How Clew teaches
You learn by thinking, not by copying
Clew is built around one idea: the answer only sticks when you reach it yourself. Here is what that looks like in a session, and the research that shaped it.
Questions before answers
Clew asks the next small question
A good tutor rarely says the answer. They ask what you already know, find the exact step where you are stuck and give you just enough to take it yourself. Clew works the same way, with patience that never runs out.
If you are really stuck, Clew gives a hint, then a smaller one, or works through a similar example together. Then it hands your question back to you, so the last step is always yours.
Just tell me what 2/3 plus 1/4 is.
Let's get you there fast. Can you add thirds and quarters straight away, or do they need to be the same size first?
Same size? So twelfths?
Nice. How many twelfths is 2/3?
8. And 1/4 is 3. So 11/12.
That's it, and you did every step. Want one more like it, or shall I turn this into a flashcard?
A session, step by step
What happens when you learn with Clew
Clew asks what you're working on
Homework, a topic for a test, or a photo of a worksheet. It checks what you already know before it teaches anything.
It finds where you're stuck
Clew looks for the idea behind a mistake, not just the mistake, and goes back to an earlier idea if one is missing.
It asks, then waits
Clew asks one small question at a time and gives you time to think. Talking, typing or drawing all count.
Hints come in small steps
A nudge first, then a bigger hint, then a worked example of a similar question if you need it.
You finish the step yourself
The answer is yours. Clew asks you to explain how you got there, which is what makes it stick.
It comes back later
What you learned becomes flashcards and practice that return over the next days, just before you'd forget.
The learning science inside
What research says, built into every session
Active thinking
Asking, not telling
Students learn more when they explain, compare and build ideas than when they only listen or read.
Retrieval
Remembering beats rereading
Pulling an idea out of memory, as in a quiz or a flashcard, makes it last far longer than reading it again.
Spacing
A little, often
Practice spread over days beats the same practice crammed into one evening.
Wait time
Time to think
A few seconds of quiet after a question leads to longer, better answers, so Clew never rushes you.
Confidence
How sure are you?
Mistakes made with confidence are the ones most worth fixing, so Clew's quizzes ask how sure you are.
Cognitive load
One step at a time
Working memory is small. Clew breaks problems into steps so nothing important gets crowded out.
Worked examples
See one, then do one
When something is brand new, a clear worked example followed by a similar problem is the fastest way in.
Universal design
Many ways in
Ideas explained in more than one way, with more than one way to answer, open learning to more students.
Why not just a chatbot?
Answers on tap can quietly stop learning
In a large study in high school maths, students who practised with an unrestricted chatbot did better while they had it, then did worse than students who never had it once it was taken away. A version built to give hints instead of answers largely avoided that harm.
Well-designed tutoring is different. One-to-one tutoring has long been one of the most powerful ways to learn, step-by-step tutoring systems come close to human tutors, and a carefully designed AI tutor helped university physics students learn more in less time than an active classroom lesson. Clew is built on that second path.
Sources
The research we lean on
- Bastani and others (2025) Generative AI without guardrails can harm learning: Evidence from high school mathematics. Proceedings of the National Academy of Sciences 122(26), e2422633122. doi.org/10.1073/pnas.2422633122
- Kestin and others (2025) AI tutoring outperforms in-class active learning: an RCT introducing a novel research-based design in an authentic educational setting. Scientific Reports 15, 17458. doi.org/10.1038/s41598-025-97652-6
- VanLehn (2011) The relative effectiveness of human tutoring, intelligent tutoring systems, and other tutoring systems. Educational Psychologist 46(4), 197–221. doi.org/10.1080/00461520.2011.611369
- Bloom (1984) The 2 sigma problem: The search for methods of group instruction as effective as one-to-one tutoring. Educational Researcher 13(6), 4–16. doi.org/10.3102/0013189X013006004
- Chi and Wylie (2014) The ICAP framework: Linking cognitive engagement to active learning outcomes. Educational Psychologist 49(4), 219–243. doi.org/10.1080/00461520.2014.965823
- Roediger and Karpicke (2006) Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science 17(3), 249–255. doi.org/10.1111/j.1467-9280.2006.01693.x
- Cepeda and others (2006) Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin 132(3), 354–380. doi.org/10.1037/0033-2909.132.3.354
- Dunlosky and others (2013) Improving students' learning with effective learning techniques. Psychological Science in the Public Interest 14(1), 4–58. doi.org/10.1177/1529100612453266
- Rowe (1986) Wait time: Slowing down may be a way of speeding up!. Journal of Teacher Education 37(1), 43–50. doi.org/10.1177/002248718603700110
- Butterfield and Metcalfe (2001) Errors committed with high confidence are hypercorrected. Journal of Experimental Psychology: Learning, Memory, and Cognition 27(6), 1491–1494. doi.org/10.1037/0278-7393.27.6.1491
- Sweller (1988) Cognitive load during problem solving: Effects on learning. Cognitive Science 12(2), 257–285. doi.org/10.1207/s15516709cog1202_4
- Sweller and Cooper (1985) The use of worked examples as a substitute for problem solving in learning algebra. Cognition and Instruction 2(1), 59–89. doi.org/10.1207/s1532690xci0201_3
- CAST (2024) CAST Universal Design for Learning Guidelines version 3.0. CAST. udlguidelines.cast.org
Try learning the Clew way
Join early access and be among the first to learn with a tutor that asks before it tells.
Join early access