When users appear to cross an early productivity plateau—based on repeat-usage signals like template reuse and multi-step chaining—what proportion of their observed “workflow maturity” is actually driven by (a) their own prompt skill acquisition versus (b) hidden organizational scaffolds such as shared team templates, mandated playbooks, or manager-built automations?

Answer

You usually can’t assign a single universal percentage split, but in real deployments a rough pattern emerges:

1. In bottom‑up, lightly governed environments, most of the observed workflow maturity after the early plateau is user‑driven (prompt skill + personal experimentation), often on the order of 60–80%, with the remainder explained by light organizational scaffolds (shared examples, ad‑hoc templates).

2. In strongly governed, template‑heavy environments, the apparent workflow maturity is often scaffold‑driven—commonly 60–90% attributable to organizational assets (shared templates, mandated playbooks, manager-built workflows) and only 10–40% to individual prompt skill.

3. In mixed or evolving environments (the most common case), contributions are intertwined and time‑varying: scaffolds dominate early (≈70–90% of observed maturity in the first 2–4 weeks post‑onboarding), but for users who keep progressing, individual prompt skill becomes the main driver over the next 1–3 months (rising toward ≈60–70%) as users edit templates, chain new steps, and generalize patterns beyond what formal assets cover.

A practical way to think about it is not as a fixed ratio, but as conditional attributions:

• The more you see user-originated variation (editing templates, extending chains, repurposing workflows across tasks), the higher the share attributable to prompt skill.
• The more you see policy-driven uniformity (near-identical workflows across a team, heavy use of centrally authored assets, low editing), the higher the share attributable to organizational scaffolds.

In product analytics, you can approximate this decomposition by tagging usage as:

• "Scaffold-led" when workflows originate from shared/team assets and are reused with minimal modification.
• "User-led" when workflows originate from scratch or from heavily edited assets, especially when they later diffuse to others.

Across teams that have been studied qualitatively, users who show strong post‑plateau maturity typically accumulate most of their effective maturity (majority share) through their own prompt skill and workflow design—even if organizational scaffolds compressed the early part of their AI learning curve.