Most ed-tech rollouts fail the same way: a long pilot, a polished training day, a champion who leaves, and silence by November. This playbook is designed against that pattern. It is built for a department head with eight to twenty teachers and a 30-day window.
Week 1: pick the wedge
Do not roll out the whole product. Pick the single feature that solves the most acute pain in your department right now. For most ELA and Humanities teams, that is Grading. For Science, it is Quiz Maker or Lab Report grading. For interdisciplinary teams, it is AI Assistants for parent communication.
- Department head identifies the wedge feature
- Two volunteer teachers commit to using it for two weeks
- Success metric defined: hours saved per week, or feedback turnaround time
Week 2: the volunteers run it for real
No training day. The two volunteers use the wedge feature on their real grading or planning workload. The department head checks in twice that week, fifteen minutes each, to surface friction. The goal is not to validate the tool; it is to find the three things that are confusing so you can warn the rest of the department.
Week 3: department-wide kickoff
30 minutes at the regular department meeting. The two volunteers demo, the department head shares the metric, and every teacher commits to one use this week. That is it. No certification, no badges, no follow-up homework.
Week 4: keep the metric visible
Department head posts the running metric weekly: hours saved, feedback turnaround, whatever you chose. This is the part most rollouts skip. The tool is only worth keeping if the metric stays good.
Common objections and how to answer them
"What about student data privacy?"
Good question, raise it before parents do. TeacherAI's privacy posture is documented on the privacy page; the short version is that student work is not used to train models. Bring the page to the meeting.
"This is going to replace teachers."
It is going to replace the part of teaching that nobody actually wanted: the unread comments, the same-day-as-the-test grading, the Sunday-evening lesson assembly. The teaching part is harder to replace because it lives in judgement.
"Students will use it to cheat."
Students already are. The right response is to redesign assignments (see the AI-written homework field guide in this library) and to use process-visible writing tools, not to wait it out.
What to track to know it is working
- Hours teachers report spending on grading per week, before and after
- Average time from assignment due to feedback returned
- Number of parent communications sent per teacher per month
- Teacher-reported sense of Sunday-evening workload
What not to do
- Do not require everyone to use every feature
- Do not run a six-month pilot before deciding
- Do not centralize all account management in one person who might leave
- Do not measure success by login counts; measure it by hours saved