SchoolAI is an education-focused AI platform that lets teachers design guided conversations and activities for students, observe their progress, and intervene when it matters. Its central classroom feature, Spaces, can support explanations, inquiry, practice and formative assessment. The important distinction is not that students can chat with AI. It is that the teacher sets the learning objective, shapes the prompts, watches the work and checks whether students can still reason without the tool.
For a US middle-school teacher, the practical starting point is one short, supervised activity—not a promise to transform an entire course. This guide explains what SchoolAI does, how to design a genuinely guided lesson, what to measure, and when to choose a different approach.
What is SchoolAI, and how does it work in a classroom?
SchoolAI is designed around teacher-created learning experiences. In a Space, students interact with an AI assistant called Dot under instructions and constraints chosen by their teacher. The company describes use cases including tutoring, discussions, simulations, bellringers and exit tickets. A teacher can browse premade activities in Discover or create an activity from a learning goal, then preview it before sharing. See SchoolAI's current Spaces guidance for the actual workflow.
During a Space, Mission Control gives the teacher visibility into student activity and generated learning signals. These signals can help teachers decide whom to question, reteach or challenge. They should not be mistaken for a verified diagnosis, a final grade or a direct reading of a child's emotions. The teacher still needs to look at student work and ask clarifying questions.
SchoolAI also has Sidekick, a more open-ended AI learning assistant. According to the SchoolAI help center, a teacher-designed Space is typically better for a specific lesson with a known outcome, while Sidekick supports broader student questions. Feature access varies by school plan. The current pricing matrix lists student portals and broader Sidekick access under its Scale offering, so a teacher should not assume every account has identical capabilities.
The key difference from a general chatbot
A general chatbot can generate explanations and answers. A well-designed SchoolAI Space organizes those interactions around a teacher's plan: a defined standard, a short sequence of questions, a deliberate level of support and a final demonstration of learning. It also provides a view of what students are doing while they work.
That structure is useful, but structure alone does not guarantee learning. Students may still copy fluent answers, accept errors, or appear successful inside a guided task without understanding the subject afterward. The educational value depends heavily on the quality of the activity and the teacher's follow-up.
Begin with a learning problem, not with a technology feature
Before opening SchoolAI, write a one-sentence answer to this question: What should my students be able to explain or do independently by the end of the lesson?
Compare two plans:
Weak goal: Students will use an AI chatbot to learn about ecosystems.
Strong goal: Students will explain how a reduction in one population may affect at least two other organisms in a food web, using evidence from a diagram.
The second goal is observable. It can be assessed on paper as well as on a screen. It also leads to better AI instructions: invite an initial hypothesis, ask for reasoning, provide a hint only when needed, and finish with a new example.
Choose a narrow task you already teach, such as comparing fractions, evaluating sources, revising a claim or interpreting a science diagram. Select one class or a small group for the first trial. Check district approval, device access and accommodations before introducing a new platform to students.
How to create a guided SchoolAI Space: six practical steps
1. Choose a curriculum-aligned outcome
Use your district's approved curriculum or applicable state standard. Define one success criterion in language students can understand. A seventh-grade science class might use: "I can predict one consequence of removing an organism from a food web and justify my prediction."
Do not ask AI to determine the curriculum on its own. The standard and the important scientific relationships should come from teacher-reviewed materials.
2. Design the questioning sequence
A productive experience asks students to think before it helps them. Organize your prompt around four moves: attempt, explain, receive a bounded hint, try again. Reserve an explicit explanation or a worked model for situations where instruction is required.
You can adapt this example when creating a Space:
You are supporting a Grade 7 science lesson on food webs. Start by asking the student what might happen if a population of frogs declines. Ask the student to explain the effects on one prey population and one predator population. Do not immediately provide the full answer. If the student is stuck, offer one relevant clue, then ask for a revised explanation. Use clear, age-appropriate language. If the student expresses a misconception, point to a relationship in the food web and ask them to test it. Finish with a different food web scenario that the student must answer without a hint. Never invent facts from a diagram you have not been given; ask the student to describe missing information.
This is an instructional design example, not a guarantee that an AI model will obey every instruction. Test it yourself.
3. Preview the student experience
SchoolAI's documentation specifically recommends using preview before sharing a Space. Try different student behaviors: an accurate answer, an incomplete explanation, an irrelevant question, and a request for the finished solution.
Check four things. Is the scientific content correct? Are hints small enough to preserve thinking? Does the activity respond respectfully to confusion? Can the student eventually demonstrate the outcome without AI assistance? Rewrite any prompt that encourages premature answer-giving.
4. Share only through approved classroom channels
Follow your school's access and consent requirements. Share the activity through the supported joining method or learning management system, rather than asking pupils to create unapproved personal accounts. Explain the rules: no personal stories, passwords, medical details, other students' information, or copied private assignments unless the school has authorized the use.
Students should know that AI-generated text can be wrong, that conversations may be visible to authorized educators, and that the teacher—not the chatbot—is responsible for classroom decisions.
5. Use Mission Control as a reason to talk to students
As students engage, look for patterns worth investigating. Three pupils making the same causal mistake might need a mini-lesson. One student repeatedly requesting the answer might need a simpler first step. Another who finished quickly might need a harder transfer problem.
Treat dashboard labels as prompts for human observation, not unquestionable conclusions. Ask, "Show me the part of the diagram that supports your idea," instead of assuming that a colored status indicator tells the whole story.
6. Close with independent evidence of learning
End the Space, then ask students to complete one fresh task without AI. The task should be comparable to the original objective but use different content. Collect the explanation and inspect the reasoning, not merely the final choice.
If learners cannot explain what they did without help, the guided interaction may have produced task completion rather than durable understanding. That is important information for the next lesson.
A ready-to-use 30-minute lesson for Grade 7 science
Here is a modest classroom sequence that can be adapted to science, history or language arts. The purpose is to test whether guided AI questioning helps students make their own explanations.
| Time | Teacher and student activity | Evidence to collect |
|---|---|---|
| 0–4 min | Show a simple food web; students write an individual prediction | Initial reasoning, including uncertainties |
| 4–7 min | Explain the Space objective, privacy rules and expected use of hints | Students can restate the task |
| 7–17 min | Students use a teacher-previewed Space to test and revise explanations | Questions, misconceptions and revisions |
| 17–22 min | Teacher discusses a common error and models one causal link | Students' verbal justifications |
| 22–27 min | Students solve a new food web problem without AI | Independent transfer response |
| 27–30 min | Exit ticket: "What changed in your thinking, and how did you check it?" | Reflection and next-step needs |
During the ten-minute Space portion, the teacher should circulate rather than simply project the dashboard. A student who types "more animals will die" may be guessing; a student who names a changed predator-prey link may be reasoning. Ask both pupils to justify their claims.
The independent task matters most. Without it, a lively conversation with the AI may be confused with mastery.
Five classroom uses where guided AI may actually help
Formative assessment before a new unit
Create a short Space that asks students to explain what they already know, rather than immediately teaching. Teachers can identify common starting points and plan small groups. Validate findings with a conventional prompt or discussion.
Reading comprehension beyond recall
Ask students to infer a character's motive or compare two claims, pointing to specific text evidence. The AI can request justification rather than only mark an answer correct. Keep the original text available and ask students to distinguish quotation from inference.
Mathematics explanations
For an error such as adding denominators when adding fractions, design a prompt that asks the student to compare visual models and articulate why a particular rule fails. The AI should not simply reveal the correct computation. After a short guided sequence, students solve a new problem independently.
Language support for multilingual learners
A Space can offer simplified vocabulary, translation or alternate explanations, features described in SchoolAI's Spaces help documentation. Verify translations and ensure the learner still encounters the essential disciplinary language. Adapt access, not the intellectual goal.
Student reflection and planning
A short end-of-lesson interaction can ask, "What was your hardest step? What strategy helped? What will you try tomorrow?" Teachers can use the responses to open a coaching conversation. These prompts should not become informal mental-health screening; sensitive disclosures require the school's established human response procedures.
What teachers should watch for: five failure modes
1. The AI supplies the answer before the learner thinks. Add an initial-attempt requirement, reduce the number of hints, and make the final response independent.
2. A confident error becomes part of the lesson. Check factual content and worked examples in preview. When there is uncertainty, use teacher-approved readings and diagrams instead of open-ended generation.
3. Students become dependent on prompts. Deliberately remove scaffolds. Ask pupils to outline a strategy on paper, compare alternatives and decide which tool they need, if any.
4. The dashboard replaces dialogue. Analytics can prioritize attention, but a short student explanation is often more revealing than an automated engagement score.
5. The tool increases inequality. Check accessibility, reading level, school devices, connectivity, language support and accommodations. Provide an equivalent nondigital route to the same objective when needed.
None of these problems is unique to SchoolAI. They reflect familiar instructional challenges in a new interface.
Student privacy and safeguarding: a checklist before your first lesson
SchoolAI states that it maintains FERPA, COPPA and SOC 2 controls, and publishes trust and compliance information. These are vendor claims and documentation to review; they do not remove a district's duty to complete its own procurement, data governance and legal checks.
Before introducing any AI tool to minors, confirm:
The district or school has authorized this product for the intended ages and purposes.
Required agreements, data processing terms and parent notices are in place.
Staff understand who can see conversations, alerts and student-level records.
Students are not asked to disclose sensitive information that is unnecessary to the learning task.
There is a human escalation route for safeguarding alerts or distress.
Teachers know how to correct an AI error and report inappropriate outputs.
Students with disabilities, multilingual learners and pupils with limited device access have equitable alternatives.
UNESCO's guidance on generative AI in education urges age-appropriate, human-centered implementation and protection of learners' data. Schools should use that principle alongside their own policies and applicable law, not treat a vendor safety page as the final word.
Can SchoolAI improve learning? Ask a narrower question
SchoolAI presents product research and impact claims on its website. Those claims deserve examination of their methods, populations and comparison groups. A vendor study showing productive conversations is not necessarily proof that the product causes better exam performance, longer retention or stronger transfer across schools.
For an everyday school decision, ask a narrower question: When the same teacher teaches the same objective, does a carefully designed Space help students produce better independent explanations than the usual activity?
Plan a five-day classroom pilot:
Day 1 — Baseline: Give a short, independent explanation task and score it with a simple rubric.
Day 2 — Teacher design: Build and test one guided Space against common misconceptions. Check accessibility and privacy approvals.
Day 3 — Guided activity: Run the Space once. Record time spent, teacher interventions and observed errors.
Day 4 — Independent challenge: Give a comparable task with no chatbot and no new hints.
Day 5 — Reflection: Compare the baseline and transfer work cautiously; ask students what helped, what confused them and what they could do next alone.
This is a small quality-improvement exercise, not a controlled scientific experiment. Different task difficulty, student familiarity, teacher effects and practice can all influence results. Avoid making causal claims from one short pilot.
An assessment rubric that measures reasoning
| Criterion | 0 — Not yet | 1 — Developing | 2 — Independent |
|---|---|---|---|
| Scientific accuracy | Incorrect relationship | Partly correct | Accurate relationship |
| Use of evidence | No evidence | General evidence | Specific diagram evidence |
| Explanation | Unsupported conclusion | Partial causal explanation | Clear causal chain |
| Transfer | Cannot start a new case | Needs a hint | Applies reasoning to a new case |
Read the actual student explanation before scoring. A completed AI conversation is not itself evidence of independent performance. Where meaningful, revisit a concept after a delay; the US Institute of Education Sciences practice guide supports approaches including spaced learning, explanatory questions and retrieval practice.
SchoolAI versus traditional teaching: it is not an either-or choice
Direct instruction remains useful when learners need an accurate model, essential vocabulary or an explicit demonstration. Peer discussion creates opportunities for disagreement, collaboration and spoken reasoning that an individual AI conversation cannot reproduce in the same way. Teacher questioning allows an educator to interpret gestures, history, classroom relationships and students' previous work.
A Space is a supplementary learning environment. Consider using it when individual practice, feedback timing or the need to surface multiple misconceptions presents a real difficulty. Use direct teaching, discussion, manipulatives, books, paper or an approved alternative when those approaches are better suited to the learning goal.
A technology purchase should never substitute for teacher development. The lesson still needs a coherent explanation, reasonable practice, feedback, classroom routines and meaningful assessment.
The Learning Coach perspective: technology should build autonomy
Learning Coach emphasizes a shift from simply delivering explanations to helping learners identify goals, recognize their own resources, reflect on strategies and act more independently. In the classroom, a Teacher Coach adopts this stance while retaining the full responsibilities of a teacher. That is different from an instructional coach who supports other educators, and different from a software assistant.
A coaching-oriented question changes the interaction. Instead of asking AI to "fix this answer," a student might be asked: "What did you try first? Where did your reasoning change? What evidence supports your conclusion? What will you do without a hint next time?"
For teachers designing Spaces, this translates into an instructional progression:
Observe: Get a genuine first attempt before offering a solution.
Question: Invite the learner to explain a choice or identify a contradiction.
Support: Provide the smallest useful explanation, model or clue.
Release: Remove the help and ask for a different example.
Reflect: Help the learner name a strategy that can be reused.
This is a way of applying learning coaching to AI-supported lessons, not a claim that SchoolAI itself is a certified coach or that its AI conversations have the same function as professional human coaching.
Teachers interested in developing these pedagogical skills can explore the Learning Coach training and specialization pathway. Program availability and admission conditions should be checked on the current training page. The English Learning Coach book collection, including Stop Teaching, Start Coaching, offers a related perspective on learning independence. These are Learning Coach educational resources, not independent evaluations of SchoolAI.
For a complementary look at AI-assisted teacher feedback, see Brisk Teaching: How AI Helps Teachers Save Time and Give Feedback. It examines a different workflow and the importance of students acting on the feedback they receive.
Frequently asked questions
Is SchoolAI free for teachers?
SchoolAI currently describes a limited Free Trial and paid Pro/Scale plans. Its pricing page lists five Space launches per year in the Free Trial, with different capabilities under higher tiers. Check current terms before planning a schoolwide rollout; pricing, limits and trial offers can change.
Does SchoolAI replace a teacher?
No. It can deliver guided interactions and surface signals for the teacher, but the educator remains responsible for planning, instruction, assessment, wellbeing and appropriate intervention. AI output can be inaccurate even when it sounds confident.
Can SchoolAI work for elementary students?
The company markets classroom experiences across grade bands. Age suitability, consent, supervision and school approval are still essential. Younger learners may need a shorter, more structured activity and more direct teacher participation.
What is the difference between Spaces and Sidekick?
A Space is built or selected around a teacher's specific activity and learning outcome. Sidekick is a broader assistant students can question across topics. Access to the student portal and Sidekick depends on an institution's plan and setup.
How can teachers keep students from copying AI answers?
Ask for an attempt before the AI conversation, require reasoning during the interaction, limit answer-revealing prompts and collect a new, independent task afterward. A conversation log alone cannot show whether a learner has acquired a transferable skill.
Does an AI engagement indicator prove understanding?
No. Such indicators may help a teacher notice who needs attention, but understanding should be assessed through student explanations, independent work and relevant follow-up questions.
Final takeaway
SchoolAI is most promising when it helps a teacher notice a misconception earlier and give a learner the right kind of support without taking over the task. Start small: one standard, one carefully previewed Space, one brief guided interaction and one independent follow-up. Measure whether students can explain their thinking after the screen is closed.
The point is not more AI conversations. It is students who need fewer prompts because they have learned how to reason, check and continue on their own.
Resources and references
SchoolAI — Getting Started with Spaces, product help documentation, consulted October 27, 2026.
SchoolAI — Sidekick or a Space?, product help documentation, consulted October 27, 2026.
SchoolAI — Pricing, current plan descriptions, consulted October 27, 2026.
UNESCO — Guidance for Generative AI in Education and Research, independent international policy guidance.
Institute of Education Sciences — Organizing Instruction and Study to Improve Student Learning, independent US evidence-based practice guide.
Product-specific features and plan limits are time-sensitive. SchoolAI is discussed independently; this article does not indicate an endorsement, partnership or controlled product trial.
