competitive-intel

name

competitive-intel

description

Real-time competitive intelligence using live web data via the Tavily MCP. Use when the user wants to analyze competitors, build a battlecard, profile a competitor, monitor pricing, mine reviews, track hiring signals, map a competitive landscape, find positioning gaps, or research win/loss patterns; says "analyze competitor", "build me a battlecard", "what is X charging", "how do we compare to Y", "competitive landscape for Z", "competitor analysis", "market intelligence", or anything similar. Always uses live data — never answers competitive questions from training knowledge alone.

Competitive Intelligence

Live competitive analysis powered by the Tavily MCP. Every finding is backed by live web evidence — never answer competitive questions from training knowledge alone, because product positioning, pricing, and team moves change constantly.

When to use this skill

• "Analyze competitor X" / "Build a battlecard on X."
• "How does X price?" / "What's their packaging?"
• "What are people saying about X?" / "Mine their reviews."
• "Who's hiring for what at X?" / "Track their team growth."
• "Map the competitive landscape for category Y."
• "Find positioning gaps we could own."

If the user just wants a generic market report (no specific competitor), use deep-research instead.

Workflow

Six recipes — pick the one(s) that match the user's question.

1. Competitor snapshot (≈30s)

Build a quick profile of one competitor.

1. tavily_search for "<competitor> overview" with include_answer: "advanced" to get a base summary.
2. tavily_search for "<competitor> pricing" and "<competitor> features", include_domains set to the competitor's own domain.
3. tavily_extract the top results for pricing and product pages.
4. Synthesize into: positioning, pricing, target customer, key features, recent moves.

2. Pricing comparison (≈45s)

Compare pricing across multiple competitors.

1. For each competitor: tavily_map from <domain> with instructions: "pricing pages".
2. tavily_extract each pricing URL.
3. Build a markdown table: tier, price, key limits, ideal customer.
4. Highlight the gaps: who's cheapest, who's free-tier-aggressive, who's gating which features behind enterprise.

3. Review mining (≈60s)

What do real customers say?

1. tavily_search for "<competitor> review", include_domains: "g2.com,trustpilot.com,reddit.com,capterra.com".
2. tavily_extract top 10 results.
3. Cluster sentiment: top 5 praises, top 5 complaints, recurring themes, recent shifts.

4. Hiring signals (≈30s)

What's a competitor building, based on who they're hiring?

1. tavily_search for "<competitor> careers" and extract their jobs page.
2. tavily_search for "<competitor>" site:linkedin.com/jobs to catch listings hosted on LinkedIn.
3. Tag roles by function (eng, product, sales, GTM) and seniority.
4. Surface unusual signals: a sudden infra-eng hiring spike, a first solutions-engineer hire, etc.

5. Content & SEO battle (≈90s)

Where do they rank, where do they own narrative?

1. tavily_map from <domain>/blog with instructions: "blog posts".
2. tavily_search for the top topics they cover, observe SERP rank.
3. tavily_research query: "What topics does <competitor> own in search? What gaps are they leaving?" with model: "pro".

For the gap analysis specifically, hand off to content-gap-analysis.

6. Landscape map (use deep-research for the heavy lift)

For "map the competitive landscape" questions, call deep-research with a query like:

Map the competitive landscape for {CATEGORY} in {YEAR}. Group players
by segment. For each, capture positioning, pricing tier, key
differentiators, and recent moves. Audience: {AUDIENCE}.

Then layer per-competitor snapshots on top using recipe 1.

Battlecard format

When the user asks for a battlecard, structure it as:

# {Competitor Name} — Battlecard
*Updated: {YYYY-MM-DD} — sources cited inline*

## Snapshot
- Positioning:
- Target customer:
- Pricing model:
- Recent moves:

## Strengths
- …

## Weaknesses (where we win)
- …

## Common objections we'll hear
- "{objection}" → talk-track: {response}

## Talk-tracks
- Discovery questions to ask:
- Trap questions (where their answer favors us):

## Sources
[1] {url}
[2] {url}

Save to ~~document store so sales can find it later.

Always cite. Never bluff.

Every claim in a competitive deliverable must trace to a tavily_* result. If a fact can't be sourced, mark it as [unverified] and tell the user — don't guess.

For sales-objection templates and pricing-table layouts, see references/battlecard-template.md.