From 33 weeks to 13 weeks: What causes the change?
Traditionally, building a single course requires weeks or even months of multi-layered work, including research, storyboarding, script writing, media production, and multiple reviews. For most learning and development teams, that schedule was simply an accepted cost of doing business. New compliance requirements, product launches, or changing market conditions may require new training, but the instructional design pipeline moved at its own pace, no matter how urgent the need.
Bottlenecks rarely originate from a single point of failure. It was born out of an accumulation of manual steps, including subject matter experts with limited availability, instructional designers building modules from scratch, and review cycles that revealed revisions late in the process. Each step took time, and the risk of miscommunication and rework increased with each handoff between personnel.
This was less important if the training content did not change frequently. A course that is built once and left largely untouched for a year or two is a reasonable model if the underlying subject matter is also stable. That assumption no longer holds true in most industries. Products iterate faster, regulations change more frequently, and the skills employees need for their jobs change on a rolling basis rather than on a predictable annual cycle. Development processes built for a slow-moving world are finding it increasingly difficult to keep up with a faster-moving world.
Do you still need an authoring tool? How AI is reshaping enterprise learning
As AI transforms the way we create learning content, are traditional authoring tools still the right choice? Join this webinar to explore how AI is reshaping course development and what to consider before investing in your next learning technology.
What does AI actually automate?
The changes underway are not just about AI being able to generate text faster than humans can type. What has changed is that AI tools can now handle much of the structural work that used to take up the majority of development time. This means converting course outlines into modules and lessons, converting existing documents and videos into structured lesson content, creating assessment questions that align with learning objectives, and suggesting a logical ordering of material based on complexity and dependencies.
That distinction is important. Content generation alone won’t solve the bottleneck if humans still need to assemble everything into a coherent course. A more meaningful transition is to tools that handle course creation, not just content creation. This means the scaffolding, structure, and initial construction happens automatically, allowing people to focus on refinement, accuracy, and decisions that AI cannot yet make on its own.
This also changes who can meaningfully contribute to course development. When the most difficult tasks like structuring modules, drafting initial assessments, and organizing sequences are done automatically, subject matter experts who are not trained as instructional designers can participate more directly in building the training, rather than handing off their expertise to another team and hoping nothing gets lost in translation. This has a knock-on effect on accuracy, as those who understand the subject best are closer to the finished product.
What the data shows
The 2025 Business Value Study conducted by IDC is based on in-depth interviews with nine organizations using the CYPHER Learning platform and provides a useful benchmark of how transformative this shift will be. Interviewed organizations reported that the average time it takes to build a new course decreased from 33.1 weeks to 13.4 weeks, a 60% improvement. Over the same period, these organizations increased the average number of courses they offer by 4.5 times.
Combining these two numbers tells a more interesting story than using either one alone. Faster course creation doesn’t just mean the same deliverables arrive faster. This meant that organizations were creating significantly more training content without proportionally growing their teams. An IDC study found that curriculum design teams improved efficiency by an average of 65%, produced 119% more courses per team member, and more than doubled output per person.
Why speed isn’t the only thing that matters
It’s tempting to treat course creation speed as a vanity metric, a number that looks great on slide decks but doesn’t reflect deeper value. It’s more useful to think of it as responsiveness. Organizations that can launch new training in days rather than months are those that can actually respond to product updates, policy changes, or newly identified skills gaps while they are still relevant.
One organization interviewed in the IDC study described building an entire AI-assisted learning system that reduced course start times from eight weeks to four weeks. Another researcher explained that 3 to 5 days per course had been reduced to 1 day. These are not small profits. These represent a fundamentally different relationship between when a training need is identified and when it can actually be addressed.
What this means for L&D teams evaluating options
For teams currently evaluating AI-assisted authoring tools, speed claims themselves are less important than understanding what actually creates that speed. It’s worth asking: Does the tool generate a complete course structure, or just isolated content that must be assembled manually? Can it be built from your organization’s own existing materials (documents, videos, policy files), or does it just generate content from scratch? And how much manual rework will a typical team report require once the AI-generated draft is complete?
The answers to these questions are more important than a single percentage improvement because they determine whether the speed increase is real and reproducible across your organization’s specific content, or whether it is a best-case scenario that is not really the case. It’s also worth asking how the tool handles precision. Speeding up course creation is only valuable if the resulting content is still correct, and a reliable AI-assisted authoring workflow should include some mechanism to review or flag AI-generated material before it reaches learners, rather than treating the first draft as the final draft.
Impact on teams beyond individual course timelines
It’s worth separating out two related but different benefits: the time it takes to build a single course and the total capacity of a team over a year. IDC research captures both. Build time for individual courses has been reduced by 60%. However, team-level effects were probably more important. The curriculum design team had to reduce the full-time equivalent headcount required to create the same amount of courses by 65%, and the content creation team by 25%. Combined, this resulted in a 119% increase in the number of courses created per team member and more than a 2x increase in individual outcomes.
Especially for smaller L&D functions, this type of leverage can be the difference between being able to support the training needs of a growing organization and constantly falling behind. One organization in the IDC study described being able to manage thousands of courses across hundreds of customers by a single person, and directly attributed that scale to the platform’s ability to automate tasks that would have required a much larger team.
where is this going
As AI tools continue to mature, the gap between identifying training needs and getting actionable content in front of learners is likely to continue to narrow. This has implications far beyond L&D efficiency metrics. It changes the realistic expectations of training capabilities in the first place and shifts the conversation from “how do we ultimately get to this content” to “how quickly can we respond?” For organizations operating in fast-changing industries, that change can ultimately outweigh the time it takes to build an individual course.
Share with
