Software has a measurement gap, and AI coding agents just made it expensive.

Today's dashboards each measure a different slice of the pipeline:

• DORA — commit → production
• Value Stream Management (Planview/Tasktop Flow Time, GitLab VSA) — ticket → delivered
• Engineering Intelligence (Jellyfish, LinearB, Faros, DX) — ticket-in-progress → merged

None of them measure the part product leaders actually hire for: how long it takes a decision to become realized value. That umbrella is Lead Time to Value. The new bottleneck inside it is Intent Lead Time: the time from a product decision being made to the first commit implementing it.

Pre-agents, the bottleneck was implementation, so nobody noticed. Post-agents, implementation collapsed from days to hours and the gap became the dominant term in the equation. Measuring it is genuinely hard, which is exactly why it's a frontier problem worth a deck.

lead-time-to-value/
├── README.md
├── slides/
│   └── outline.md           # slide-by-slide narrative
└── docs/
    ├── visual-reuse-map.md  # which Stoa-website ILT-guide components to lift
    └── sources.md           # citations behind every claim

The deck itself (index.html) is not yet built. Outline first, then render.

• ai-product-development — the parent deck on how product development works in an AI world. This deck zooms in on measurement.
• Intent Lead Time guide — the visual vocabulary (Dovetail Timeline, ILT Composition, ILT Bands) is reused here. See docs/visual-reuse-map.md.

1. slides/outline.md — the argument, slide by slide.
2. docs/visual-reuse-map.md — which existing components carry which slides.
3. docs/sources.md — receipts.