Doesn't templating conflict with individualization?

No — it funds it. The core scaffold (layout, category structure, key word positions) is where consistency helps every student; the fringe (their people, interests, settings) is where individualization lives. Template the skeleton so your hours go to the personal parts.

How much time does AI board drafting actually save?

It moves board creation from build-every-tile to review-and-adjust. You still apply clinical judgment to every board — vocabulary selection, access, organization — but you apply it to a draft instead of a blank grid, which is reliably faster.

Managing an AAC caseload without rebuilding every board from scratch

Somewhere around the tenth AAC student, the math breaks. Each student needs boards for their settings, their vocabulary, their year — and if every board is hand-built tile by tile, your caseload’s communication infrastructure runs on your evenings. The students keep coming; the evenings don’t scale.

The fix isn’t working faster. It’s refusing to build the same board twice.

Template the skeleton, personalize the skin

Across your caseload, most boards share a skeleton: a core word set in consistent positions, category structure (people / places / feelings / activities), a feelings row, yes/no anchored in corners. That skeleton should exist exactly once, as a template — because you designed it deliberately and because consistency across your caseload means you never re-orient either.

What differs per student is the skin: their people, their interests, their lunch menu, their classroom nouns. That’s the individualization that matters clinically — and it’s maybe twenty minutes of work on top of a solid template, versus three hours from a blank grid.

Practical template set worth owning: a home/core board, a mealtime board, a morning-routine board, a medical-visit board, a field-trip board, and a seasonal board you re-skin each term. Six templates cover a startling fraction of caseload requests.

Let AI do the first draft (you keep the judgment)

The newest lever: describing a board instead of assembling it. In SpeakAnyWay, “board for a 3rd grader’s science unit on weather, 12 cells, photo-realistic symbols” produces a draft in under a minute — and a photo of a worksheet, a menu, or an old paper board can seed a board the same way. Your role shifts to what actually requires your license: auditing vocabulary selection, fixing what the draft got wrong, matching access needs. Review-and-adjust beats build-from-nothing, every time, at caseload scale.

(Worth being clear-eyed: no AI knows your student. Drafts are scaffolding for your clinical judgment, not a substitute for it.)

Stop being the single point of failure

The other caseload killer is that every board change routes through you. Build the escape hatch deliberately:

Shared access, not emailed files. Boards that live where the team — para, teacher, family — can see the current version end the “which-PDF-is-newest” problem. Updates propagate; nobody’s laminated copy is three revisions stale.
Teach paras and parents to add fringe. They know the new obsession before you do. Let them add “excavator” on Tuesday; audit on Friday. A team that can extend boards between your sessions is a team whose student doesn’t wait a week for a word — and it’s the same muscle that makes family handoffs stick.
Keep a caseload change-log habit. One line per student per week (words added, words ignored, next need) turns board maintenance from archaeology into a checklist.

The honest arithmetic

None of this reduces the clinical thinking your caseload needs — that was never the bloat. The bloat was rebuilding skeletons, re-typing core sets, and re-laminating what a synced device would have updated silently. Template the skeleton, draft with AI, distribute the fringe work, and the hours you reclaim go where they belonged all along: with the students.

This is lived experience and research, not medical advice — your SLP knows your child.