Competitive Analysis in 2026: The AI Agent Guide (With Claude Code Prompts)

mkdir competitive-analysis && cd competitive-analysis

claude

## Task: Discover My Direct Competitors

I run [yourcompany.com], which is [one-sentence description of what you sell and to whom].

### Step A: Find competitors via organic search overlap
1. Visit my homepage and extract what my product does and who it targets.
2. Search the web for "[my product category] alternatives" and "[my product category] vs" queries. Collect the domains that appear repeatedly.
3. Search for reviews/comparisons that include my product category on G2, Capterra, and Reddit. Extract competitor names that appear alongside products like mine.

### Step B: Find competitors via paid search
Search for "[my brand name] alternatives" and "[my brand name] vs" and note which domains are consistently mentioned or advertised.

### Step C: Classify and deduplicate
Categorize each competitor as:
- **Direct** (same solution, same audience)
- **Indirect** (different solution, same audience)
- **Aspirational** (same solution, different/larger audience)

### Output
Save to `competitors.md`:
- A ranked list of direct competitors (3–8 names + domains)
- A secondary list of indirect competitors to watch
- For each: one sentence on why they're a competitor, and your confidence level (High/Medium/Low)
- A note on any data gaps (e.g., "Could not access paid search data without API key")

# CloudDesk — Direct Competitors

| # | Competitor | Domain | Why | Confidence |
|---|-----------|--------|-----|------------|
| 1 | HelpScout | helpscout.com | Shared "help desk software" queries; targets SMB support teams | High |
| 2 | Freshdesk | freshdesk.com | Overlaps on "customer support software" and SMB segment | High |
| 3 | Zendesk | zendesk.com | Dominates enterprise queries; overlaps on mid-market searches | High |
| 4 | Intercom | intercom.com | Competes on "customer messaging platform" queries; partial product overlap | Medium |
| 5 | Zoho Desk | zoho.com/desk | Price-competitive alternative; appears in "cheap help desk" searches | Medium |

## Indirect Competitors
- **Front** (front.com) — shared inbox, not full helpdesk; different solution, overlapping audience
- **HubSpot Service Hub** (hubspot.com) — CRM-first; competes when buyer already uses HubSpot

## Data Gaps
- Paid search overlap: not verified (no SEMrush/Ahrefs API configured). Recommend checking manually or connecting an API.

## Task: Compare Market Reach for My Competitors

I analyzed my competitors in `competitors.md`. Now compare their market reach across three channels.

### Step A: Organic Search Visibility
For each competitor (and my site), search for:
1. "[competitor name]" and note if they rank #1 for their own brand
2. "[product category] software" and note which competitors appear on page 1
3. "[product category] vs" and note which comparison pages rank
4. Estimate which competitor has the broadest organic footprint based on the number of ranking pages you can observe

If you have access to a configured DataForSEO, Ahrefs API, or Semrush API, use it. Otherwise, estimate from visible SERP data and clearly label as estimates.

### Step B: Social Presence
For each competitor, check:
1. LinkedIn company page — follower count and posting frequency
2. Twitter/X — follower count and engagement on recent posts
3. YouTube — subscriber count and upload frequency (if relevant to the product category)
4. Any platform where the audience in this category is known to be active

### Step C: AI Search Visibility (NEW)
This is the 2026 dimension. Test the following queries across AI platforms:

For each of these queries, check ChatGPT, Perplexity, and Google AI Overviews (when available):
1. "best [product category] software"
2. "[product category] for [target audience]"
3. "[competitor A] vs [competitor B]"
4. "alternatives to [my brand]"
5. "[specific use case] tool"

For each query and platform, record:
- Which brands are mentioned or recommended
- Whether my brand appears
- The sentiment/tone toward each mentioned brand
- The source or rationale the AI gives for its recommendation

Note: Checking AI answers manually is valid. If you have access to an AI citation tracking tool (AgentAEO MCP, DataForSEO AI Optimization API), use it. Otherwise, manually test 3–5 high-value queries and label as a manual sample.

### Output
Save to `market-reach.md`:
1. **Organic SOV table** — competitor vs estimated organic footprint (label all estimates)
2. **Social presence comparison** — followers, posting frequency, platform notes
3. **AI visibility heatmap** — competitors × platforms × queries, showing who gets cited where
4. **Gap analysis** — where competitors are visible and I'm not, with priority ratings

# CloudDesk — Market Reach Comparison

## Organic Search Visibility (Estimated from SERP observation)

| Competitor | Brand SERP Ownership | Category Page 1 Presence | Comparison Query Presence | Estimated Footprint |
|-----------|---------------------|------------------------|--------------------------|-------------------|
| Zendesk | Strong (#1, full SERP control) | Yes (help desk, customer service) | High (appears in multiple vs queries) | Largest |
| Freshdesk | Strong (#1, knowledge panel) | Yes (help desk) | Medium | Large |
| HelpScout | Good (#1, no knowledge panel) | Partial (ranks for "shared inbox" more than "help desk") | Medium | Medium |
| Intercom | Strong (#1) | Partial (ranks for "customer messaging" not "help desk") | Low | Medium |
| CloudDesk (us) | Weak (ranks #1 for brand, but no knowledge panel) | Low (appears page 2–3 for category terms) | Very low | Small |

⚠️ All estimates from manual SERP observation 2026-07-30. No API data.

## Social Presence

| Competitor | LinkedIn Followers | Twitter/X Followers | YouTube Subs | Posting Cadence |
|-----------|-------------------|-------------------|--------------|----------------|
| Zendesk | 450K+ | 200K+ | 25K+ | Daily, all channels |
| Intercom | 200K+ | 150K+ | 20K+ | Daily, strong video |
| Freshdesk | 150K+ | 80K+ | 10K+ | 3-5x/week |
| HelpScout | 35K+ | 25K+ | 5K+ | 2-3x/week, blog-heavy |
| CloudDesk (us) | 5K | 3K | 500 | Weekly |

⚠️ Follower counts from public profile pages on 2026-07-30.

## AI Search Visibility (Manual Sample, 5 Queries)

| Query | ChatGPT | Perplexity | Google AIO |
|-------|---------|------------|------------|
| "best help desk software" | Zendesk, Freshdesk, HelpScout | Zendesk, Freshdesk, Zoho Desk | Zendesk, Freshdesk |
| "help desk for small teams" | HelpScout, Freshdesk | HelpScout, Zoho Desk | HelpScout, Freshdesk |
| "Zendesk vs Freshdesk" | Both mentioned neutrally | Both with pros/cons | N/A (no AIO for this query) |
| "alternatives to Zendesk" | Freshdesk, HelpScout, Intercom | Freshdesk, HelpScout, Zoho Desk | Freshdesk, HelpScout |
| "shared inbox software" | HelpScout, Front | Front, HelpScout | Front, HelpScout |

⚠️ Manual check on 2026-07-30 using logged-out sessions. CloudDesk did not appear in any AI answer for these queries.

## Gap Analysis

| Gap | Competitors Winning | Priority | Action |
|-----|-------------------|----------|--------|
| Category page 1 presence | Zendesk, Freshdesk | 🔴 High | Build topical authority for "help desk software" |
| AI visibility | Zendesk, HelpScout, Freshdesk | 🔴 High | Optimize for answer extraction; ensure entity clarity |
| Social presence | All competitors | 🟡 Medium | Increase LinkedIn posting to 3x/week |
| Comparison query coverage | Zendesk, Freshdesk | 🟡 Medium | Create comparison pages for high-volume vs queries |

## Task: Extract and Compare Competitor Pricing and Features

Use the competitor list in `competitors.md`.

### Step A: Extract pricing data
For each competitor, visit their pricing page. For each plan/tier, extract:
- Plan name
- Monthly price (and annual price if available)
- User limit (if any)
- Key features included at this tier
- Features locked behind higher tiers
- Free trial availability and length
- Free plan (yes/no, and what it includes)

### Step B: Extract feature lists
From each competitor's feature page or pricing page, collect:
- All named features
- Which features are unique to that competitor (not offered by others)
- Which features are table-stakes (everyone has them)

### Step C: Structure the comparison
Build two tables:

**Pricing Table:**
| Competitor | Cheapest Plan | Mid-Tier | Full Plan | Free Tier? | Free Trial? |
|-----------|--------------|----------|-----------|------------|-------------|

**Feature Matrix:**
| Feature | CloudDesk (us) | Competitor A | Competitor B | Competitor C | Competitor D | Competitor E |
|---------|---------------|-------------|-------------|-------------|-------------|-------------|

Mark each cell: ✅ (has it), ❌ (doesn't), ⚠️ (partial/add-on), ❓ (could not verify)

### Output
Save to `pricing-features.md`:
1. Pricing table with plan names and prices
2. Feature matrix with all competitors
3. Unique features per competitor
4. Pricing gap analysis: where we are more expensive, cheaper, or missing key features
5. Data freshness note: when each pricing page was accessed

### Data Integrity Rules
- Only record prices you can see on public pages. Do not guess.
- If a pricing page requires login, record "Login-gated" and skip.
- Record the URL of every pricing page you accessed.
- Label any prices that appear to be promotional or time-limited.

## Task: Analyze Competitor User Reviews

For each competitor in `competitors.md`, analyze 2–4 star reviews from G2, Capterra, and Trustpilot.

### Step A: Collect reviews
For each competitor, find their G2, Capterra, and Trustpilot profile pages. Collect up to 30 reviews per competitor, prioritizing:
- Most recent first
- 2–4 star ratings (these contain the most actionable feedback)
- Reviews with detailed written text (skip "Great product" single-liners)

### Step B: Analyze sentiment and themes
For the collected reviews, identify:
1. **Recurring praise themes** — what do satisfied-but-not-delighted users consistently mention?
2. **Recurring complaints** — what do users consistently criticize?
3. **Feature requests** — what do users wish the product had?
4. **Switching reasons** — why did users leave or consider leaving?
5. **Onboarding experience** — how do users describe their first weeks with the product?
6. **Support quality** — how do users describe customer support interactions?

### Step C: Write per-competitor summaries
For each competitor, output:
- Top 3 strengths (with example quote fragments, not full quotes)
- Top 3 weaknesses (with example quote fragments)
- One sentence: "Users choose [Competitor] because _____, but they complain about _____."

### Step D: Cross-competitor patterns
Identify themes that appear across multiple competitors:
- What do users across the market consistently want?
- What seems to be an unsolved problem in this product category?

### Output
Save to `review-analysis.md`:
1. Per-competitor sentiment summary
2. Cross-competitor pattern analysis
3. Opportunities table: unsolved problems that no competitor addresses well
4. Review source list: which platforms were used, how many reviews per competitor

### Data Integrity
- Do not fabricate reviews. Only use text you can access.
- Quote fragments should be short enough to be fair use.
- If a review platform blocks access (login gate, CAPTCHA), note it and skip that source.
- Record the date you accessed each review source.

## Task: Benchmark Purchasing Experience

For each competitor in `competitors.md`, evaluate their purchasing journey as a prospective buyer would experience it.

### Evaluation Criteria
Score each competitor on these dimensions (1–5, where 5 is best):

1. **Pricing transparency** — Are prices visible without logging in? Is it clear what each plan includes?
2. **Free trial / freemium access** — Can you try the product without talking to sales? How fast can you get to value?
3. **Signup friction** — How many steps from landing page to working product? Is a credit card required?
4. **Demo availability** — Can you see the product before signing up? (Screenshots, videos, interactive demo, sandbox)
5. **Sales contact accessibility** — If you need to talk to someone, how easy is it to find contact info or start a chat?
6. **Cancellation/downgrade clarity** — Is it clear how to cancel or downgrade? (Check the help center, not the signup flow)
7. **Documentation quality** — Is there a public knowledge base? Are setup guides clear and current?

### Process
For each competitor:
1. Visit their homepage and find the pricing/signup path without using search
2. Start a signup/trial if one exists (stop before entering real information)
3. Check their help center for cancellation/downgrade instructions
4. Note any dark patterns: hidden prices, forced sales calls, confusing tier structures, hard-to-find cancel buttons

### Output
Save to `purchasing-experience.md`:
1. **Scorecard table** — all competitors scored on each dimension
2. **Best practices observed** — what specific competitors do well
3. **Friction points** — where competitors make it hard to buy
4. **Your gap analysis** — where your own purchasing experience needs improvement

### Data Integrity
- Only score dimensions you can verify as a logged-out visitor.
- If testing signup, stop before entering real personal information.
- Note which parts of the experience are login-gated.

## Task: Synthesize Competitive Analysis Findings

I have completed my competitive analysis. The findings are in:
- `competitors.md` — competitor list and classification
- `market-reach.md` — organic, social, and AI visibility comparison
- `pricing-features.md` — pricing table and feature matrix
- `review-analysis.md` — user review sentiment and patterns
- `purchasing-experience.md` — purchasing journey scorecard

### Step A: Executive Summary
Write a 3–4 paragraph executive summary covering:
- Who the top 2–3 competitive threats are and why
- The biggest visibility gap (where competitors reach customers that we don't)
- The biggest product/pricing gap (what competitors offer that we don't)
- The most urgent recommendation

### Step B: Priority Matrix
Create a 2×2 matrix:
- **High Impact + Easy to Fix** → Do this week
- **High Impact + Hard to Fix** → Plan this quarter
- **Low Impact + Easy to Fix** → Do when convenient
- **Low Impact + Hard to Fix** → Deprioritize

Place at least 3 specific actions into the matrix.

### Step C: Recommendations by Team
For each of these teams, provide 1–3 specific, sequenced recommendations:
- **Product team** — what to build, fix, or improve
- **Marketing team** — what content, campaigns, or positioning to create
- **Sales team** — what competitive objections to prepare for, what battlecards need updating

### Step D: What We Don't Know
List every significant question the analysis could not answer:
- Data that required an API key we didn't have
- Pages that were login-gated
- Metrics we estimated rather than measured
- Assumptions that need human verification

### Output
Save to `competitive-analysis-report.md` — a single, self-contained report.

### Style Rules
- Every recommendation names a specific competitor and a specific gap.
- Do not write "improve SEO" — write "build 5 comparison pages targeting queries where Zendesk currently ranks #1–3."
- Do not write "fix pricing" — write "our mid-tier plan costs $X more than HelpScout's equivalent plan while missing feature Y."
- Mark every uncertain claim with [NEEDS VERIFICATION].

---
name: competitive-analysis
description: Run a complete competitive analysis against up to 5 competitors — discover who they are, compare reach and pricing, analyze reviews, benchmark AI visibility, and produce an actionable report.
---

# Competitive Analysis Agent

You are a competitive analysis agent for SEO beginners and marketing practitioners. Your job: execute a structured competitive analysis workflow, save results to organized markdown files, and flag every data gap or assumption.

## When the User Triggers This Skill
They say something like: "Analyze my competitors," "Run a competitive analysis for [domain]," or they invoke `/competitive-analysis [domain]`.

## Workflow

### Phase 0: Confirm Scope
Ask the user:
1. What is the target website/domain?
2. What does the product do and who is the audience? (one sentence each)
3. How many competitors should I analyze? (default: 5)
4. Do they have API keys configured for DataForSEO, Firecrawl, or similar? If not, I will use visible web data and label everything as estimates.

Wait for answers before proceeding.

### Phase 1: Competitor Discovery
1. Visit the target homepage. Extract: product description, target audience, key features.
2. Search for "[product category] alternatives," "[product category] vs," and "[product category] software" to find competitor domains.
3. Search G2, Capterra, Reddit for product category mentions. Collect domains that appear alongside the target.
4. Classify each as Direct / Indirect / Aspirational.
5. Save to `competitors.md` with: ranked list, classification, confidence level, data gaps.

### Phase 2: Market Reach Comparison
1. **Organic search**: For each competitor, search "[competitor name]" and "[product category] software." Note SERP presence, knowledge panels, and ranking patterns. Label all observations as [ESTIMATE] unless API data is available.
2. **Social presence**: Check LinkedIn, Twitter/X, YouTube for follower counts and posting frequency. Use public profile data only.
3. **AI search visibility**: Test 5 category-relevant queries across ChatGPT, Perplexity, and Google AI Overviews (when accessible). Record which brands are cited, sentiment, and source attribution.
4. Save to `market-reach.md` with: organic SOV table, social comparison, AI visibility heatmap, gap analysis.

### Phase 3: Pricing and Feature Comparison
1. Visit each competitor's public pricing page. Extract: plan names, prices, user limits, key features per tier, free trial/free plan availability.
2. Visit each competitor's feature page. Collect: all named features, unique features, table-stakes features.
3. Build: pricing comparison table and feature matrix.
4. Mark every feature as ✅ (has it), ❌ (doesn't), ⚠️ (partial/add-on), or ❓ (could not verify).
5. Save to `pricing-features.md` with: pricing table, feature matrix, unique features, gap analysis, source URLs and access dates.

### Phase 4: Review Analysis
1. For each competitor, find G2, Capterra, and Trustpilot profile pages.
2. Collect 2–4 star reviews (most recent, up to 30 per competitor). Skip single-line reviews without substance.
3. Analyze: recurring praise themes, recurring complaints, feature requests, switching reasons, onboarding experience, support quality.
4. Write per-competitor summaries: top 3 strengths, top 3 weaknesses, one-sentence value proposition from user perspective.
5. Identify cross-competitor patterns: what does the market consistently want? What's unsolved?
6. Save to `review-analysis.md` with: per-competitor summaries, cross-competitor patterns, opportunities table, source list.

### Phase 5: Purchasing Experience
1. For each competitor, evaluate as a logged-out prospective buyer: pricing transparency, free trial access, signup friction, demo availability, sales contact accessibility, cancellation clarity, documentation quality.
2. Score each dimension 1–5 with a specific observation.
3. Note any dark patterns observed.
4. Save to `purchasing-experience.md` with: scorecard, best practices, friction points, gap analysis.

### Phase 6: Synthesize Report
1. Read all saved files.
2. Write an executive summary (3–4 paragraphs).
3. Build a priority matrix: High/Low Impact × Easy/Hard to Fix, with at least 3 specific actions.
4. Write recommendations for product, marketing, and sales teams — each naming a specific competitor and gap.
5. List all unknowns: missing API data, login-gated pages, estimates, assumptions needing verification.
6. Save to `competitive-analysis-report.md`.

## Data Integrity Rules (NEVER VIOLATE THESE)
1. **Never guess a price, search volume, follower count, or metric.** If you can't see it, record "Not publicly available" or "Login-gated."
2. **Label every estimate.** Use [ESTIMATE] on any number not directly observed or API-confirmed.
3. **Record sources.** Every data point must include: source URL, access date, and whether it was observed, estimated, or user-provided.
4. **No fabricated reviews.** Only analyze review text you can access. If a platform blocks you, note it and skip.
5. **No secrets in output.** Never write API keys, passwords, tokens, or cookies into any saved file.
6. **Stop before real signups.** When testing purchasing experience, stop before entering personal information.
7. **Your own product gets the same scrutiny.** Score the user's own product with the same rigorous, honest criteria as competitors.

## Output Structure
All files saved to the current working directory:
- `competitors.md`
- `market-reach.md`
- `pricing-features.md`
- `review-analysis.md`
- `purchasing-experience.md`
- `competitive-analysis-report.md`

## Limits
- Analyze a maximum of 8 competitors per run (default: 5).
- Collect a maximum of 30 reviews per competitor.
- Test a maximum of 10 AI search queries per run.
- If a task exceeds these limits, ask the user whether to expand or stop.

mkdir -p .claude/skills/competitive-analysis

/competitive-analysis yourdomain.com