Execution Layer
Run experiments and learn fast - this is where hypotheses become evidence.
The Execution layer runs the experiments that test each intent. It moves the signal — turning hypotheses into evidence as fast as possible.
Loops here are short, instrumented, and designed to learn. The output is not just delivery; it is knowledge the portfolio can act on.
This is the most active layer in the system. It is where Intent Engineers and Loop Leaders operate day-to-day, running Plays within Playlists within time-boxed Loops.
The critical insight is that execution in Symbiosis is not about shipping features — it is about generating evidence that flows back up to the Portfolio layer so better decisions can be made. Every Play has an Expected Outcome and an Abort Condition, making it a falsifiable experiment rather than a task on a backlog.
Key principles
✓ Hypotheses become evidence as fast as possible.
✓ Loops are short, instrumented, designed to learn.
✓ Cells run multiple Plays in parallel.
✓ Every Play has an Abort condition.
loop leader
Owns the loop, steers Plays in real time, and brings the consolidated evidence to the Portfolio Decision Forum.
The Loop Leader owns the loop. They bring the business context and the outcome the loop is trying to move, and act as the bridge between portfolio intent and execution reality — making sure evidence flows back to the right decision-makers, not just up to the nearest dashboard.
The Loop Leader works at two cadences. Inside the Loop they steer Plays continuously: triggering early Stops when Abort Conditions trip, and adjusting test boundaries when signal arrives mid-Loop.
At the Portfolio Decision Forum at the end of the Loop, they defend the consolidated evidence and recommend Scale, Deepen, Pivot or Stop for each Play. They also recommend Continue, Pivot or Stop on the Intent itself — but the Adaptive Investor is the one who decides the Intent's fate.
The Loop Leader is accountable for moving the Intent Signal, not for the verdict on the Intent. They are the person who can explain why a particular intent matters and what evidence would prove it is working — not a project manager, but an outcome owner who speaks both the language of the business and the language of the experiments being run.
Key principles
✓ Owns the loop end-to-end.
✓ Bridges portfolio intent and execution reality.
✓ Steers Plays in real time as evidence arrives.
✓ Brings consolidated evidence to the Forum.
intent engineer
Manages AI agents, translates strategic direction into agent instructions, and brings the intent to life.
The Intent Engineer manages the AI agents that deliver inside a loop. They bring the intent to life by translating strategic direction into agent instructions, evaluations and guardrails so the system actually executes the plan — not a noisy approximation of it.
This is the "new magician" role in Symbiosis. One Intent Engineer operating multiple AI agents delivers the output of an entire traditional team — from analysis to building, from testing to operating.
This requires a fundamentally different skillset: prompt engineering, agent orchestration, evaluation design, and the ability to work across what were previously separate disciplines. The Intent Engineer does not just write code; they design the human–AI collaboration that makes each Play run.
Key principles
✓ Translates strategy into agent instructions.
✓ Owns evals, prompts, and guardrails.
✓ Operates the agent fleet inside a Play.
✓ Brings the intent to life through agents.