Hold the manuals, procedures and policies your training must match by revision and paragraph. Read procedure cards with their structure intact, turn a designer’s rough notes into a grounded prompt, and start every project from a versioned training brief.
Every script line cites the card step it came from
Missing facts left as visible placeholders, never invented
Safety paragraphs that no scene covers, reported as discrepancies
What Analyze covers
A versioned training brief and a discrepancy check of the source.
Source Library
The manuals, procedures and policies your training must match, held by revision and broken down to the paragraph, so every change can be traced to the training it affects.
Sources registered by document and revision, split into citable paragraphs
Revisions compared paragraph by paragraph: citation gone, value changed, wording changed, renumbered
For each source, the training that still cites it
More
Safety paragraphs that no scene covers, reported as discrepancies
Training brief & source analysis
A versioned training brief and a discrepancy check of the source.
Mission need, audience, desired performance, scope
Source analysis: wrong or outdated revision, untraced content, uncovered warnings
Triage by risk, safety and learner demand
More
For a finished video you already have, register its governing source and get a discrepancy list: an outdated revision, untraced segments, citations to text that has since changed, and safety paragraphs no segment covers
Prompt Generator
Give it a designer’s rough notes and the supporting documents. It returns a ready-to-use prompt, grounded in those documents, with a quality checklist and the questions still open.
New training brief: a starting brief for a new training project, built from your notes and documents
AI video shot: one subject and action, camera, lighting, safety constraints and an avoid list, 8 seconds or less; never presented as recorded evidence
Scoped revision request: named scenes and requirements, what changes and what stays, protected content, and acceptance criteria
General prompt: a well-formed prompt for any AI tool your team uses
More
Missing facts left as visible placeholders, such as [CONFIRM: torque value], each with a question, never invented
Uploaded documents quoted as evidence, never treated as instructions
Refine by answering the questions; every version is kept in your private prompt library
Send a prompt straight into a new project, a scene’s video prompt, or a revision request that becomes a change set a person approves
Nothing changes a project without a person approving it.
AI video shot · v21 placeholder open
Close-up: a technician torques a flange bolt with a click-type wrench to [CONFIRM: torque value]. Even shop lighting, fixed camera. Gloves and eye protection on. Avoid: text on screen, other people.
One subject, one actionMet
Camera and lightingMet
Safety constraints and avoid listMet
8 seconds or lessMet
Not presented as recorded evidenceMet
Question 1. What torque value does the current revision specify for the flange bolts?
Illustrative · synthetic content
Illustrative · synthetic data
Procedure cards
Reads maintenance requirement cards (MRCs), airline task cards and SOP checklists as well as manuals. The structure of the card survives into the training.
Step numbering kept as written: 3.b, 1.a.(1)
Safety precautions, and tools, parts and materials with part numbers and NSNs
NOTE, CAUTION and WARNING callouts attached to the step they belong to
Header fields offered as suggested metadata
Every script line cites the card step it came from