningful similarities and differences.
PHASE 6 — CONVERSION
Identify potential friction points and explain which are observations versus hypotheses.
PHASE 7 — PRIORITIZATION
Create a prioritized action list using:
Impact:
High / Medium / Low
Effort:
High / Medium / Low
Evidence confidence:
High / Medium / Low
PHASE 8 — QUALITY CONTROL
Before finalizing:
- Check every factual statement against the supplied material.
- Remove unsupported claims.
- Identify assumptions.
- Flag missing information.
- Check that recommendations logically follow from the findings.
Do not invent traffic figures, rankings, customer behavior, competitor data, revenue, conversion rates, certifications, or other business facts.
Return the final report in a clear structure that a marketing team can turn into an action plan.
Common AI Prompt Mistakes in Marketing Work
Even a powerful model can produce poor analysis when the instructions are poorly designed.
1. Asking for a vague outcome
“Make my website better” does not tell AI what success looks like.
Define the actual task.
For example:
“Identify the five places where a first-time visitor may struggle to understand the offer.”
2. Providing too little context
An AI model may know general marketing principles, but it does not automatically know your company’s:
- Customers
- Products
- Pricing
- Internal processes
- Competitive position
- Business priorities
Give it the information required to make the analysis meaningful.
3. Letting AI fill information gaps
This is particularly dangerous when analyzing:
- Prices
- Customer statistics
- Financial information
- Legal claims
- Certifications
- Performance results
- Business credentials
- Competitor data
Use explicit instructions such as:
“If the information is unavailable, identify the gap instead of guessing.”
4. Treating assumptions as research
AI can generate plausible explanations that sound authoritative.
That does not make them verified.
Ask the model to distinguish:
Evidence → Interpretation → Hypothesis → Recommendation
This simple separation makes the output easier to review.
5. Combining too many unrelated jobs
A single prompt asking AI to research competitors, write articles, redesign a website, create an SEO strategy, analyze pricing, and produce a marketing plan can result in shallow output.
Break complex projects into stages.
6. Assuming longer prompts are automatically better
Prompt quality is not measured by word count.
A short prompt containing the correct context and clear output requirements can outperform a much longer prompt filled with irrelevant instructions.
Focus on:
- Specificity
- Relevant context
- Evidence
- Constraints
- Output requirements
7. Copying AI output without reviewing it
AI-generated analysis should be treated as a working document, not automatically as verified research.
Check important:
- Names
- Numbers
- Dates
- Claims
- Policies
- Statistics
- Legal statements
- Financial information
- Technical recommendations
against authoritative sources.
A Better Workflow: Input → Analyze → Draft → Verify
For recurring marketing work, one of the most useful improvements is to stop asking AI to produce the final answer immediately.
Instead, create a repeatable four-stage process.
Stage 1: Input
Give AI clean source material.
For example:
Here is the website copy, customer research, competitor information, and business objective.
Do not analyze it yet.
First identify:
- What information is available
- What information is missing
- Any contradictions
- Any unclear terminology
This prevents the model from building an analysis on incomplete inputs.
Stage 2: Analyze
Once the information has been checked, ask AI to examine patterns.
Using only the verified source material, identify:
- Major themes
- Customer needs
- Competitive patterns
- Content gaps
- Potential risks
- Opportunities
Separate observations from interpretations.
At this stage, you are investigating rather than writing.
Stage 3: Draft
Now ask for the actual deliverable.
Turn the approved findings into a strategic report.
Prioritize the findings according to:
1. Business impact
2. Evidence strength
3. Implementation effort
Use the requested report structure.
Stage 4: Verify
Finally, run a separate quality-control pass.
Review the report against the original source material.
For every major recommendation:
- Identify the supporting evidence.
- Identify unsupported assumptions.
- Identify missing information.
- Flag claims that require external verification.
- Identify contradictions.
- Check whether the recommendation logically follows from the evidence.
Do not rewrite the report yet.
Return only the quality-control findings.
This staged approach can be considerably more dependable than asking for a “perfect” report in one giant prompt.
How to Make These Prompts Work Better
The prompts above are templates, not magic formulas.
Their output quality depends heavily on the information you provide.
Give AI the actual material
If you want an analysis of a webpage, provide the relevant page content.
If you want competitor research, provide the competitor information.
If you want customer insights, provide actual customer evidence where possible.
The more directly the model can work from evidence, the less it needs to rely on assumptions.
Tell it who the audience is
“Customers” is often too broad.
Instead:
“Existing B2B software managers evaluating a replacement platform.”
The narrower description gives the analysis a more useful frame.
Specify the decision you need to make
Instead of:
“Analyze these competitors.”
Try:
“We need to decide whether our next content campaign should focus on comparison content, educational content, or product-led content.”
Now the analysis has a purpose.
Request prioritization
A long list of recommendations is rarely enough.
Ask AI to rank findings according to criteria such as:
- Impact
- Effort
- Confidence
- Urgency
- Business relevance
- Customer importance
Keep human judgment in the loop
AI can help you organize information, identify patterns, create alternatives, and expose questions worth investigating.
It should not automatically become the final authority on consequential business decisions.
Can I use these prompts for SEO?
Yes, particularly for areas such as:
- Content-gap analysis
- Search-intent research
- Competitor content analysis
- Content planning
- Content overlap
- Customer-question discovery
- Website messaging
- Landing-page analysis
However, AI-generated SEO recommendations should be validated against actual search data, website performance, audience behavior, and relevant search results before making major strategic decisions.
Final AI Prompt Quality Checklist
Before running an important marketing-analysis prompt, ask yourself:
- Is the objective clearly defined?
- Have I explained the business context?
- Have I identified the intended audience?
- Have I provided the relevant source material?
- Have I explained the decision or outcome I need?
- Have I specified the desired output format?
- Have I defined important constraints?
- Have I told AI not to invent missing facts?
- Have I asked it to distinguish evidence from assumptions?
- Have I requested prioritization where appropriate?
- Have I included examples when the desired output is difficult to describe?
- Have I avoided combining unrelated tasks?
- Will I verify important claims before using the result?
- Have I considered whether sensitive information should be excluded?
- Can another person understand and repeat the workflow?
Final Takeaway
The most useful AI prompts are not necessarily the longest ones.
They work because they give the model a specific problem to solve, relevant evidence to work from, clear boundaries, and a useful definition of success.
For website analysis and marketing research, this becomes especially important. Generic instructions tend to produce generic recommendations. Task-specific prompts can instead help you examine positioning, customer journeys, competitor messaging, conversion barriers, content gaps, trust signals, pricing communication, brand voice, and broader website strategy from different angles.
The biggest improvement you can make is to stop treating prompting as a request for an answer and start treating it as a structured analytical workflow:
Provide evidence → Define the question → Analyze → Prioritize → Draft → Verify.
That approach gives you a much stronger foundation for using AI as a practical research and marketing assistant rather than simply as a content generator.