Reported operating use · Creator publishing and educational media
Jakob Nielsen’s reported training-video production workflow
This is a creator’s reported media-production workflow, not evidence of enterprise adoption, independently verified learning quality or Nuvepro delivery. The enterprise design below is proposed separately.
Produce a training video from an authored research article
Jakob Nielsen (personal training-video production)
Creator-reported completed production, published 13 November 2026; no enterprise customer adoption or independent quality validation established.
A published third-party example, not a Nuvepro customer case or evidence of a partnership.
The workflow at a glance
A creator turns an article into a narrated lesson
Illustrative summary of the published account
Reported work components. Their arrangement does not imply a fixed execution order or simultaneous work.
- Jakob Nielsen
Use an article as the lesson basis
Nielsen reports adapting his article about research questions into a narrated training video.
- Speech, image and video tools
Produce narration and visual assets
The account lists synthetic narration, reference images, background video, motion graphics and sound effects.
- Editing agent and author
Edit and revise the production
Agentic editing is paired with Nielsen’s editorial iterations, including changes to background ambience and sound levels.
- Creator production
Create captions and a portrait variant
Nielsen reports multilingual captions and a portrait version that reuses the same production. These are variants of one creator project.
What the source reports
Jakob Nielsen reports adapting his article about research questions into a training video. He describes synthetic narration, generated images and background clips, agent editing and motion graphics, plus editorial retakes. Captions and a portrait version extend the same production. The tool list does not establish a fixed execution order.
Reported technology: AI narration, visual generation and editing tools with author editorial work. Scope and outcomes are source-reported.
What the enterprise environment would need
The following is a proposed project design for an enterprise doing similar work. Published details are marked; the remaining systems, data and scenarios would be agreed for your project.
Proposed practice environment
Build and review one source-grounded lesson package
Illustrative project design to scope with your enterprise.
- 1 · Set up the environment
Fictional lesson and approved storyboard
Use an SME-approved fictional script and storyboard, learning objective, asset rights register, pronunciation list, caption reference and specified formats in a disposable project. Freeze their versions before generation.
- 2 · Preparation and agent contributions
AI assembles a reviewable lesson draft
Agents prepare narration, visuals, edit decisions and captions from the approved brief. Keep unsupported statements and missing asset rights unresolved; produce a preview rather than publishing.
- 3 · Independent checks
Independent content and media checks
A separate SME compares claims with the approved script and sources. An independent reviewer checks caption text and timing against the narration, pronunciation, asset permissions and required formats against prepared references.
- 4 · People decide acceptance
The learning owner accepts an exact version
The accountable learning owner accepts or rejects the independently checked script, preview, asset manifest and caption version together. Acceptance is limited to the scoped practice package and grants no publication authority.
A reviewable evidence package
A versioned lesson preview, script, asset provenance, captions, independent findings and owner disposition.
If a check fails → withhold acceptance.
An unsupported claim, caption mismatch, hidden example in a crop, stale script, missing asset permission or owner rejection withholds acceptance. Repair the affected package, rerun independent content and media checks, then return the exact revised version to the learning owner.
Failure or owner rejection returns to review
- Hold
- → 2 · Revise work
- → 3 · Rerun independent checks
- → 4 · Owner review
Operating boundary. Designed checks, not executed. Fictional materials and draft-only outputs. No public upload, live learner data, claimed learning gain or readiness proof. SME, rights and accessibility owners must confirm the reference rules.
- Approved lesson source and permitted synthetic assets
- Proposed: a training owner supplies a fictional source article, learning objectives, approved terminology and synthetic asset/voice permissions. Freeze the source version and a subject expert’s expected teaching points.
- Reviewable media drafts
- Proposed: people approve a storyboard before an agent prepares narration, visual assets, a rough cut and caption drafts in an isolated workspace. Retain source references and unresolved claims; enable no distribution endpoint.
- Independent instructional and media checks
- Proposed: a subject reviewer separate from the generating agent checks each teaching claim against the approved source. Separately check caption text and timing, visual relevance, asset permissions and the portrait variant against approved references.
- Training-owner acceptance and repair
- Proposed: the accountable owner accepts or rejects the exact independently checked media package. Rejection or a changed source withdraws acceptance; repair the affected script, media or captions, repeat independent checks and return the new version to the owner.
A proposed simulation
An illustrative practice design derived from this work, not an executed assessment.
Proposed: assemble a source-linked training-video draft and caption/portrait variants from fictional approved lesson material, retaining human instructional review.
Review gate: The training owner accepts only the exact package whose teaching claims, captions, asset permissions and variants passed independent checks. Practice acceptance grants no publication authority or learner-readiness claim.
Challenge the result: A changed source contradicts a narrated instruction after review. Withhold acceptance, repair script, audio, captions and affected variants, rerun independent checks and return the revised package to the training owner.
What needs checking
- Source boundary: this is an author’s production account. A linked video is not an inspected assessment of instructional quality, captions or learning outcomes.
- Proposed: test a correct lesson, an unsupported claim, an obsolete procedure, a missing caption segment and a portrait crop that hides a necessary example. Use expert-authored expected content and independently checked caption/timing references. Designed, not executed.
- Proposed: change the source or replace an asset after review. Revoke the old acceptance, repair every affected version and repeat source, caption and permission checks before owner review. Designed, not executed.
Review environment fidelity, human capability and agent performance separately. The enterprise’s accountable owners decide acceptance of the shared workflow and any operating use.
How an enterprise could build similar work
A proposed build order for an enterprise doing similar work, written by Nuvepro from the published account and our interpretation of the environment and checks above. It is not how Jakob Nielsen (personal training-video production) built or operates this work.
- Freeze lesson evidence and permissionsProposed: agree fictional source material, objectives, asset rights and independently prepared teaching references.
- Approve a storyboard and draft mediaProposed: people approve the outline; an agent drafts narration, visuals, rough cut and captions without a publishing endpoint.
- Check instruction and media independentlyProposed: separate reviewers verify source fidelity, captions, timing, visual meaning and permitted assets against approved references.
- Repair every affected versionProposed: reject unsupported claims, missing captions or changed sources; repair the package and rerun independent checks.
- Training owner reviews the exact packageProposed: accept or reject the checked version for the scoped practice only. Distribution and learner readiness are separate decisions.
Your systems, data, rules and accountable owners decide the real project. Agree acceptance criteria, escalation owners and operating responsibility with the people accountable for the work before anything runs on live records. Treat the first run as a rehearsal on test records: compare the draft output with an independently prepared answer, and record what the people reviewing it change.
Questions about this example
- Is this an enterprise customer or FDE deployment?
- No. It is Jakob Nielsen’s self-reported creator production. The source does not establish an enterprise client or FDE delivery.
- Does the tool list establish a production sequence?
- No. The source describes components and editorial retakes; their arrangement must not imply a fixed execution order.
- Were the video and its captions independently checked?
- No. This curation used the public article section, not an inspection of the video, audio, captions or project files.
- What would the proposed enterprise practice produce?
- An offline draft media package with source-linked claims, captions, review results and exact-version owner disposition. Its checks are designed, not executed, and acceptance grants no distribution authority.
Shape this around your enterprise
Bring a similar workflow and the people responsible for it. Together, define the records and systems to use, the result your team will build or configure, and how its owners will check it. Nuvepro scopes the simulation, environment and preparation. Agree any existing-agent evaluation, implementation assistance and production responsibilities separately.
Sources and limitations
Curation review: . This is the date this source example was reviewed, separate from source publication dates.
- The public article section was read; the linked video, audio and project files were not inspected.
- Costs, quality and performance remain author reports. Corporate localization and full-course scenarios are projections, not delivered enterprise cases.
- The source does not establish a formal independent instructional review, caption-verification process or final approval rule. Video variants share one production account.
- The separate enterprise practice and its checks are fictional and unexecuted. It proves neither learner readiness nor permission to distribute media.
- UX Roundup, 13 November 2026: Making an Agentic Training VideoSource publication: 2026-11-13
More sales and marketing examples
- An author’s reported Substack notes planning and review workflowGenAI Unplugged (Dheeraj Sharma’s publication)
- Reported student-recruitment CRM work and unresolved handoffsPima Community College (as reported by Three Sonorans)
- AI ad campaign management: Advolve exampleAdvolve
- AI car buyer qualification: Cars24 exampleCars24