AI is used today in large parts of course production. That is why the most important question is not whether a training agency uses AI, but how the technology is used, how the result is quality-assured and where the line is drawn. Here we describe how we work.
What we automate
The heavy lifting in the script
AI helps us transcribe expert interviews, structure extensive source material and produce first drafts of scripts and voice-over text. That makes production faster, but every script is reviewed and developed further by both instructional designers and subject-matter experts before it becomes a finished training.
Images with a shared expression
In Graphics Studio, the image workshop in Care of Skills Create™, we produce illustrations based on a shared style, chosen colours and visual references. The images are saved with their recipes so we can build on them when new course parts are added. We review both what each image shows and how the series works together. We have developed the interface and workflow for the producer's work with components and modules. Image generation and AI processing are performed by external services such as Recraft and fal.ai, independent products from other vendors. Read more about how we work with consistent AI images.
Audio and translation
AI voices are often used for drafts and versioning, and in more and more projects also for finished productions when the quality is high enough. For translations, AI does the first draft, while the final language review is always done by a person who has the language as their mother tongue.
Variants and updates
Code-based video production makes it possible to create a large number of versions from the same template. AI also helps generate, for example, alt texts, summaries and quiz suggestions from finished content. Work that previously took hours can often be done in a few minutes.
What we never let go of
The instructional design
What participants should be able to do after the training, how the content should be structured and which exercises and examples give the best effect is still human work. The instructional design is developed together with the client and starts from the audience's real needs. AI can create content about a subject, but it does not know which challenges your employees face in everyday work.
Fact checking
All AI-generated material is reviewed by both people and subject-matter experts before it is used in a training. That applies whether the content consists of text, images, audio or video. In training where correct information is critical, there is no room for inaccuracies.
Judgement
Tone, relevant examples and the ability to cut what does not add value cannot be automated. Quality is often less about adding more content and more about creating focus by removing what is not needed.
The risk everyone in the industry should talk about
When it becomes easier and cheaper to produce training, the amount of content also increases. But more courses do not automatically mean better learning. On the contrary, organisations risk being filled with training that takes time without creating real value. AI's greatest strength is therefore not to produce more, but to free up time for better pedagogy, higher quality and more targeted training.
What it means for your budget
A smaller share of the budget needs to be spent on manual production steps and a larger share can be used where it makes the greatest difference — for example analysis, instructional design, scenario work and follow-up. At the same time, production goes faster. When a first draft can be produced in a few days instead of several weeks, there is more time for review, improvements and quality assurance. That is often where the best training is shaped.
Want to see what it looks like in practice? We are happy to show a real production flow, with AI's contribution and human decisions side by side. Get in touch.