A job aid, not a course
A complaint handler at a medical device manufacturer has an investigated complaint in front of them and needs to know whether it is reportable, to whom, and by what date. This answers that in under a minute, for FDA and the EU at once, and hands back a record they can paste into the complaint file.
Try the tool first. Why it is a job aid rather than a course, how it was built, and the accessibility and deployment notes all follow underneath it.
Is this complaint reportable, and by when?
The tool runs in the frame below and scrolls inside it. Open it full screen instead →
Why it is a job aid
Courses build capability. Job aids support people who already have it.
Choosing correctly between the two is most of the work in performance consulting, and it is the decision clients are really paying for. Nobody needs to be taught reportability every time a complaint lands. They need the right answer at the moment they are deciding, under a clock, with a record of how they got there.
So this teaches nothing on first read. It asks the fewest questions that resolve the determination, then gets out of the way. The reasoning sits underneath, in a panel the user can open on the second or third pass, once they want to understand rather than decide.
A course covering the same material would look completely different. It would build the judgment this tool assumes: why the awareness threshold is organizational, what goes wrong when it is not, and practice on genuinely borderline cases with feedback on the reasoning rather than the answer.
Build information
How it was made
What it demonstrates
- Knowing when performance support beats training, and designing for the moment of decision
- Entering a regulated domain and getting the details right, which is the job an instructional designer does with a subject matter expert
- Two jurisdictions assessed in parallel, because the same facts can be reportable in one and not the other
- An honest stop where the decision needs a human, rather than a tool that pretends to answer everything
- Scope stated plainly, including what was deliberately left out
How it was built
- One self-contained HTML file, 94 KB, no frameworks and no network calls
- A conditional question set that skips what it does not need, so most assessments take four questions rather than eleven
- Date arithmetic that distinguishes calendar days from working days, because the two regimes use different units
- Poppins, Lato and Noto Sans Mono subset to the characters actually used and embedded as WOFF2, 40 KB for all seven faces
- The printable reference is generated programmatically, so a regulation change is a text edit rather than a redraw
Accessibility
- Every text pair meets WCAG 2.1 AA, most reach AAA
- Colour carries deadline severity, but never alone: each state also has a word and a shape
- Fully keyboard operable, with focus moved to each new question and returned on close
- Changes announced to screen readers, and external links flagged as opening in a new tab
Deployment
- Drops into Storyline as a web object, or into any LMS as a standalone page
- Sets course variables on completion where a player exists, and fails silently where one does not
- Works offline, which matters because an embedded tool cannot assume a connection
- Stores and transmits nothing, which is the first question anyone in this field asks
Built from publicly available regulation: 21 CFR Part 803, EU MDR Articles 87 and 88, and the MDCG vigilance guidance. It is a demonstration piece and not regulatory advice. The scope note on the tool says so, because people who work in regulated environments expect to see scope stated and get suspicious when it is not.
Want something like this for your team?
Tell me what people need to be able to do differently. I will tell you what it takes.