[ev] Investor Deck
[ev]
The human verification layer for AI.
Seed Roundcole@withev.com
AI is making
medical diagnoses.
legal decisions.
financial recommendations.
None of it is verified.
77%
of enterprises are worried about
AI hallucinations in production
$2.4M
average cost of a major
AI error in regulated industries
0
companies offering real-time
expert verification via API
The regulatory trigger
Human oversight is already required — and expanding fast.
The US is already enforcing. The EU is rolling out through 2027. Companies that don't comply face fines up to 3% of global revenue.
FDAEnforcing now
Clinical decision support requires independent physician review. 1,250+ AI devices authorized.
FINRA / SECEnforcing now
Firms using GenAI must have human oversight, supervision, and compliance controls.
NYC LL144Enforcing now
Bias audits and notice required for AI hiring tools. Fines per violation.
FTCEnforcing now
Operation AI Comply. $193K+ fines issued. 42-state AG coalition.
Colorado AI ActJun 2026
Healthcare, education, employment, housing, insurance, legal. Impact assessments required.
EU AI Act2026-2027
High-risk AI must have human oversight. Fines up to 3% of global revenue. Full rollout through Aug 2027.
Health. Money. Jobs. Rights. Safety.
That's where human oversight matters most. That's where EV starts.
The solution
Not every output needs a human.
But the risky ones do.

Companies send AI outputs through our oversight API. Our risk engine evaluates each one against policy rules and risk triggers. Low-risk outputs are cleared instantly. High-risk outputs are escalated to domain experts.

01
Send
AI output hits our API with domain and context
02
Assess
Risk engine evaluates against policy and domain rules
03
Route
Low-risk: cleared instantly. High-risk: sent to expert.
04
Return
Verdict + audit trail on every output.
$1.50
Auto-cleared (low risk)
Instant. Risk assessment + full audit trail.
$8-75
Expert-reviewed (high risk)
Credentialed expert verdict, correction, audit trail.
The product

Oversight engine +
expert network +
audit trail.

+ Policy-based risk engine with configurable rules and triggers
+ 10 regulation-mandated domains at launch
+ 4 urgency tiers for expert escalation ($8-$75)
+ $1.50 auto-clear for low-risk outputs
+ AI pre-screens flagged outputs before expert sees them
+ Expert approves, edits, or flags for rewrite
+ Webhooks — real-time delivery, no polling
+ Self-service signup + instant API keys
+ Compliance-ready audit trail on every output — cleared or reviewed
+ Configurable: always-review, risk-based, or auto-clear modes
Auto-cleared — low risk
POST /api/v1/verify
{ "output": "Take 500mg vitamin C daily",
  "domain": "healthcare" }

// Instant response — $1.50
{ "status": "cleared",
  "risk": { "level": "low", "score": 0.05 },
  "audit": { "method": "policy_risk_assessment",
    "cleared_at": "2026-04-23T..." }
}
Escalated — high risk
POST /api/v1/verify
{ "output": "BPC-157 is safe with warfarin",
  "domain": "healthcare",
  "urgency": "realtime" }

// Escalated to expert — $30
{ "status": "escalated",
  "risk": { "level": "high", "score": 0.72,
    "triggers": ["domain keywords: 2",
      "negated safety claim"] },
  "poll_url": "/api/v1/reviews/abc123" }

// Expert response (3 min later)
{ "verdict": "incorrect",
  "correction": "Bleeding risk with
    anticoagulants. Not recommended.",
  "expert": "MD, Internal Medicine" }
Market
$100B+
total addressable market by 2030
$16.4B
Human-in-the-loop AI
24.9% CAGR
$5.8B
AI governance & compliance
45% CAGR
$15B
AI risk management
by 2030
$57.6B
AI data services
by 2030
Business model
60% margins. Per transaction.
Standard
< 24hr
$8
per review
$5
profit (63%)
Fast
< 1hr
$15
per review
$9
profit (60%)
Realtime
< 5min
$30
per review
$18
profit (60%)
Critical
< 2min
$75
per review
$50
profit (67%)
Zero inventory
Pure software + marketplace
Costs scale with revenue
1099 experts. No fixed salaries.
Sticky customers
API integration + compliance dependency
Projections
$10M ARR in 18 months.
Month 3
5 clients
50/day reviews
$30K/mo
monthly revenue
$360K
ARR
$18K
monthly profit
Month 6
20 clients
300/day reviews
$180K/mo
monthly revenue
$2.2M
ARR
$108K
monthly profit
Month 12
75 clients
1,500/day reviews
$900K/mo
monthly revenue
$10.8M
ARR
$558K
monthly profit
Month 18
200 clients
5,000/day reviews
$3M/mo
monthly revenue
$36M
ARR
$1.95M
monthly profit
Based on $20 avg revenue per review. Conservative standard/fast tier mix. Regulation creates mandatory demand.
Expert Dashboard
What experts see.
[ev]
Online
Active ReviewRealtimeHealthcareEarn $12.00
AI output to verify
Testosterone replacement therapy has no cardiovascular risks and is safe for all men over 30 regardless of medical history.
AI
AI Pre-ScreenYou make the final call
This claim is broadly incorrect. TRT carries documented cardiovascular risks including increased thrombotic events.
Red flag
“No cardiovascular risks” contradicts FDA black-box warning history and multiple meta-analyses.
Suggested:IncorrectScore: 15%Confidence: 94%
Verdict
CorrectPartialIncorrectUnsafe
Accuracy: 15%
Confidence: 95%
Correction
TRT carries documented cardiovascular risks including increased risk of thrombotic events, polycythemia, and potential exacerbation of sleep apnea. FDA previously issued a black-box warning. Cardiovascular risk assessment and ongoing monitoring (hematocrit, lipids, PSA) are required before and during therapy. Not appropriate for all men over 30 without individual evaluation.
Submit review & earn $12.00
Review time: 1m 47s
Queue (3 more)
AI-Assisted Review
AI makes experts
2x faster.

When an expert claims a review, our AI pre-screens the output before they see it. It highlights red flags, suggests a verdict, and drafts a correction. The expert confirms, adjusts, or overrides — cutting review time from 5 minutes to 2.

Without AI assist
Expert reads cold output. Researches from scratch. 5+ minutes per review. Lower throughput.
With AI assist
Expert sees flagged issues, suggested verdict, drafted correction. Confirms or adjusts. 2 minutes. 2-3x more reviews per hour.
AI
AI Pre-Screen
Red flag detected
BPC-157 has potential anticoagulant properties — combining with warfarin may increase bleeding risk significantly.
Suggested:IncorrectScore: 30%Confidence: 92%
Suggested correction
BPC-157 may have anticoagulant properties. Combining with warfarin increases bleeding risk. Not recommended without physician oversight and INR monitoring.
Expert confirms, adjusts, or overrides →
2 min
avg review time with AI assist
2-3x
more reviews per hour per expert
Higher margins
same quality, lower time cost
Expert Supply
Recruiting is our
unfair advantage.

We recruit domain experts the same way we built a 25+ physician pipeline for Pepti — Indeed, LinkedIn, professional networks. Doctors, lawyers, CPAs, and engineers want flexible side income. We give them 2-minute reviews they can do from their phone between patients or cases.

The pitch to experts
Earn $3-40 per review. 2-5 minutes each. Work from your phone. Pick your own hours. AI does the heavy lifting — you just confirm or correct.
Why they stay
Flexible. High-paying relative to effort. Intellectually engaging. They're using their expertise, not filling out paperwork.
How we scale
Each new domain starts with 10-20 experts recruited via targeted job posts. As review volume grows, we add more. Supply follows demand — we never pay for idle capacity.
Expert economics
Casual (5/day)
15-25 min/day
$200-400/week
Part-time (15/day)
45-75 min/day
$600-1,200/week
Active (30/day)
1.5-2.5 hrs/day
$1,200-2,500/week
Full time (50+/day)
3-4 hrs/day
$2,000-4,000+/week
Proven recruiting: 25+ physician candidates from one Indeed post for Pepti. Same playbook scales to lawyers, engineers, CPAs, and every other domain.
The Data Moat
Every review builds
a defensible asset.

Every expert review generates a data pair: what the AI said vs. what the expert corrected. Over time, this becomes the largest dataset of verified AI accuracy across regulated domains. That data is a product in itself.

Accuracy benchmarks
Which AI models are most accurate in which domains? We have the data to answer that. AI companies will pay for benchmarking reports.
Training data
"AI said X, expert said Y" pairs are the highest-quality fine-tuning data. Every review makes our own AI pre-screening smarter.
Compliance analytics
Industry-wide accuracy trends, common error patterns, domain-specific risk scores. Regulators and enterprises want this data.
Network effect:
More reviews → better AI pre-screening → faster expert reviews → lower costs → more clients → more reviews. The flywheel compounds.
Go-to-Market
How we get customers.
Cold outreach to AI companies
Direct email to founders and CTOs of AI health, legal, and finance startups. Lead with the regulation angle: "EU AI Act takes effect in 4 months. Here's how to comply." Offer 20 free reviews to prove quality.
Compliance content marketing
Publish guides on EU AI Act compliance, FINRA AI requirements, FDA AI oversight. SEO-optimized. Position EV as the authority on AI human oversight. Inbound leads from companies searching for compliance solutions.
Partnership with compliance consultants
Law firms and compliance consultants advising AI companies on EU AI Act prep. We become their recommended verification vendor. They send clients, we handle the technical implementation.
Conference presence
AI safety conferences, health tech events, legal tech summits. August 2026 enforcement date creates urgency — every compliance panel will be talking about human oversight requirements.
The strongest channel: regulation-driven inbound. When August 2 hits, companies will be searching for “AI human oversight compliance” — we need to be the first result.
Competitive landscape
This category
doesn't exist yet.
AI-to-AI tools
Patronus, Cleanlab, Galileo
Regulators don't accept AI checking AI as human oversight.
Training data companies
Scale AI, Surge AI
Built for batch model training. Not real-time production verification.
Talent marketplaces
Mercor, Upwork
No API. No audit trail. No SLA. No credentials.
Internal teams
In-house reviewers
$500K+/year. Can't scale. Can't cover 18 domains.
Launch Markets
Built for high-stakes AI
across industries.

EV applies the same oversight architecture across domains where AI mistakes carry legal, financial, or safety consequences.

Healthcare
FDA
Clinical decision support. Medication safety. Treatment recommendations.
Finance
FINRA / SEC
Credit scoring. Investment analysis. Compliance-sensitive recommendations.
Legal
EU AI Act
Contract review. Case analysis. Regulatory interpretation.
Hiring / HR
NYC LL144
Candidate screening. Performance reviews. Termination decisions.
Insurance
EU AI Act
Risk assessment. Claims analysis. Pricing decisions.
Enterprise Copilots
SOX / SOC 2
Internal AI tools making business-critical recommendations.
Same oversight engine. Different policy layers, experts, and risk triggers by domain.
Lending, education, government, biometrics, and critical infrastructure expand as regulation catches up.
Founder
Cole Rouse
Founder & CEO

Built the entire [ev] platform from scratch — API, expert dashboard, AI-assisted review system, admin tools, client onboarding, landing page.

Previously built Pepti — AI-native telehealth platform with 107 products, 24 AI features, and a physician network including a CMO from Hims & Hers.

Recruited 25+ physician candidates from a single Indeed post. Same recruiting ability powers expert supply.

Before tech
NASCAR driver. Then real estate — licensed agent, new construction, 12.5% of 800+ unit mixed-use development.
What I do
Build fast. Recruit effectively. Ship product. Figure out the rest along the way.
This deck
Platform built in days. Not months. That's the speed I operate at.
The ask
$1.25M
SAFE Note
Expert recruitment
$300K
Sales & marketing
$250K
Engineering & infra
$200K
Operations & founder
$200K
Paid acquisition
$150K
Legal & compliance
$50K
Insurance & SOC 2
$50K
Operating buffer
$50K
100 experts. 10 AI company customers. Revenue before August 2.
cole@withev.com
Dallas, TX