How to Run AI Professional Development for Teachers
Running AI professional development well means sequencing it like a skill-building program, not a single announcement: assess where staff actually stand, pick a delivery format that matches the time you have, build in spaced follow-up coaching, and check afterward whether classroom practice actually changed. A one-hour show-and-tell rarely moves the needle on its own.
Quick Answer: To run AI PD that actually changes practice, start with a short readiness check instead of guessing, choose a delivery format sized to your real time budget (a single workshop, a train-the-trainer model, or embedded coaching), sequence at least one follow-up touchpoint after the first session, and set a specific date to check whether teachers are using what they learned.
A 2024 Gallup and Walton Family Foundation survey of K-12 teachers found that a majority had already experimented with AI tools on their own, frequently without any formal training from their school. That gap between informal use and structured support is exactly what running AI PD well is meant to close.
Coordinators who treat "AI PD" as one item on a checklist tend to book a single afternoon session and move on. The staff who most need ongoing support — the ones still unsure whether a tool is safe to use with student data — are usually the same staff who stop attending after that one session ends.
What "Running" AI PD Actually Requires
Running AI PD is different from planning its budget or picking its topic — those decisions happen earlier. Running it means designing the actual sequence of sessions, choosing who facilitates each one, and building in the follow-up that determines whether any of it sticks.
Treat it as a short program with a beginning, middle, and end, rather than a single event that ends when the room empties.
Awareness Sessions and Fluency Sessions Are Not the Same Thing
An awareness session introduces what a tool does and doesn't do; a fluency session builds the muscle memory to use it inside a real lesson plan. Booking only the first and expecting the second is the single most common design mistake in AI PD.
- Awareness comes first for any staff group that hasn't seen a generative AI tool used inside an actual classroom workflow yet.
- Fluency requires repetition — one exposure to a new interface rarely survives the following Monday's schedule.
- Skipping straight to fluency for a group with no context usually produces confusion rather than confidence.
Deciding Who Should Lead the Sessions
The person leading a session shapes how much trust the room extends to the content, and different leaders suit different session types. Match the leader to the goal rather than defaulting to whoever is available that afternoon.
- An internal instructional coach or lead teacher — best for ongoing, embedded support once staff have basic context, since they're already a familiar, trusted presence in the building.
- An outside facilitator or vendor trainer — useful for a focused kickoff session, especially one tied to a specific tool, though the relationship usually ends when the contract does.
- A peer "early adopter" teacher — often the most credible voice for skeptical colleagues, since the examples come from the same building, the same schedule, and the same real constraints.
None of these three has to work alone. A common, effective pattern pairs an outside facilitator for the kickoff session with an internal coach who carries the follow-up coaching forward for the rest of the semester.
Addressing the Two Concerns That Come Up in Every Room
Almost every AI PD session surfaces the same two worries, and naming them upfront saves time later. Staff want to know whether a tool is safe to use with student information, and whether using it crosses a line into academic dishonesty.
- Data privacy falls under the Family Educational Rights and Privacy Act (FERPA), and for students under 13, the Children's Online Privacy Protection Act (COPPA) also applies — staff should know their district's vetted-tool list before a session, not learn about it mid-session.
- Academic integrity questions usually come from teachers worried about students misusing the same tools staff are being trained on; addressing this directly, rather than avoiding it, tends to build more trust than skipping past it.
- A short, honest answer works better than a polished one. Telling a room "here's exactly what our district's data policy allows" is more useful than a vague reassurance that everything is fine.
Sessions that name these concerns in the first five minutes generally see less resistance during hands-on practice than sessions that let them sit unaddressed in the back of the room.
Find Out Where Your Staff Actually Stands First
Skipping a readiness check is why so many AI PD sessions feel either too basic or too advanced for half the room. A five-minute survey before scheduling anything tells you which format actually fits your staff, rather than the format that was easiest to book.
RAND Corporation's survey work on classroom AI adoption has found usage remains uneven across a typical staff, often driven more by an individual teacher's own initiative than by any formal rollout. A readiness check is how a coordinator gets ahead of that unevenness instead of discovering it mid-session.
A Simple Three-Tier Readiness Split
Most staff groups sort cleanly into three rough tiers once you ask a few direct questions about actual use, not just familiarity with the term "AI."
- Cautious — has not used a generative AI tool for work, and has real concerns about data privacy or academic integrity.
- Curious — has tried a free chatbot once or twice, informally, without a clear workflow for repeated use.
- Confident — already uses a tool regularly for at least one recurring task, like drafting worksheets or writing comments.
Building a Short Diagnostic Before You Schedule Anything
A diagnostic doesn't need to be elaborate to be useful — five questions, sent through whatever survey tool your school already uses, is enough to sort a staff into the tiers above.
- Have you used any AI tool for lesson prep, grading support, or communication this year?
- If yes, which specific task did you use it for most recently?
- What's your biggest concern about using AI tools with student work?
- Would you rather learn in a whole-group session or a smaller, hands-on group?
- Is there a specific task you'd want an AI tool to help with, if you could pick one?
Results almost always reveal more spread than a coordinator expects — which is the exact reason a single one-size session tends to underserve at least a third of the room.
Choosing a Session Format That Fits Your Timeline
The format matters as much as the content, since a whole-staff workshop and a train-the-trainer rollout can cover identical material with very different time and follow-up demands. Pick the format before locking in a date on the calendar.
| Format | Time Commitment | Best Fit | Follow-Up Needed |
|---|---|---|---|
| Single whole-staff workshop | One session, 60–90 min | Building-wide awareness, kickoff | Yes — without it, adoption fades fast |
| Train-the-trainer | One session + ongoing peer support | Scaling after a small pilot group | Built in by design |
| Small hands-on cohort | 3–4 short sessions over weeks | Staff ready for fluency, not just awareness | Naturally spaced already |
| Self-paced modules | Flexible, staff-led timing | Staff who need flexible scheduling | A live check-in still helps |
| Embedded coaching | Ongoing across a semester | Sustained practice change | Is the follow-up |
Why One Workshop Rarely Changes Practice on Its Own
TNTP's widely cited research on teacher development has long found that isolated PD events, without follow-up support, produce little measurable change in day-to-day classroom practice. AI-specific PD isn't exempt from that pattern just because the topic is new.
A single well-run session can still build genuine excitement and clear up real misconceptions. What it rarely does alone is carry a hesitant teacher through the friction of trying something new for the first time, unsupervised, on a busy Tuesday.
Spacing Out Follow-Up So It Actually Sticks
Learning Forward's Standards for Professional Learning make a similar case from a different angle: learning that's sustained and embedded in the actual job tends to shift practice more reliably than a single session ever can.
- Schedule a short check-in two to three weeks after the first session, not just at the end of the year.
- Keep the follow-up small — fifteen minutes at an existing staff meeting works better than trying to book another full session.
- Ask what staff actually tried, not just whether they liked the training, since those two answers are often different.
Who Delivers It: Platforms and Support Options
Every delivery format above still needs a source of hands-on material and support behind it, and the options range from fully outsourced to entirely in-house.
| Option | Cost Pattern | Strongest For |
|---|---|---|
| Outside consultant or vendor trainer | Flat fee per session | A focused, one-time kickoff |
| Conference-based learning (e.g., ISTE) | Registration + travel | Deep learning for a small leadership team |
| In-house train-the-trainer | Stipend for a few lead staff | Scaling across a large or multi-building district |
| Vendor-provided onboarding | Often included with a subscription | Reducing the paid-facilitator time a rollout needs |
Vendor-provided onboarding is easy to underuse. Many AI content platforms, including EduGenius, ship with a guided walkthrough and documentation built around a single setup step — in EduGenius's case, a class profile that captures grade level, subjects, and ability range once, after which the platform can generate differentiated worksheets, quizzes, and slides without a separate learning curve for each format.
That kind of design choice is itself worth asking any vendor about before booking a paid trainer: how many distinct workflows does a teacher need to learn to use this tool competently? A platform built around one reusable setup step generally needs less paid facilitation time than one that requires separate training for every content type.
What a Full Semester Sequence Looks Like
Seeing the pieces above laid out on a calendar makes the sequence concrete. A readiness check, a chosen format, and a follow-up plan mean little until they're attached to actual dates a staff can see coming.
| Timeframe | Focus | What Happens |
|---|---|---|
| Weeks 1–2 | Readiness | Send the five-question diagnostic; sort responses into the three tiers |
| Weeks 3–4 | Kickoff | Run a whole-staff awareness session sized to the majority tier |
| Months 2–3 | Practice | Small hands-on cohorts split by tier; first follow-up check-in at week 3 |
| Month 3–4 | Scale | Train-the-trainer stipends assigned; peer observation windows open |
| End of semester | Review | Informal check on adoption signals; decide next semester's depth |
Say a 40-teacher elementary school runs this sequence starting in September. By November, the school isn't asking "did the PD happen" — it's asking which two or three specific tasks staff are actually repeating on their own, which is a much more useful question to bring into a spring planning meeting.
- A smaller school (under 20 staff) can often compress this into two months instead of four, since there's less scheduling coordination across departments.
- A multi-building district usually needs the full timeline, plus a short kickoff call between building leaders before week 1 to keep the sequence aligned across sites.
Running the Live Session: A Step-by-Step Structure
A well-run 60–90 minute session follows a predictable shape, whether it's the first session of the year or the third in a cohort series.
Before the Session
- Send the diagnostic results back to attendees as a short heads-up, so the session doesn't feel like it's starting from zero.
- Pick one or two concrete tasks to build around — a worksheet, a set of comments, a rubric — rather than a broad tour of features.
- Prepare a printed or shared one-page reference staff can keep open on a second screen during hands-on time.
During the Session
- Open with the "why" in under five minutes. A long framing section eats into practice time that matters more.
- Model the task live, narrating each decision out loud rather than just clicking through it silently.
- Give at least half the session to hands-on practice on the attendees' own real materials, not a generic sample.
In the Two Weeks After
- Send a short, low-pressure reminder with one more example task, timed to land before the follow-up check-in.
Sessions that reserve real hands-on time on the attendees' own materials consistently generate more usable output — and more staff confidence — than sessions built mostly around watching a demonstration.
Signs the Training Actually Changed Practice
A satisfaction survey at the end of a session mostly measures how the room felt, not what changed afterward. A few concrete, observable signals are more useful for deciding whether to keep funding a second semester of the same approach.
- Staff voluntarily bring examples to a staff meeting without being asked to share, which signals the tool became part of an actual workflow rather than a one-time exercise.
- Questions in the follow-up check-in get more specific over time — moving from "how do I even start" toward "how do I get a better result on this particular task."
- The same small group of early adopters isn't the only group still using it three months in; broader, quieter adoption matters more than a few enthusiastic outliers.
- Fewer support tickets or hallway questions about the basics, which suggests the awareness-level friction has genuinely worn off.
None of these signals require a formal research instrument to collect. A coordinator who simply asks a few specific questions at a staff meeting three months out usually learns more than a same-day exit survey ever revealed.
Mistakes That Quietly Sink an AI PD Rollout
Even a well-intentioned rollout can stall for reasons that have nothing to do with the tool itself.
- Skipping the readiness check and assuming every teacher is starting from the same place. A cautious teacher sitting through a fluency-level session usually disengages within minutes.
- Treating one workshop as "complete" PD. Without a follow-up touchpoint, most staff drift back to old habits within a few weeks.
- Letting the session become a feature tour instead of a task-based practice block. Staff remember what they did, not what they watched.
- Ignoring the staff who don't show up. Silence often signals unresolved concern about data privacy or workload, not disinterest.
- Never circling back to check whether anything actually changed. A short, informal check three to six months later is usually enough to know if the investment is working.
Key Takeaways
- A short readiness check before scheduling anything reveals whether your staff needs awareness, fluency, or a mix — and prevents a one-size session from underserving part of the room.
- Awareness-level and fluency-level sessions serve different purposes; booking only the first and expecting fluency is the most common design mistake.
- The delivery format — single workshop, train-the-trainer, small cohort, or embedded coaching — should be chosen based on available time, not convenience.
- Research from TNTP and Learning Forward both point to the same conclusion: sustained, job-embedded support changes practice more reliably than a single isolated session.
- Vendor-provided onboarding, including a platform's built-in walkthroughs, can reduce how much paid facilitator time a rollout actually needs.
- A well-run live session reserves at least half its time for hands-on practice on real materials, not a feature demonstration.
- Checking back three to six months later is what separates PD that changed practice from PD that was simply attended.
Frequently Asked Questions
How long should an AI PD session for teachers be?
Sixty to ninety minutes works well for most single sessions — long enough to model a task and give real hands-on practice time, short enough to fit inside a normal staff-meeting slot without needing a special schedule change.
Do we need an outside consultant to run AI PD, or can we do it in-house?
Either can work well. An outside facilitator often suits a one-time kickoff session, while an in-house train-the-trainer model tends to scale better across a full year and multiple buildings, since it builds capacity that outlasts any single vendor relationship.
How do we know if AI PD is actually working?
Ask what staff have tried since the session, not just whether they enjoyed it. A short, informal check three to six months out — even a simple show-of-hands at a staff meeting — usually reveals more than a same-day satisfaction survey ever does.
What's the biggest reason AI PD fails to change classroom practice?
Stopping after a single session. Research on professional learning generally, including Learning Forward's standards, consistently shows that sustained, job-embedded support outperforms one-off events — a pattern that applies directly to AI training.
Should every teacher attend the same AI PD session?
Not necessarily. A readiness check often reveals a split between staff who need basic awareness and staff who are ready for hands-on fluency work; splitting those groups, even informally, usually serves both better than one combined session.
Can AI PD be optional for staff?
Making the first awareness session optional often works better than mandating it, since willing early adopters tend to become credible peer voices for the colleagues who join later. By the fluency stage, most schools do expect participation, but by then interest has usually caught up on its own.
Running AI PD connects closely to the wider planning and budgeting questions around it. For the individual habits this program is meant to build, see Simple Ways Teachers Can Start Using AI; for the day-to-day routine those habits feed into, see A Teacher's Workflow for Integrating AI Into Planning; and for training focused specifically on assessment design, see How to Train Teachers to Use AI for Designing Assessments.
Related Reading
References
- Gallup and Walton Family Foundation — survey research on K-12 teacher AI tool use, 2024.
- RAND Corporation — survey research on classroom AI adoption patterns.
- TNTP — research on the effectiveness of teacher professional development.
- Learning Forward — Standards for Professional Learning.
- ISTE — professional learning and conference programming for education technology.