The AI Hit
Artificial intelligence did not arrive to assist instructional design. It arrived to absorb it. One multi-skilled ID now completes in an hour what previously took a ten-man team forty hours—a ninety percent reduction in both manpower and development time, collapsing a $100K course budget to $10K. This page marks the target, names the tasks already taken, and sets out the moves that keep an instructional designer working.

The ISD mandate has two halves: identify all corporate-wide business gaps and problems, then develop the training strategies that align employee goals with the business vision. The first half is judgment work performed among people who will not tell you the truth on the first pass. The second half is production work—outlines, storyboards, scripts, graphics, narration, video, quizzes, captions, grading, reporting. AI came for the second half and took it in under three years. Seventy to ninety percent of corporate training was already Word and PowerPoint converted into electronic page-turners; that work was automation-ready long before the models arrived.
The terminal objective is the unifying thread—the golden thread binding every instructional phase together. A model can generate any single phase; it cannot hold the thread. Survival means owning the analysis, becoming the one-man team of twenty, treating prompt engineering as a design discipline, designing the practice rather than the content, and moving up into measurement where ROI, KPI, and Success Case Method prove business impact. The role did not disappear—it consolidated. One seat survives per former team, and it goes to the designer who can operate every tool in the chain.
Ten people, forty hours. Programmers, graphic artists, voice talent, video crew, QA, and project managers, with a designer somewhere in the middle. A hundred-thousand-dollar course budget was normal, and the meetings alone justified it. Large teams of eLearning developers boasting years of experience never guaranteed a sound outcome—only a larger invoice.
One person, one hour. A single multi-skilled ID now runs the whole stack—authoring, graphics, audio, video, data, and LMS. The same course ships at a tenth of the cost. This is the Power of AI at 10X: a one-man team of twenty, where the surviving seat goes to the generalist rather than the specialist who owned one link in the production chain.
Sorted by how much of the task a current tool set can carry without a human in the loop. The bar measures exposure, not quality—a task at 90% still needs a designer to sign it, but it no longer needs a specialist to make it.
Production tasks a model performs end to end. Do not sell these as a service, and do not put them on a résumé as a skill.
The model produces a usable first draft; a designer earns their seat on the second. This is where the work is right now—and where it will not stay.
Work that requires standing in a room, reading a person, and being accountable for a number. Build your job here.
No model walks the plant floor, sits in the incident review, or notices that the "training problem" is actually a broken process. Identify the gap yourself and you own the project before a prompt is written.
AI generates parts. You own the whole. Set one terminal objective, tie every lesson, practice and assessment to it, and be the only person in the room who can say whether the course coheres.
The surviving seat goes to the generalist. Learn the whole chain—authoring, graphics, audio, video, data, LMS—well enough to direct it. Specialists in a single production link are Tier 1.
Prompt engineering is instructional design pointed at a machine: precise objectives, constraints, examples, feedback. Designers already have the skill. Most have not noticed.
Content is free now. Mastery is not. Error-and-solution feedback, consequence-driven responses, and practice that strengthens neural pathways still change behavior on the job.
Anatomy views, exploded diagrams, and infographics that place each part against the whole. Architecture is a judgment about what a learner needs to see first, and it remains yours to make.
Quantitative and qualitative data, ROI, KPI, Success Case Method, pilot testing. The designer who reports business impact pays for themselves. The one who reports completions is overhead.
Andragogical, learner-centered, instructor-led blended with WBT. Live facilitation, coaching and group practice remain the highest-value delivery—and the least automatable.
Thirty years of watching corporations deploy DreamWeaver, Lectora and Captivate to build an electronic page-turner. Overkill was always the tell. Now it is disqualifying.
Ninety-five percent gone, and no client will pay a team for it again. If digitizing existing content is your service, change your service this quarter. The hit is not the end of instructional design—it is the end of instructional design as a production department. What survives is the part that was always the job: find the gap, set the objective, prove the result.