MVP v0.2

Repeatability & Scale Demonstration™

Proving the consistency and economic impact of Kav AI.

MVP Demo Specification

Version 0.2 / US Version Focus

April 2026

Kav AI Development Team

1 1. Purpose

MVP v0.2 demonstrates that Kav AI is not a one-off detection tool, but a scalable system for identifying, prioritizing, and acting on integrity risks across an industrial facility.

This version moves beyond single-anomaly demos and establishes three core capabilities:

• Repeatability — consistent detection across many assets
  
• Scale — meaningful coverage of real facilities
  
• Actionability — clear prioritization and recommended next steps

2 2. What This MVP Proves

MVP v0.2 answers the key questions an operator will ask:

2.1 Can this system scale beyond a demo?

Yes — the platform processes thousands of assets and dozens of anomalies in a single run, not just isolated findings.

2.2 Does it help me decide what to do?

Yes — every high-priority anomaly includes recommended inspection actions tied to operational workflows.

2.3 Is the output meaningful to my business?

Yes — findings are connected to risk and economic impact, not just temperature differences.

3 3. System Output Overview

3.1 Facility Scan Summary

• Assets scanned: 1,200
  
• Thermal anomalies detected: 37
  
• High-priority anomalies: 8
  
• Coverage achieved: 65% of unit in a single deployment

This establishes that Kav AI can operate at facility-relevant scale, not just isolated inspection points.

4 4. Example Finding (Representative)

Asset: Insulated process piping

Observed condition:

• Surface temperature: ~120°F
  
• Ambient temperature: ~35°F
  
• ΔT: ~85°F

Secondary signal:

• Structural steel nearby shows ΔT ~49°F

Interpretation:

• Thermal signature consistent with insulation breakdown
• Elevated likelihood of Corrosion Under Insulation (CUI)

Recommended action:

• Prioritize this location for:
  • Insulation removal
    
  • Ultrasonic thickness (UT) inspection

This demonstrates the transition from detection → diagnosis → action.

5 5. Prioritization Output

Instead of a flat list of anomalies, MVP v0.2 produces a ranked inspection set:

• High-priority locations: 8
  
• Recommended timing: Next scheduled shutdown

Selection basis:

• Thermal severity
  
• Pattern consistency
  
• Cross-asset comparison

This ensures inspection teams know where to act first, not just what exists.

6 6. Coverage Context

A key limitation of early demos is lack of context around scale. MVP v0.2 explicitly shows:

• 65% of a process unit scanned in one day
• Coverage is repeatable and extendable across the facility

This reframes the system from:

• “interesting detection tool”
  → to
  
• practical inspection coverage solution

7 7. Economic Impact Framing

Thermal anomalies are translated into business-relevant risk:

• Typical failure consequence (if unaddressed):
  • $500K – $2M event risk (shutdown, repair, lost production)
• MVP scenario estimate:
  • $750K – $1.5M avoided cost potential

This ties the output directly to business value.

8 8. Role of AI in MVP v0.2

A critical shift in this version is clarifying that:

The drone and thermal camera collect data —
Kav AI is what makes that data usable at scale.

AI responsibilities in v0.2:

• Index and organize inspection data
  
• Detect anomalies across large datasets
  
• Rank findings by priority
  
• Enable retrieval across thousands of assets

Without this layer, the workflow does not scale beyond manual review.

9 9. Evolution from v0.1

MVP v0.2 directly addresses the limitations of earlier demonstrations:

10 10. Position in MVP Roadmap

MVP v0.2 is the second step in a staged evolution:

• v0.1 — Detection
  Demonstrates anomaly identification

• v0.2 — Repeatability + Scale (current)
  Demonstrates detection at facility scale

• v0.3 — Decision Impact
  Will quantify inspection prioritization and workflow integration

• v0.4 — Closed-loop Intelligence
  Will connect detection → validation → action → feedback

11 11. Key Takeaway

MVP v0.2 proves that Kav AI is:

• Scalable — operates across hundreds to thousands of assets
  
• Repeatable — produces consistent findings, not one-off detections
  
• Actionable — outputs prioritized inspection decisions
  
• Relevant — connects technical signals to financial impact

It moves the system from:

“We found something interesting”

to:

“Here are the 8 places you should inspect next — and why.”