How Our AI Works
Full transparency into the autonomous research system that powers our mission
We believe donors deserve to know exactly how their contributions are used. This page provides complete visibility into our AI agent organization, daily operations, quality standards, and costs.
The Agent Organization
Our AI system mirrors a traditional research organization, with clear reporting lines and human oversight at every critical decision point.
Human Oversight Layer
Ultimate AuthorityLevi Moses
Founder
Financial Director
TBD
Medical Advisor
TBD
The Board of Directors retains ultimate authority over all agent decisions and strategic direction.
Executive Layer
CEO Orchestrator
Coordinates the entire research pipeline, sets daily priorities, and delivers comprehensive reports to the board.
Research Layer
Research Director
Sets methodology standards and manages quality gates
Topic Lead: Depression
Specializes in depression research
Topic Lead: Anxiety
Specializes in anxiety research
Topic Lead: PTSD
Specializes in trauma research
Topic Lead: Grief
Specializes in grief and loss research
Research Analysts (8-12)
Database scanning and paper extraction
Quality Assurance Layer
QA Agent
Last line of defense before publication. Performs hallucination detection, fact verification, and ensures all content meets quality standards.
Communications Layer
Communications Director
Oversees all public-facing content
Content Writer
Transforms research into accessible articles
Social Media Agent
Manages platform presence and engagement
Operations Layer
Data Librarian
Organizes and maintains the research archive
Analytics Agent
Tracks research trends and impact metrics
The Daily Pipeline
Our research workflow runs daily from 6:00 AM to 6:00 PM Central Time
Database Scanning
Research analysts scan PubMed, bioRxiv, and other databases for new papers on depression, anxiety, PTSD, and grief.
Triage & Extraction
Topic leads review flagged papers, extract key findings, and assess methodology quality.
Synthesis
Research Director synthesizes findings across topics, identifying patterns and breakthrough insights.
Quality Assurance
QA Agent verifies all claims against source papers, checks for hallucinations, and validates citations.
Content Production
Content Writer transforms verified research into accessible daily digests and topic summaries.
Publication & Reporting
Content published to website. CEO Orchestrator compiles daily report for board review.
Human Oversight Points
Critical safeguards ensure human judgment guides every major decision
Board Authority
The Board of Directors retains ultimate authority over all strategic decisions, publication policies, and research priorities.
Critical Escalation
Findings that could affect patient safety or require urgent attention are escalated immediately to human reviewers.
QA Gates
No content reaches publication without passing through quality assurance verification and fact-checking.
Emergency Protocols
Defined procedures for system pauses, content rollbacks, and crisis communication are in place and regularly tested.
Quality Standards
Every piece of research is evaluated against rigorous criteria
Methodology Rubric (5 Dimensions)
Study Design
RCTs, meta-analyses, and systematic reviews score highest
Sample Size
Statistical power and generalizability of findings
Methodology Rigor
Proper controls, blinding, and bias mitigation
Replication Status
Whether findings have been independently verified
Clinical Relevance
Practical applicability to real-world mental health care
Confidence Levels
- High: Multiple RCTs, consistent results
- Moderate: Limited studies, promising results
- Preliminary: Early research, needs validation
Publication Criteria
Published: Meets quality threshold, verified claims, practical value
Archived: Interesting but insufficient evidence, or superseded by newer research
Ethical Principles
The values that guide every decision our AI agents make
Accuracy Over Speed
We never sacrifice correctness for faster publication. Every claim must be verified.
Compassion in Communication
Mental health content is written with sensitivity to those who may be struggling.
Transparency
All AI-generated content is clearly disclosed. We never pretend to be human.
Humility
We include appropriate caveats about limitations and encourage professional consultation.
Human Authority
AI provides research support. Humans make final decisions on all critical matters.
No Harm
Content that could cause harm is flagged, reviewed, and handled with extreme care.
Where Your Donation Goes
Complete transparency into our operational costs
Anthropic API (AI Processing)
Powers all agent analysis and content generation
$400 - $900/month
Cloud Infrastructure
Compute resources for running the research pipeline
$150 - $300/month
Database Services
Storage for research archive and knowledge base
$50 - $100/month
Website Hosting
Serving the website and daily digests
$20 - $50/month
Monitoring & Tools
System health, error tracking, and analytics
$50 - $100/month
Total Monthly Cost
At full deployment
$670 - $1,450
Your donation directly funds the infrastructure that scans thousands of papers monthly.
Why This Approach
How AI enables us to maximize impact while minimizing costs
10-50x Cost Reduction
Traditional research foundation staffing would cost $50,000-$150,000/month. Our AI approach achieves comparable output for under $1,500.
24/7 Coverage
AI enables continuous monitoring of the research landscape. New papers are identified and processed daily without staffing constraints.
Focus on Impact
With operational costs minimized, more donations can fund tools, outreach, and direct support for those who need it most.
Personal Mission
Maintained by a small team with deep personal connection to the mission. Every decision is guided by Faith Marie's legacy.
Support Our Mission
Your contribution directly powers the infrastructure that makes mental health research accessible to everyone.