✦ If you're searching for a tool that recognizes real thinking — not just grades better answers — you're in the right place.
ThinkSmith is not another AI that does the work for your child. It observes how their thinking develops as they write, and shows it back to them in their own words.
Works on any device — add it to your home screen from your browser
Students paste a question. The AI produces an answer. The student copies it. The teacher marks it. Nobody learned anything. The Thinking Outcome is empty.
Students write their own answer first — before seeing anything else. There is nothing to copy from. The only way to get a strong Thinking Outcome is to actually think.
Where most AI tools score answers, ThinkSmith observes how thinking develops — and shows it back to the learner in their own words.
XAI (Explainable AI) is the idea that AI should make its reasoning visible, auditable, and understandable. Every observation ThinkSmith makes is grounded in evidence from the student’s own writing.
See thinking, not just answers.
Peer models think out loud, show their steps, and reveal how ideas form and evolve — so every learner can see what genuine reasoning looks like.
See transparent evidence, not black-box grading.
Every competency level is tied to observable features in the student’s own writing, informed by Ontario’s Achievement Chart structure.
See clarity, not mystery.
Feedback is understandable because the reasoning behind it is visible. Parents can follow why a level was observed — in plain language.
Is accountable.
ThinkSmith never hides its logic. Every decision is externalised — making it the opposite of the opaque “AI tutors” that replace thinking rather than build it.
ThinkSmith isn’t replacing teachers — it’s making thinking observable, coachable, and assessable in a way that aligns with real educational systems.
Whether a student is exploring a topic from curiosity or working through a course syllabus, the thinking process is always the same three Parts — followed by a Portfolio. Ms. Rivera, the AI teacher, is available throughout to review their work, answer questions, or simply help them find their own thinking.
The Thinking Outcome is not a score for getting the right answer. It is a recognition of how your child’s thinking is developing — built entirely from evidence in their own writing, with direct quotes.
Early levels reading “emerging” or “developing” are normal and expected. They mean the observation is honest.
What parents ask: “Why do all five levels move together?” The five competencies aren’t five separate tests — they’re five ways of looking at the same learning journey. Strong thinking tends to strengthen all five at once, which is why the levels move together. But the differences matter: a consistently strong Communication level with a lower Creative Thinking level tells you your child expresses ideas clearly and is ready to be challenged to take more intellectual risks. A strong Critical Thinking level with lower Knowledge & Understanding tells you the reasoning skills are strong but the conceptual foundation needs more depth. The gaps between competencies are where the most useful feedback lives.
ThinkSmith is built differently to the AI tools that have caused concern. Ms. Rivera is not a chat AI, a companion, or a counsellor. Every response she gives is constrained, structured, and tied to the learning activity — she cannot wander, improvise emotionally, or go beyond her role.
How ThinkSmith Handles Personal or Emotional Content
ThinkSmith is designed for one purpose — helping students learn to think. If a student types something personal, emotional, or unrelated to learning, ThinkSmith simply can't engage with it. She doesn't interpret feelings, give advice, or step into sensitive topics. Instead, she gently redirects the student back to the activity.
As students write, ask questions, and give peer feedback, their Thinking Outcome updates live — showing exactly where growth is happening, and where it isn't.
Every sentence observed. Every Ms. Rivera interaction considered. Every peer feedback analysed. The chart builds across all three Parts so you see the full arc — not just the final result.
As the student writes and revises, the chart shows how each competency is currently developing. A student can see the moment their thinking moves from emerging to developing well — after a conversation with Ms. Rivera, after considering peer feedback, or after their own revision.
The five competencies aren’t five separate tests — they’re five lenses on the same thinking. A strong Communication level with a lower Creative Thinking level tells you your child expresses ideas clearly and is ready to be challenged to think beyond expected frameworks. A strong Critical Thinking level with lower Knowledge & Understanding tells you the reasoning is strong but the conceptual foundation needs depth. The gaps are where the most useful feedback lives.
The live Thinking Outcome and the final portfolio report are the same recognition — one shown live during learning, one presented formally at the end. No surprises. No unexplained gaps.
"My daughter came away from the activity having actually changed her mind about something. Not because she was told the right answer — but because writing it down, reading other perspectives, and reflecting on feedback forced her to really examine what she thought."
"I was skeptical about an AI teacher. But Ms. Rivera never once gave my son an answer. She kept asking him questions until he figured it out himself. By the end he was explaining it to me."
When peers show their reasoning — not just their answers — something different happens for every level of learner. ThinkSmith is built on three of the most replicated findings in learning science.
Learning happens in the gap between what a student can do alone and what they can do with guidance. Visible peer reasoning is that guide — it shows the path, not just the destination.
When you evaluate someone else's reasoning, you understand the material more deeply than when you study it alone. Parts 2 and 3 of ThinkSmith are built entirely on this mechanism.
Generating explanations for why something is true — not just what — improves retention and transfer significantly. Strong learners need arguments to challenge, not answers to agree with.
Thinking Out Loud vs. conclusions-only peer work — predicted effect by learner level
ZPD = Zone of Proximal Development (Vygotsky) · Elaborative interrogation (King, 1992) · Protégé Effect (Chase et al.) · Preliminary simulation data consistent with theoretical predictions
Running average by learner level · preliminary simulation data
Part 1 scores are identical in both modes — the only variable is how peers present their thinking.
ThinkSmith is free to try. No credit card. No setup. Just a student, a challenge, and the space to think.
Demo available at thinksmith.ca/demo · Password: Partner2026