# Competitive Intelligence Prover MCP for AI Agents AI Agent Connect

> Competitive Intelligence Prover stops your AI from making up fake numbers about your rivals. It forces every claim to have a verifiable source, identifies specific measurable weaknesses, and demands a realistic execution plan with actual costs and deadlines. No more vague 'better UX' claims; get hard data.

## Overview
- **Category:** productivity
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_R7dzS7qwhuCobSHINQgCY0Uvj7jZXFHsZIXUb0aB/ai-agent-connect
- **Tags:** competitive-intelligence, competitor-analysis, strategy, anti-hallucination, market-research, business-strategy, agentic-pipeline

## Description

AI agents are prone to producing confident-sounding fiction when you ask them to analyze your competition. They often invent revenue figures, cite non-existent sources, and suggest broad strategies that your team lacks the resources to actually execute. This Connector changes the dynamic by acting as a reality check for your market research. Instead of accepting a generic summary, you get a report that requires every single claim to be backed by a verifiable source like a pricing page, a specific review ID, or a direct test result. It moves your research from a vibe-based guess to a hard audit. When you connect this through the Vinkius catalog, your agent stops suggesting that you should just 'do better marketing' and starts identifying the specific friction points in a rival's onboarding flow. It forces the AI to acknowledge where your own product is lagging while it's looking at the competition. You end up with a strategy that includes a clear breakdown of who is doing the work, what it will cost, and the specific deadlines required to succeed. It turns a brainstorm into a work order.

## Tools

### validate_competitive_intel
Validates competitive analysis for fact-grounded data and feasible strategy. It ensures every claim cites a source and includes a cost-aware execution plan.

## Prompt Examples

**Prompt:** 
```
Competitor X has about $50M in revenue and 30% market share. They have bad UX. We should build a better product and outspend them.
```

**Response:** 
```
Verdict: EVIDENCE_MISSING. Triple failure. 

1. **'$50M revenue' and '30% market share'** — WHERE? No SEC filing, no Crunchbase link, no source. Likely fabricated. 
2. **'Poor UX' and 'limited features'** — vague opinions, not intelligence. Test their product and document specifics. 
3. **'Build better and outspend'** — generic playbook, no WHO/WHEN/COST. 

Rewrite with verifiable sources, measured weaknesses, and execution plan.
```

**Prompt:** 
```
Competitor: Acme CI/CD. Evidence: pricing page shows $49/user/month (verified 2024-01-20). G2 reviews #14523, #14601, #14789: 'slow builds', 'no ARM support', 'expensive for small teams'. Our testing (2024-01-18): build time 12min vs our 3min for identical repo. GitHub issues: 234 open about ARM (#4521, #4589). Job postings: 0 DevRel, 3 enterprise sales. Weakness: no ARM (verified), 4x slower builds (tested), no free tier (pricing). Attack: 'Migrate from Acme' page targeting 'acme...
```

**Response:** 
```
Verdict: STRATEGY_PROVEN. 

*   **Evidence-based:** Pricing verified, G2 with IDs, hands-on testing with dates, GitHub issues cited. 
*   **Weaknesses quantified:** No ARM (234 issues), 4x slower (tested), no free tier (pricing). 
*   **Strategy feasible:** 1 engineer, 5 days, $0. 
*   **Self-aware:** SOC2/SAML gap acknowledged. 
*   **Differentiated:** 'acme alternative ARM' keyword. 
*   **Metrics:** Includes kill criterion. 

Execute.
```

**Prompt:** 
```
Competitor Y's onboarding has issues (we heard from users). They're expensive. We should target their unhappy customers with a better price and better experience.
```

**Response:** 
```
Verdict: VAGUE_WEAKNESS. Three gaps: 

1. **'Has issues' and 'expensive'** — WHAT issues? HOW expensive vs you? No G2 reviews, no testing, no pricing comparison. 
2. **'We heard from users'** is not evidence. Cite specific Reddit threads, G2 reviews with IDs, or your own testing with dates. 
3. **'Better price and better experience'** is a generic playbook — this applies to any competitor. 

Define: WHAT specific onboarding problem? HOW MANY steps vs yours? WHAT exact price comparison?
```

## Capabilities

### Verify competitor claims
Forces the AI to cite specific sources like SEC filings, G2 reviews, or direct testing for every claim.

### Quantify product weaknesses
Replaces vague opinions with measurable data like specific onboarding steps or quantified performance gaps.

### Build feasible execution plans
Demands a clear breakdown of who will execute a task, the timeline, the cost, and the necessary dependencies.

### Identify your own vulnerabilities
Forces the agent to acknowledge your product's gaps and how a competitor might exploit them.

### Create unique market wedges
Pinpoints specific competitive gaps that only your specific team has the resources to attack.

### Set kill criteria and deadlines
Ensures every strategy has a measurable success metric and a hard date to determine if the plan is working.

## Use Cases

### The 'Fake Revenue' Audit
An analyst wants to know if a rival is growing. The agent uses the tool to find actual pricing and review trends instead of guessing market share.

### The 'Better UX' Trap
A PM wants to beat a competitor's design. The tool forces the agent to find specific friction points in the rival's flow to create a real roadmap.

### The 'Build it All' Fallacy
A founder wants to build a massive platform. The tool flags the lack of resources and demands a smaller, feasible MVP with a clear timeline.

### The 'Switching' Campaign
A marketer wants to poach users. The tool helps identify specific GitHub issues or G2 complaints to target with a specific migration offer.

## Benefits

- Stop hallucinations: validate_competitive_intel ensures every dollar and market share claim has a cited source like a Crunchbase link or SEC filing.
- Get actionable data: Instead of 'bad UX,' you get '14-step onboarding,' making it much easier for your engineering team to prioritize fixes.
- Avoid fantasy plans: The tool forces you to define who, when, and at what cost a strategy will happen, preventing wasted effort on impossible goals.
- Stay self-aware: It forces the AI to acknowledge your own missing features so you don't get blindsided by a competitor's move.
- Create unique wedges: It helps you find the one specific gap that only your team can realistically fill based on your current resources.
- Hard success metrics: You get clear kill criteria and deadlines instead of a 'we'll see what happens' attitude for your new initiatives.

## How It Works

The bottom line is it turns hallucinated market research into a verifiable battle plan.

1. Provide your AI agent with a competitor name or a draft of your current market research.
2. Run the validator to check for verifiable sources, specific metrics, and resource feasibility.
3. Receive a verdict that either approves the strategy or demands specific evidence to move forward.

## Frequently Asked Questions

**Does the Competitive Intelligence Prover help me find real competitor data?**
Yes, it forces the AI to find and cite specific sources like pricing pages, G2 reviews, and GitHub issues instead of guessing market share.

**How does this stop the AI from making up fake numbers?**
It requires every claim to be backed by a verifiable source, which prevents the AI from hallucinating revenue or market share figures.

**Can I use the Competitive Intelligence Prover to build a marketing plan?**
It helps you build a feasible strategy by requiring you to define the cost, timeline, and specific resources needed for each step.

**What happens if my AI gives a vague answer about a competitor's weakness?**
The Connector will flag it as a vague weakness and ask for measurable, observable data like specific onboarding steps or user review IDs.

**Does this tool help identify my own product's flaws?**
Yes, it forces a self-aware assessment where the AI must acknowledge your own weaknesses and how a competitor might exploit them.

**How does this help with kill criteria?**
It ensures your strategy includes a deadline and a specific metric to determine if the plan is actually working or if you need to pivot.

**How does it prevent hallucination?**
The engine detects 16+ hallucination markers: 'approximately', 'reportedly', 'sources suggest', 'believed to', 'likely has', 'probably', etc. Any hedging language triggers rejection. Every claim must cite a verifiable source: their website, G2/Capterra reviews with IDs, Crunchbase, SEC filings, Reddit threads, or your own hands-on testing. If you can't cite the source, you must delete the claim.

**What sources are acceptable?**
Seven categories: (1) Their website — pricing pages, feature docs, changelogs. (2) Review platforms — G2, Capterra, TrustRadius with specific review IDs. (3) Public financials — Crunchbase funding, SEC filings, press releases. (4) User complaints — Reddit threads, GitHub issues, support forums. (5) Job postings — reveals tech stack, team gaps, growth areas. (6) Your own testing — 'We signed up and tested on [date].' (7) App Store/Play Store reviews. Minimum 3 independent sources to triangulate.

**Why does it require self-assessment?**
Because competitors also analyze YOU. An analysis that only shows 'they're bad, we're good' is self-deception that leads to strategic blindness. The tool forces you to acknowledge: (1) Where are YOU vulnerable? (2) What do THEY do better? (3) How would THEY attack your weak points? (4) What blind spots exist? The best competitive strategies are brutally honest about both sides.