What is an FDE?
Define the role through the outcome chain and delivery evidence.
A living bilingual handbook for turning AI capability into an operable, adopted, and transferable business outcome.
Start with accountability, process, evidence, and two working templates, then expand into technology, organizations, jobs, and field cases.
Where accountability begins and ends
How a project moves from field discovery to transfer
How to prove launch, adoption, and outcomes
Record stage claims, five evidence layers, gates, and publication controls
Use it for kickoff, weekly review, stage gates, and transfer
Cases, playbooks, templates, jobs, and community feedback live inside the relevant chapters instead of becoming disconnected top-level channels.
Define the role through the outcome chain and delivery evidence.
Why AI products have renewed the need for high-touch delivery.
Compare product, solutions, implementation, customer success, and Applied AI roles.
How procurement, governance, private deployment, and organizations change delivery.
From entering the field and defining the problem to operations and transfer.
Data, permissions, legacy systems, procurement, compliance, and adoption friction.
When autonomy helps and when a deterministic workflow is enough.
Knowledge boundaries, permissions, freshness, citations, and retrieval evals.
Replace demo impressions with failure modes, datasets, thresholds, and regression.
Least privilege, human approval, audit, and high-risk actions.
Interfaces, rollback, SLOs, incidents, adoption, and transfer.
Prove completion through problem, solution, quality, operations, outcomes, and transfer.
Map capability to artifacts, project evidence, and observable behavior.
Move from engineering, data, product, solutions, consulting, or internal IT into FDE work.
Read the real accountability, not only the job title.
Create product feedback without falling into unlimited custom services.
Explain why demos fail to become durable outcomes.
Compare projects with a shared evidence card instead of packaging success stories.
Keep cases, methods, templates, and role maps correct through real field feedback.
Every important claim in the handbook should eventually resolve to reviewable material.
Real users, workflow, baseline, and loss
Constraints, architecture decisions, assumptions, and alternatives
Evals, failure modes, safety, and launch thresholds
Release, adoption, reliability, cost, and incidents
Business results, confirmed ownership, and operating capability
The project evidence card and 10-step delivery canvas are ready for kickoff, review, and retrospectives. Later appendices will add evals, permissions, launch gates, capability models, job maps, and correction logs.