L&D teams vs course creators: how each briefs AI

Build, launch and sell quality courses in minutes

Start Free

Corporate learning teams and independent course creators open the same AI tool and type very different first messages. One is usually building training for colleagues; the other is building a product for strangers who will pay for it.

I compared the first prompts of the two groups in a random sample of 10,000 people who created a course with AI on Coursebox between May and November 2026. Corporate teams name their audience in 70% of briefs and give the AI their own documents in 44%. Course creators state what learners will be able to do in 44% of briefs, more than any other group.

What do L&D teams and course creators build with AI?

Corporate L&D teams mostly build job skills (32%), compliance and safety (26%) and onboarding (18%) training, according to Coursebox's analysis of 10,000 classified first prompts (May to November 2026). Course creators mostly build courses about earning income (46%), personal development (25%) and job skills (19%).

The learners differ just as much. 94% of corporate prompts are for employees, while 88% of creator prompts are for the general public. Corporate courses run inside an organisation that already knows its learners; a creator's course has to win them first.

How do their AI briefs differ?

Corporate L&D briefs lead with the audience and the assessment, while course creator briefs lead with the outcome and the structure. Corporate teams name the audience in 70% of first prompts and ask for assessment in 30%; creators state learning outcomes in 44% and set out their own module outline in 26%.

Both patterns make sense for the job. A corporate course has to land with a defined group and prove it worked, so audience and assessment come first. A paid course is sold on its promise, so the outcome and the shape of the course come first.

Who gives the AI their own material?

Corporate L&D teams are more than twice as likely as course creators to build from their own documents: 44% of corporate briefs point the AI at source material, against 17% for creators. 34% of corporate first prompts arrive with a file attached, against 8% for creators.

Creators still write the longer briefs, a median of 74 words against 57. Their method, examples and stories go into the prompt itself rather than into an attached document.

Which languages do they work in?

Course creators on Coursebox work in Portuguese as often as in English: 39% of creator first prompts are in Portuguese and 37% in English. Corporate L&D work is more English-led, at 49% English and 17% Portuguese.

For anyone selling courses, the creator market is already multilingual. A course written natively in Portuguese reaches an audience that a translated English course may not.

What can each group borrow from the other?

Each group already does something the other could use. These are the habits I would borrow, based on what the most complete briefs in the data had in common:

• If you are in a corporate team: state what learners should be able to do by the end. Only 27% of corporate briefs do, and it decides what every module is for.
• If you are in a corporate team: sketch the module outline you want. Creators do this twice as often, and it keeps the AI close to the structure you already have in mind.
• If you create courses to sell: attach your own material. Your notes, slides or transcripts carry the method a buyer is paying for, and only 17% of creator briefs include any.
• If you create courses to sell: ask for assessment. 12% of creator briefs do, and a quiz or exercise per module gives buyers proof of progress.
• For both: name the audience. It is the single detail most likely to change the course the AI produces.

How I measured this

The data is anonymised and aggregated. No names, email addresses or company details were used in the analysis, no prompt text is published, and any group smaller than 30 is left out. It covers the first prompts of 37,284 people on Coursebox between 14 May and 9 November 2026, with bots, staff and test accounts removed. Categories come from an AI classification of a random sample of 10,000 first prompts written in people's own words; a manual check of 40 classifications from each of the two AI models found about 92% correct, and 304 prompts labelled by both were checked by hand to calibrate the design-decision figures. Corporate L&D and course creator groups are people whose role was clear from their prompt: 496 and 642 in the full sample, 324 and 413 in the calibrated design-decision sample. The full methodology is in the report.