Super-Cyborg ID: How I Broke Washington’s Child Advocacy Rules into Bite-Size, Visual, Ready-to-Use Training

I saw a posting on LinkedIn for an “AI-powered Instructional Designer” and couldn’t help smiling—because this is already my day-to-day. Here’s a quick look at how I used Napkin AI to atomize Washington State’s guardian ad litem duties (RCW 13.34.105 and related rules) into a library of snackable visuals that double as a reference kit and the first sprint of an online module.

The problem to solve

Advocates finish training and ask, “Okay… what do I do now?” The big manuals and webinars teach what and why; they rarely deliver the exact small steps you need at 8:15 a.m. before a home visit or five minutes before a hearing.

The super-efficient workflow (Napkin AI at the core)

1. Source of truth: Load the relevant RCW sections + program policies into my notes (and a private, searchable repo).
2. Atomic tasks: Chunk the law into tiny “do/decide/say” actions (e.g., “Meet child privately,” “Verify tribal status,” “Flag non-compliance”).
3. Scenario anchors: Tie each action to where it lives in the real workflow (intake → first 30 days → ongoing monitoring → hearing day → closure).
4. Napkin AI drafting: For each action, I prompt Napkin AI to produce one visual card that blends: plain-language step, legal citation, risks, and “what to say” sample.
5. Rapid variation: I give Napkin AI text from the large document and it has me chose from between almost 20 style variations (diagram, decision path, timeline, checklist, script, etc.).
6. Down-select: Score variations for clarity, speed, and printability; keep 1–2 best per action.
7. Human guardrails: Quick SME check against the law and local practice; adjust wording for tone and trauma-informed language.
8. Publish fast: Export as PNG/SVG or PDF to screenshot later for a reference kit and drop the same cards into a Rise/Storyline or LMS micro-module with H5P, whatever the advocates can manage and access!

A sample Napkin AI prompt I use

“Turn this duty into a single, skimmable visual card for volunteers. Show: 1) When to do it, 2) Exact steps, 3) What to document, 4) Risk if skipped, 5) 1-line script I can say. Keep it at a 9th-grade reading level. Offer 20 visual variations (flow, decision tree, checklist, timeline, side-by-side ‘Do/Say’).”

What the output looks like (one card, many uses)

• Front: “Meet with the child—privately” → 3 steps, 1 line from RCW, a tiny decision branch (“If safety concern → coordinate with supervisor”), and a one-sentence script.
• Back (or tooltip): Documentation bullets + cross-references (“See: first-30-days checklist; hearing prep kit”).
• Formats: PNG for printing, SVG for slides, and an accessible HTML block for the LMS.

The 20 “flavors” I can spin up instantly

1. One-card checklist
2. Do/Say micro-script
3. 30-day timeline
4. Decision tree
5. Red-flag callouts
6. Court-day run-sheet
7. First-visit route map
8. Evidence log mini-table
9. HIPAA/Confidentiality quick rules
10. Report paragraph template
11. Outcome matrix (A/B/C)
12. Escalation ladder
13. Before/After example
14. Myths vs Facts
15. Quality checklist (QA)
16. Tribal status flow
17. Email snippet bank
18. Phone script
19. “If/Then” cheat card
20. Self-check reflection prompt

Why this “micro-first” approach works

• Cognitive load: Each card answers a single job-to-be-done.
• Transfer to performance: The same asset prints to a clipboard, lives in a mobile LMS, and slots into a chatbot for just-in-time help.
• Speed to value: From statute → usable visual in minutes, not weeks. (this project would have taken a week using my go-to Canva or Powerpoint! (can you say…Gamma AI??)

Is this the future—IDs as super cyborgs?

Kind of, yes—and it’s a good thing (with guardrails). AI collapses the distance from “source of truth” to “ready-to-use job aid.” The human ID still does the work machines can’t: define the workflow, choose the moment of need, ensure tone and equity, and verify compliance. Think power tools, not autopilot.

Guardrails that matter (especially in child advocacy)

• Ground every card in the exact statute/policy; no free-floating “advice.”
• Protect confidentiality (align to local interpretations of RCW 13.50.100).
• SME validation on anything legal or clinical.
• Accessibility & language equity: Alt text, plain language, and translations.
• Analytics with empathy: Use feedback to reduce friction, not to police volunteers.

What this enables next

• A reference kit you can hand to a brand-new advocate on day one.
• A modular course built from the same cards—each becomes a 3–5 minute micro-lesson with a quick scenario check.
• A chat assistant that answers “What do I do next?” by pulling the exact card + script.
• Ongoing updates when laws change—regenerate visuals, keep the links stable.

Check out the examples below I was able to create!