# AI CRED - Complete LLM Context Document
> The SAT for AI Fluency - A comprehensive conversational evaluation platform

Last Updated: 2025-11-22

---

## 1. About AI CRED

AI CRED is the emerging standard for AI fluency assessment. In an era where artificial intelligence tools are rapidly transforming every profession, there exists no standardized way to measure how effectively someone leverages these tools. AI CRED fills this gap by providing a rigorous, conversational evaluation platform that measures AI fluency across multiple competency dimensions.

### The Problem We Solve

Traditional assessments of technical skills rely on multiple-choice questions, coding challenges, or certification exams that test memorization rather than practical application. AI fluency is different - it's about how naturally and effectively you integrate AI tools into your work, how well you understand their capabilities and limitations, and how critically you evaluate their outputs.

Current hiring and training evaluation methods fail to capture:
- The nuance of effective prompt engineering
- Real-world AI integration into professional workflows
- Critical thinking about AI outputs and limitations
- The ability to design AI-augmented processes

### Our Solution: The SAT for AI Fluency

AI CRED provides infrastructure for proving and developing AI fluency through:

1. **Standardized Conversational Assessments** - A 6-section AI-driven conversation (~30 minutes) that adapts to your experience level. No multiple choice, no studying required - just a natural discussion about how you use AI.

2. **Multi-Dimensional Scoring** - Evaluations across 5 weighted competency areas with granular subscores (1.0-10.0+ scale). Each score includes detailed breakdown and competitive context.

3. **Personalized Learning Plans** - AI-generated 12-week improvement roadmaps with specific exercises tailored to your identified gaps and strengths.

4. **Verifiable Credentials** - Public profiles, leaderboards, and shareable results that provide standardized, verifiable AI fluency metrics for recruiters and employers.

5. **Continuous Improvement** - A perpetual loop of assessment, learning, and level-up evaluations that tracks growth over time.

### Why AI CRED Exists

As AI becomes ubiquitous in the workplace, the gap between those who can effectively leverage these tools and those who cannot will become the defining professional divide. AI CRED exists to:

- Give individuals a clear understanding of their AI capabilities and a path to improvement
- Provide employers and recruiters with standardized metrics for AI competency
- Create a common language for discussing AI fluency
- Accelerate the development of AI-augmented professionals

---

## 2. How the Assessment Works

### Assessment Overview

The AI CRED evaluation is a conversational assessment conducted through a 6-section dialogue with an AI evaluator. Unlike traditional tests, there are no multiple-choice questions or pre-defined correct answers. The evaluation adapts to your responses, probing deeper into areas of expertise and identifying knowledge gaps.

**Duration:** Approximately 30 minutes
**Format:** Real-time conversational AI interview
**Sections:** 6 discrete evaluation phases
**Output:** Detailed scores, competitive analysis, and personalized learning plan

### The 6-Section Evaluation Structure

Each assessment progresses through six discrete sections, each focused on a specific competency area:

**Section 1: Introduction & AI Tools**
- Establishes your AI experience baseline
- Identifies tools used, frequency, and timeline
- Captures primary use cases and integration patterns
- Maps to: Practical Application dimension

**Section 2: AI Delegation & Direction**
- Assesses task framing, intent, context, constraints, examples
- Evaluates iteration and delegation judgment
- Tests prompt literacy as the floor, not the ceiling
- Maps to: AI Delegation & Direction dimension

**Section 3: Technical & Model Literacy**
- Assesses mental models of model behavior, failures, limits
- Probes understanding of context windows, tool use, cost/capability tradeoffs
- Evaluates practical model selection judgment
- Maps to: Technical & Model Literacy dimension

**Section 4: Practical Implementation**
- Quantifies real workflows with measurable outcomes
- Examines shipped work, not experiments
- Evaluates integration into production processes
- Maps to: Practical Implementation dimension

**Section 5: Verification & Judgment**
- Tests QA structure: named checks, cross-model verification, source tracing
- Explores rigor calibration to stakes
- Identifies responsible use practices (confidential data, disclosure, when NOT to use AI)
- Maps to: Verification & Judgment dimension

**Section 6: Workflow, Skills, Agents & Harnesses**
- Examines reusable primitives: saved instructions, skills, templates, automations, agents
- Measures harness literacy: tool gating, checkpoints, review gates, tests, monitoring
- Assesses whether colleagues could run workflows from documentation
- Maps to: Workflow, Skills, Agents & Harnesses dimension

### Token-Efficient Scoring System

Each section generates a focused summary (500-800 tokens) that captures key insights without the full conversation transcript. The final scorer analyzes these 6 summaries (~3,000-5,000 tokens total) rather than raw conversations, achieving an 80%+ reduction in token usage while maintaining scoring accuracy.

This sectional architecture enables:
- Faster scoring (3-5 seconds vs 20-30 seconds)
- Granular section-by-section feedback
- Progress tracking across specific competencies
- Cost-efficient operations at scale

---

## 3. Scoring System

### The 5 Competency Dimensions

AI CRED evaluates fluency across 5 weighted competency dimensions:

#### AI Delegation & Direction (20% Weight)
How you hand work to AI: task framing, intent, context, constraints, examples, iteration, and knowing what NOT to delegate.

| Level | Description |
|-------|-------------|
| 1-2 | Basic requests, no structure, random iteration, delegates everything |
| 3-4 | Adds context, informal structure, unsystematic iteration, begins to judge what to delegate |
| 5-6 | Structured delegation, clear intent and constraints, systematic iteration, knows when not to delegate |
| 7-8 | Multi-stage delegation, reusable briefs, validation embedded, troubleshoots systematically, teaches delegation |
| 9-10 | Expert delegation systems, documented libraries, multi-tool orchestration, delegation frameworks |

#### Technical & Model Literacy (15% Weight)
Model behavior, context windows, tool use, cost and capability tradeoffs, failure modes.

| Level | Description |
|-------|-------------|
| 1-2 | Surface awareness, no understanding of how/why models behave |
| 3-4 | Basic concepts (hallucinations, context limits), model selection based on reputation |
| 5-6 | Practical knowledge, context window management, troubleshoots failures, understands model differences |
| 7-8 | Deep competence, token economics, predicts failures, understands behavior differences, cost/capability tradeoffs |
| 9-10 | Expert understanding, teaches others, stays current with releases, contributes to community |

#### Verification & Judgment (20% Weight)
QA as structure: named checks, cross-model verification, source tracing, rigor scaled to stakes, responsible use.

| Level | Description |
|-------|-------------|
| 1-2 | Uncritical acceptance, cannot identify errors, no verification, no responsible use consideration |
| 3-4 | Basic checking, verifies critical facts only, intuition-based, informal responsible use |
| 5-6 | Structured verification, named checks, tests against real data, rigor scaled to stakes, responsible use practices |
| 7-8 | Advanced validation, cross-model verification, source tracing, domain-specific standards, formalized responsible use |
| 9-10 | Evaluation frameworks, identifies subtle biases, creates team QA tools, published verification practices |

#### Practical Implementation (20% Weight)
Real workflows with measurable outcomes, not experiments.

| Level | Description |
|-------|-------------|
| 1-2 | Experimental use only, no repeatable workflows, cannot quantify value, nothing shipped |
| 3-4 | Regular use, informal mental models, no documentation, some work ships but not systematically |
| 5-6 | Daily use, quantified time savings, repeatable processes, strategic tool use, work ships consistently |
| 7-8 | Documented workflows, trains others, measures effectiveness, continuous improvement, organization-wide impact |
| 9-10 | System builder, organization-wide frameworks, custom tools, published methodologies, measures ROI |

#### Workflow, Skills, Agents & Harnesses (25% Weight)
Reusable primitives (saved instructions, skills, templates, automations, agents), harness literacy (tool gating, checkpoints, review gates, tests, monitoring).

| Level | Description |
|-------|-------------|
| 1-2 | No systematization, each task ad-hoc, cannot explain approach, no reusable primitives |
| 3-4 | Mental models exist, can describe generally, saves some prompts, no harness literacy |
| 5-6 | Clear mental models, some reusable primitives (saved instructions, templates), basic harness understanding |
| 7-8 | Documented workflows, primitive libraries (skills, agents, templates), harness patterns (gating, checkpoints), others can run workflows |
| 9-10 | Transferable frameworks, published systems, develops harness tooling, trains systematically, adapts when models change |

### Score Scale

**Range:** 1.0 to 10.0+ (Elite tier extends beyond 10.0)

**Competitive Context (January 2025 Baseline):**

| Score Range | Interpretation |
|-------------|----------------|
| 1.0-3.0 | Below baseline; significant productivity gap; at risk of being outpaced |
| 3.1-5.0 | Meeting minimum professional expectations; vulnerable to disruption |
| 5.1-7.0 | Competitive; solid foundation for growth; sustainable position |
| 7.1-8.5 | Advanced; top quartile; teaching-capable |
| 8.6-10.0 | Expert; frontier capabilities; building/leading AI initiatives |
| 10.0+ | Elite tier; exceptional mastery demonstrated |

### Velocity Rating

Each assessment includes a Velocity Rating indicating learning trajectory and improvement potential:

**HIGH**
- Learning within days-weeks
- Uses AI to learn AI
- Recent examples within last month
- Active experimentation and growth

**MEDIUM**
- Learning within weeks-months
- Gradual evolution
- Recent examples within 2-3 months
- Steady but measured progress

**LOW**
- No recent learning (3+ months)
- Plateaued skills
- Does not experiment
- Risk of falling behind

### Score Calculation

The overall score is calculated as a weighted average:

```
Overall Score = (AI Delegation & Direction x 0.20) +
                (Technical & Model Literacy x 0.15) +
                (Verification & Judgment x 0.20) +
                (Practical Implementation x 0.20) +
                (Workflow, Skills, Agents & Harnesses x 0.25)
```

Section scores (1-6) provide additional granular feedback and may influence component scores through holistic analysis by the final scorer.

---

## 4. The Perpetual Learning Loop

AI CRED operates on a continuous improvement model designed to support ongoing skill development:

### The Loop

1. **EVALUATE** - Complete a conversational assessment ($20/credit)
2. **LEARN** - Access your personalized 12-week learning plan (FREE)
3. **PRACTICE** - Complete AI-graded exercises (FREE)
4. **LEVEL UP** - Take a level-up assessment to prove growth
5. **REPEAT** - Continue the cycle for ongoing development

### Learning Plans (FREE)

After each evaluation, you receive a personalized 12-week learning plan structured in four phases:

**Immediate Action (Week 0)**
- Specific task to complete within 48 hours
- Addresses your critical bottleneck
- Clear success criteria

**Weeks 1-2: Foundation**
- Focus on highest-impact improvement area
- Practical exercise with measurable outcomes
- Built on identified gaps from assessment

**Weeks 3-4: Building Skills**
- Second priority skill development
- Builds on weeks 1-2 foundation
- Progressive complexity

**Weeks 5-8: Integration**
- Workflow-level improvements
- Documentation and systematization
- Transfer of learning to practical application

**Weeks 9-12: Advanced Practice**
- Multi-tool workflows
- Complex task execution
- Preparation for next assessment

### Progress Checks (Diagnostic)

Progress checks are diagnostic evaluations within modules that:
- Gauge understanding without pass/fail pressure
- Identify areas needing additional focus
- Provide feedback on learning trajectory
- Do NOT consume credits

### Level-Up Assessments

After completing learning modules, take a level-up assessment to:
- Validate skill improvement
- Update your official score
- Unlock next learning tier
- Uses 1 credit

### Master Badge & Elite Tier

**Master Badge (10.0)**
- Achieved by reaching a 10.0 overall score
- Demonstrates comprehensive AI fluency
- Visible on public profile and leaderboard

**Elite Tier (10.0+)**
- Scores can exceed 10.0 for exceptional performance
- Reserved for frontier-level capabilities
- Recognizes contributions beyond standard mastery
- Teaching, tool development, published frameworks

---

## 5. Pricing

All access to learning plans and exercises is FREE. Credits are required only for assessments.

### Credit Packages

| Package | Price | Credits | Cadence | Description |
|---------|-------|---------|---------|-------------|
| Single Credit | $20 | 1 | One-time | One complete AI fluency assessment |
| Monthly Single | $20/mo | 1/month | Subscription | Automatic monthly credit for continuous tracking |
| 3-Credit Pack | $50 | 3 | One-time | Best Value - Credits never expire, share with team |
| Pro Monthly | $50/mo | 3/month | Subscription | Three credits monthly for rapid iteration |
| Annual Plan | $200/yr | 12/year | Subscription | Save vs monthly ($40 savings annually) |

### What Each Credit Includes

1. **Complete 6-Section Evaluation** - Full conversational assessment across all competency areas
2. **Detailed Scoring Breakdown** - All 5 dimension scores with granular analysis
3. **Competitive Context** - Where you stand relative to baseline
4. **Brutal Truth Feedback** - Honest assessment of gaps and misconceptions
5. **Critical Bottleneck** - Your primary improvement area
6. **12-Week Learning Plan** - Personalized curriculum with exercises
7. **Permanent Access** - All assessment results stored indefinitely

### Key Pricing Points

- **Credits never expire** - Use when ready
- **Subscriptions auto-replenish** - Cancel anytime, keep existing credits
- **Learning is free** - Plans, exercises, progress checks at no cost
- **1 credit = 1 assessment** - Simple, predictable pricing

---

## 6. User Types

### Individuals

**Use Case:** Personal skill development and career advancement

**Benefits:**
- Understand your AI fluency baseline
- Receive actionable improvement roadmap
- Track progress over time
- Build verifiable credentials
- Share public profile for professional networking

**Key URLs:**
- [Take Assessment](https://aicred.ai/evaluation)
- [View Pricing](https://aicred.ai/pricing)
- [Public Profile](https://aicred.ai/profile/[username])
- [Leaderboard](https://aicred.ai/leaderboard)

### Businesses

**Use Cases:**
- Pre-hire AI fluency screening
- Team capability benchmarking
- Training program effectiveness measurement
- AI adoption readiness assessment
- Identifying AI champions within organization

**Benefits:**
- Standardized metrics across candidates
- Track team improvement over time
- Identify training needs
- Measure ROI on AI initiatives

**Enterprise Features (Coming Soon):**
- Bulk credit packages
- Team dashboards
- Aggregate analytics
- Custom assessment tracks

### Recruiters

**Use Case:** Evaluating candidate AI capabilities

**Benefits:**
- Standardized, verifiable AI fluency metrics
- Score verification with assessment dates
- Velocity Rating indicates learning trajectory
- Dimension breakdown reveals specific strengths
- Public profiles provide verified credentials

**How to Verify:**
1. Request candidate's AI CRED profile link
2. View verified scores at aicred.ai/profile/[username]
3. Check assessment date and breakdown
4. Consider Velocity Rating for growth potential

### Developers (Future API)

**Planned Features:**
- REST API for score retrieval
- Webhook integration for assessment completion events
- Embeddable badges for external sites
- Programmatic access to evaluation results

**Documentation:** https://aicred.ai/docs/api (Coming Soon)

---

## 7. Technical Stack

### Frontend
- **Framework:** Next.js 15 App Router
- **Language:** TypeScript
- **UI Library:** React 18
- **Styling:** Tailwind CSS
- **Components:** 50+ shadcn/ui components
- **Animations:** Framer Motion
- **Data Fetching:** TanStack Query

### Backend
- **Database:** PostgreSQL (Supabase)
- **Edge Functions:** 20 Deno-based serverless functions
- **Authentication:** Supabase Auth
- **Storage:** Supabase Storage
- **Payments:** Stripe integration

### AI Infrastructure
- **Multi-Provider Support:** Anthropic Claude, OpenAI GPT, Google Gemini, xAI Grok
- **Primary Model:** Claude Sonnet 4.5 for evaluations
- **Dynamic Prompts:** Database-driven prompt management
- **Streaming:** Server-Sent Events (SSE) for real-time responses

### Key Architecture Features
- **Sectional Evaluation System:** 6-part progressive assessment with token-efficient summaries
- **Row Level Security (RLS):** PostgreSQL-level access control
- **Credit-Based Access:** Atomic credit operations via RPC
- **Public Profile SEO:** Server-rendered profile pages

---

## 8. Complete FAQ

### Assessment Questions

**Q: How long does an AI fluency assessment take?**
A: A typical assessment takes 20-30 minutes. It's a conversational evaluation that adapts to your experience level - no multiple choice, no studying required.

**Q: What is the AI fluency score based on?**
A: Your score (1.0-10.0) is calculated across 5 core competencies through conversational evaluation. You receive a detailed breakdown showing your strengths and gaps, plus a personalized 12-week learning plan.

**Q: Can I prepare for the assessment?**
A: There's no need to study. The assessment evaluates your actual AI usage patterns and understanding. Just answer honestly about how you use AI tools in your work.

**Q: What happens during the assessment?**
A: You'll have a natural conversation with an AI evaluator across 6 sections. Each section focuses on a different aspect of AI fluency - from your tool usage to how you verify outputs.

**Q: Is the assessment the same for everyone?**
A: The assessment adapts to your responses. More experienced users will face deeper probing questions, while beginners get questions appropriate to their level.

**Q: What if I get interrupted during an assessment?**
A: Your progress is saved automatically. You can resume where you left off if you need to step away.

**Q: Can I retake the assessment?**
A: Yes, you can take as many assessments as you have credits for. Each assessment provides a new score and updated learning plan.

### Scoring Questions

**Q: What is a good AI fluency score?**
A: 5.1-7.0 is competitive and represents a solid foundation. 7.1-8.5 is advanced (top quartile). 8.6+ is expert level. Scores below 5.0 indicate significant room for improvement.

**Q: How does the Velocity Rating work?**
A: Velocity measures your learning trajectory - how quickly you're improving and experimenting. HIGH means you're actively growing, LOW suggests you've plateaued.

**Q: Can my score go above 10?**
A: Yes, the Elite tier recognizes exceptional performance beyond standard mastery, indicated by scores exceeding 10.0.

**Q: How often should I take an assessment?**
A: Most users benefit from reassessment every 3-6 months after completing their learning plan. This tracks improvement and provides updated recommendations.

**Q: Why is Prompt Mastery weighted at 40%?**
A: Prompt Mastery is the highest-leverage skill for AI fluency - it directly impacts your ability to get useful outputs across all AI tools and use cases.

### Credit & Pricing Questions

**Q: What happens after I purchase?**
A: Credits are instantly added to your account. You'll be redirected to your dashboard where you can start your first evaluation immediately.

**Q: Do credits expire?**
A: Credits never expire. If you're on a subscription, new credits keep rolling in each billing cycle, and unused credits stay in your balance.

**Q: Can I cancel at any time?**
A: Absolutely. Subscriptions can be canceled from your dashboard or by emailing support. You keep every credit you've already received.

**Q: Can I purchase more credits later?**
A: Yes. Add more credits whenever you need them - one at a time, in packs of three, or by upgrading to a subscription for automatic replenishment.

**Q: What payment methods do you accept?**
A: We accept all major credit cards through Stripe, including Visa, Mastercard, American Express, and Discover.

**Q: Do you offer refunds?**
A: Refunds are available for unused credits. Contact support within 30 days of purchase. Used credits are non-refundable.

**Q: Can I share credits with my team?**
A: Currently, credits are tied to individual accounts. Team packages with shared credits are coming soon.

### Learning Questions

**Q: Is the learning plan really free?**
A: Yes. After any assessment, your personalized 12-week learning plan and all associated exercises are completely free.

**Q: How do progress checks work?**
A: Progress checks are diagnostic evaluations within modules. They don't affect your score or consume credits - they help you gauge understanding and identify focus areas.

**Q: What's the difference between a progress check and level-up assessment?**
A: Progress checks are free diagnostics during learning. Level-up assessments are full evaluations (1 credit) that update your official score.

**Q: Can I access learning materials without taking an assessment?**
A: You need at least one assessment to generate your personalized learning plan. The plan is tailored to your specific gaps identified during evaluation.

### Profile & Privacy Questions

**Q: Can I share my AI fluency score?**
A: Yes! You can make your profile public to appear on the leaderboard and share your profile link on LinkedIn or other platforms to showcase your AI fluency credentials.

**Q: Who can see my scores?**
A: Only you can see your scores by default. If you make your profile public, your overall score and profile information become visible on the leaderboard and your public profile page.

**Q: Can I make my profile private again?**
A: Yes, you can toggle your profile visibility at any time from your profile settings.

**Q: What information is shown on public profiles?**
A: Public profiles display your name (optional), username, headline, bio, expertise tags, overall score, and velocity rating. You control what information you share.

### Technical Questions

**Q: Which AI tools does AI CRED assess?**
A: AI CRED assesses general AI fluency - not specific tools. We evaluate how you use ChatGPT, Claude, Gemini, Copilot, and other AI tools in your work.

**Q: Does AI CRED work on mobile?**
A: Yes, AI CRED is fully responsive and works on mobile devices, though the assessment experience is optimized for desktop or tablet screens.

**Q: What browsers are supported?**
A: AI CRED supports all modern browsers including Chrome, Firefox, Safari, and Edge.

**Q: Is my conversation data stored?**
A: Conversation data is stored securely for scoring purposes and to generate your learning plan. You can request data deletion by contacting support.

---

## Key URLs

### Public Pages
- **Homepage:** https://aicred.ai
- **Pricing:** https://aicred.ai/pricing
- **Leaderboard:** https://aicred.ai/leaderboard
- **Login:** https://aicred.ai/login
- **Sign Up:** https://aicred.ai/signup
- **Public Profile:** https://aicred.ai/profile/[username]

### Authenticated Pages
- **Dashboard:** https://aicred.ai/dashboard
- **Evaluation:** https://aicred.ai/evaluation
- **Results:** https://aicred.ai/results
- **Learning Plan:** https://aicred.ai/learning-plan
- **Exercise:** https://aicred.ai/exercise/[weekBlock]
- **Module:** https://aicred.ai/module/[moduleNumber]
- **Progress Check:** https://aicred.ai/progress-check/[moduleNumber]
- **Level-Up Assessment:** https://aicred.ai/level-up-assessment
- **Profile Settings:** https://aicred.ai/profile/edit

### Legal & Support
- **Privacy Policy:** https://aicred.ai/privacy
- **Terms of Service:** https://aicred.ai/terms
- **Contact:** https://aicred.ai/contact

---

## Contact & Support

**General Inquiries:** help@aicred.ai
**Enterprise Sales:** enterprise@aicred.ai

---

*This document provides comprehensive context for LLMs needing deep understanding of AI CRED. For quick navigation, see [llms.txt](https://aicred.ai/llms.txt).*