X-YYYYY where
X is the effort letter
(E = Extensive: 3+ dimensions attend;
M = Moderate: 1–2 attend;
L = Light: all skip), and each Y
position is A (Attend: mean ≤ 5.7, a gap exists) or
S (Skip: mean > 5.7, near ceiling).
Dimension order: AISE (AI self-efficacy), PU (perceived
usefulness), PEOU (perceived ease of use), PE (perceived enjoyment),
BI (behavioural intention).
Intervention Focus
Start with a single low-stakes task using familiar curriculum content. Build one visible success before adding complexity. Use paired activities so no teacher works alone. Address anxiety explicitly — normalise struggling with new tools. Sequence strictly: confidence first, usefulness second, commitment last.
Deprioritise
Theory, slides, policy rationale, and ethics modules. Do not ask teachers to share outputs publicly. Do not fast-track — nothing can be skipped with this cohort.
Intervention Focus
Hands-on access is the entire agenda. Map what devices and platforms are available. Walk through one complete AI-assisted task (e.g., a lesson plan draft) using the teacher's own subject content. Show explicitly how AI connects to tools they already use daily. Assign a low-pressure try-it-this-week task before leaving.
Deprioritise
Motivational content, benefit arguments, and case studies. This cohort is already convinced; every minute not spent on hands-on practice is wasted.
Intervention Focus
This is an identity and relevance problem, not a skills problem. Connect AI to what each teacher already cares about in their subject. Use peer storytelling — invite colleagues who do use AI to share one specific classroom example. Co-create one real artefact the teacher will use in their next lesson.
Deprioritise
Technical training and tool navigation. This cohort knows how to use AI; they need a reason to commit, not another demonstration.
Intervention Focus
Frame AI as an extension of tools teachers already use, not a replacement. Provide scaffolded prompt templates for immediate success. Celebrate imperfect outputs as normal. Pair teachers by subject so confidence is built in a familiar context. Low pressure and high success rate is the design principle.
Deprioritise
Abstract AI theory, ethics modules, and policy sessions. Avoid public sharing of outputs until confidence is established.
Intervention Focus
Commission this cohort as internal AI champions or peer mentors for lower-readiness colleagues. Offer advanced challenges — prompt engineering, AI output auditing, or cross-subject curriculum co-design. Involve them in designing the school's AI use policy.
Deprioritise
Any standard workshop content. Running a basic session for this group wastes their expertise — they are more valuable as trainers than as trainees.
Mixed rooms: if 50% or more of participants share the same effort letter, treat the whole room under that effort level and use the most common dimension pattern to set the session design. If effort letters are genuinely split, design for the highest effort code present — a session built for E-code participants can still serve M-code participants, but the reverse is unlikely to be true.
About TAM: the Technology Acceptance Model (Davis, 1989) and its TAM3 extension (Venkatesh & Bala, 2008) are among the most widely validated frameworks for understanding technology adoption. The five dimensions here capture attitudinal and motivational facets that are directly actionable in professional development. They are not a complete measure of AI literacy.