Builders Who Embed, Not Experts Who Advise
Founding engineers, not deck consultants. That's the new consulting paradigm.
The old consulting model: experts produce analysis and recommendations. Deliverables are documents. Implementation is a separate engagement (if it happens at all). Value is measured in insight quality.
This model sells the new with the old paradigm. It treats AI as another topic to analyze rather than a capability to deploy.
The Old Model
| Aspect | Traditional Consulting |
|---|---|
| Output | Decks, reports, recommendations |
| Relationship | Expert → client |
| Value metric | Insight quality |
| Implementation | "That's a separate engagement" |
| Knowledge transfer | Documents, presentations |
| Economic model | Time & materials |
| Success looks like | Client says "great analysis" |
The New Model
| Aspect | Capability Consulting |
|---|---|
| Output | Tools, infrastructure, encoded taste |
| Relationship | Founding engineer → co-builder |
| Value metric | Capability installed |
| Implementation | IS the engagement |
| Knowledge transfer | Working systems, methodology |
| Economic model | Outcome-aligned |
| Success looks like | Client doesn't need us anymore |
What "Founding Engineer" Means
We're more like temporary co-founders than consultants:
- We write code (or the AI equivalent)
- We make product decisions (not just recommend them)
- We're accountable to outcomes (not deliverables)
- We're temporary-permanent (long enough to matter, not forever)
This is different from T&M bodies. Different from strategy consulting. Closer to venture studio or EIR—but for capability, not company building.
Why the Shift Happened
AI changed the economics:
- Building is cheap → The constraint isn't "can we build" but "do we know what to build"
- Tools compound → Building our own tools makes us faster with each engagement
- Taste is the moat → Generic AI is commodity; configured AI is differentiation
- Implementation IS insight → You learn what works by building, not analyzing
When This Works Well
This model works best when there's appetite to build, not just analyze. Teams that want to ship something and learn from it. Organizations willing to make decisions without exhaustive consensus.
It's not the right fit for every situation. Sometimes you need comprehensive analysis first. Sometimes the constraint is stakeholder alignment, not capability. That's fine—different situations call for different approaches.
Implication
The shift is from "experts who advise" to "builders who embed." The value isn't in the recommendation—it's in the working system you have when we're done.