Faculty

Introduction

A note to faculty

You have spent the last two years watching something subtle happen to the work your students submit. The work is often correct. It is sometimes more polished than what the same students produced a year ago. But the questions that arrive in office hours have flattened, and the moments of genuine confusion that used to lead to understanding have grown rarer.

Exams, papers, problem sets, and weekly assignments are giving noisier signals than they used to. Students can produce the right answer — or a fluent argument — without having internalized the reasoning that makes it theirs. This shows up most acutely in courses where this week's understanding is next month's prerequisite, and in any course whose real purpose is to build a capability the finished artifact only stands in for: constructing an argument, working a proof, weighing evidence, reasoning toward a defensible conclusion.

Classy AI was built in response to that drift. It sits at the intersection of two facts: Bloom's 1984 finding that one-to-one tutoring produces roughly two standard deviations of gain over conventional instruction, and the simple economic constraint that has kept that level of personalization out of reach until now. It is an attempt to deliver one-to-one tutoring without sacrificing the cognitive rigor that makes learning transfer to new problems.

The friction is the feature. Productive struggle is what converts information into transferable understanding.

01 · The product

What Classy AI is

An infinitely patient tutor, online whenever students choose to study — no scheduled tutoring hours.

Classy AI is a personalized learning platform that coaches students through their own reasoning rather than producing the reasoning for them. It is built on five pillars, each drawn from cognitive science and chosen because it has survived decades of replication.

Pillar	What it does
Socratic questioning	Guides students to construct ideas through dialogue, modeling the kind of one-to-one conversation a skilled TA or faculty mentor would lead in office hours.
Adaptive scaffolding	Calibrates its scaffolding to what the student demonstrates in conversation, rebuilding from where understanding broke down.
Adaptive quizzing	Students generate short quizzes from any section on demand; question difficulty adapts to measured quiz performance.
Recency-aware review	Review recommendations resurface each student's weakest and least-recently-practiced sections at proficiency-keyed intervals.
Bounded learning	Grounds course sessions in your course materials rather than the open internet.

What a session looks like

A student in your course opens Classy and types something simple: I don't understand why this works. Classy asks the student a question. The student answers. Classy asks another.

Within a few exchanges, it has located the actual point where understanding broke down. Perhaps a missing intuition from an earlier unit, or a brittle grasp of a concept the current material quietly rests on. It begins rebuilding from there.

After the dialogue, the student can generate a short quiz on the section they just worked through — question difficulty adapts to their measured performance, and the results feed their per-section progress. The next session begins where the last one ended.

02 · Why this matters

The capability your course builds is the thing at risk

General-purpose generative AI poses a particular problem for coursework, and the problem is fundamentally the same across disciplines even though it surfaces differently. A course does not exist to produce artifacts — problem sets, papers, lab reports, proofs, response essays. It exists to build the capability those artifacts stand in for: the reasoning, the analytical habits, the ability to construct and defend a conclusion. A student who outsources the artifact can submit correct, even polished work while building none of the underlying capability.

This is sharpest in courses that are cumulative — where mastery of week 4 is the working memory of week 9, and a student who outsources the derivation in October arrives in February without the schemas to even parse the new material. But it is not confined to those courses. A seminar paper assembled by a chatbot leaves the same gap as an unearned derivation: the student never did the thinking the course was designed to develop.

Students using general-purpose AI to complete coursework can produce correct artifacts while developing brittle, non-transferable understanding. The artifact looks like learning but it is far from it.

The cost of a shortcut is invisible at submission and obvious later — when the capability that should have been built isn't there.

Where Classy is architecturally different

Classy AI is built to decline to do the work for the student. That design choice is the center of the product — and it is a design choice, not a guarantee.

• Guidance, not answer delivery. Classy is instructed to work with the student on the structure of the problem, the relevant principles, and the next defensible step rather than produce a finished solution. This is how the tutor is built to behave, not an enforced block — which is why the corpus scoping below matters: instructors keep answer keys and solution material out of what they share.
• Misconception-first dialogue. When a student's question or claim rests on a misunderstanding, Classy is instructed to surface and localize it with a question before explaining, so the grounding an idea depends on is repaired rather than papered over.
• Quizzing as part of the loop. Students generate quizzes from course sections on demand, with question difficulty adapting to measured quiz performance.
• Recency-aware review. Review recommendations resurface weak concepts at proficiency-keyed intervals informed by the spacing literature.
• Bounded to your materials. Classy draws from the materials you have chosen for the course, limiting hallucinations and staying aligned with your lesson plan. The tutor speaks your course's notation and conventions, citing the section it is drawing from.

03 · Bounded learning

The tutor is bounded to your materials.

Being built to guide rather than give the answer is one half of what makes Classy AI different. The other half is what the tutor draws on when it does respond. A general-purpose LLM has been trained on most of the public internet, which gives it broad coverage but introduces two specific problems in a course context. It has no commitment to your course's notation or framing, and the completeness of what it generates cannot be guaranteed against any particular reference.

Bounded learning is the architectural response. The tutor is given the materials you have chosen for the course, retrieves the passages relevant to the student's question, and is prompted to draw from those passages first. The pre-trained knowledge of the underlying model remains available as a fallback when the corpus does not cover the question, but it is no longer the primary source. The approach is drawn from the SocratiQ research framework (Jabbour et al., 2025), which documents the design choice directly: "relying solely on the pre-trained knowledge of the language model introduces the risk of generating content that is inconsistent or not related to the course content curated."

Why this matters in your course

In a course, the difference between generally correct and course-correct is the entire point. A literature course built around close reading will scaffold differently than one built around historical context. A history course that frames a period through economic forces will scaffold differently than one that frames it through ideas. A statistics course that teaches inference through simulation will scaffold differently than one that teaches it through formulas. A bounded tutor speaks the language of your lecture; an unbounded tutor speaks the language of the average textbook on the internet.

How the binding works

Commitment	What it does
Corpus scoping	In a course session, retrieval is limited to the one document the session is opened against — a class-shared document or the student's own upload. Solutions handouts, the answer key to next week's assignment, instructor notes — if you do not share them, the tutor cannot retrieve them (it can still reason from the model's general knowledge).
In-context grounding	Retrieved passages are passed to the tutor as the primary source. The system prompt instructs the model that course materials take precedence, and that conflicts should be resolved in favor of the corpus.
Citation	When the tutor draws from a specific section, it names the section. Students can verify; instructors can audit.
Graceful fallback	When the curated corpus does not cover the question, the tutor falls back to general knowledge but signals that it is doing so.

Bounded learning is the architectural commitment that the tutor speaks the same language as your lecture.

This is the architectural answer to the question faculty often ask first: how do I know the tutor won't teach my students the wrong thing? The answer is that you choose what it teaches from.

One boundary worth stating plainly: bounded learning applies to document-grounded sessions — the sessions students open against your shared course materials. Students can also open a general study session without a document; those sessions draw on the model's general knowledge and are separate from course-grounded work.

04 · The educator dashboard

Defensible evidence of where understanding breaks down

The instrument many faculty have wanted for years.

Every Classy AI course comes with an educator dashboard built for instructors and institutions. It is a diagnostic instrument, engineered to feel like a tool a department chair could put in front of a curriculum committee. Four views, sequenced so that an instructor opening it for the first time lands on a view that orients them.

View	What it shows
Class overview	Stat cards for active members, average proficiency, total quizzes taken, active-this-week. Top five struggling concepts on the left, recent activity feed on the right.
Concepts	A single-column heatmap of every section across the course's shared documents, sorted by struggle score. Click any section to expand: every student who has engaged with it, sorted strugglers-first, alongside coarse Bloom's-taxonomy bands (lower-order vs. higher-order engagement), shown only when enough evidence exists.
Students	Per-student summary stats with click-through to per-section progress, quiz performance, and session activity. Used when exam performance has raised a question and you want a finer-grained read than the gradebook supports.
Settings	Class controls: rename, rotate join code, manage shared documents, manage roster, archive.

What gets measured, and the struggle score

Two signals drive the concept view: demonstrated proficiency per section (a decayed mean over recent quiz attempts, bounded to [0,1], with null for unmeasured rather than zero) and coarse Bloom's-taxonomy engagement bands (lower-order versus higher-order, reported only when enough evidence exists). The struggle score is computed directly from the first:

struggle_score = 1 − avg_proficiency

One question, answered plainly: how poorly is the class performing on this section, based on demonstrated quiz evidence. Sections with no evidence report insufficient data rather than a guessed score. A faculty member can look at any number on the dashboard and trace it back to the quiz attempts that produced it.

A percentage grade tells you a student got 73%. Bloom mastery bands tell you whether the missing 27% is lower-order recall or higher-order reasoning — reported only when enough evidence exists to say.

Why instructors and institutions care

• Lecture revision. Identify where the class is collectively stuck, by concept, before the next lecture. The struggle-sorted concept view answers one question fast: what should I cover differently on Tuesday?
• Targeted remediation. Two students with the same midterm grade can have very different gap profiles. The per-student view, with per-section progress and Bloom mastery bands, targets office-hour time where it will move the needle.
• Curriculum committee evidence. Concept-level engagement data across a cohort gives departments a defensible empirical basis for curriculum redesign.
• Insight, not surveillance. The dashboard is read-only from the instructor side. No keystroke logging, no time-on-page, no full chat transcripts surfaced to instructors — evidence drill-downs show at most a short, capped excerpt of the student's own work. Topic reports you download show what your class asked about by date, document and section. They don't include students' names, emails or IDs. An instrument for the instructor, not a behavior-modification system for the student.

05 · In your course

How Classy fits the work you are already doing

Classy AI works in concert with faculty. Lecture, seminar, lab, recitation, and office hours remain the center of gravity. Classy meets students in the moments between — at 11pm before an assignment is due, on the bus, the morning of a midterm — when a question would otherwise go unasked.

Concrete instructor workflows

Use case	How it works
Pre-class preparation	Assign a short Classy session before a flipped-classroom discussion or seminar. Students arrive with prerequisite gaps already surfaced and partially closed.
Supplemental assignment support	Students working through an assignment use Classy in place of an answer-key shortcut. The dialogue is structured to keep the reasoning load on the student.
Targeted remediation	Identify students whose exam performance suggests a specific gap; route them to a focused Classy sequence without pulling them from class or singling them out.
Independent study	Advanced students explore topics beyond the syllabus at their own pace, with the same Socratic discipline.

The research, briefly

Finding	Source
One-to-one tutoring produces ~2σ of gain over conventional instruction.	Bloom (1984)
Active retrieval, even unsuccessful, produces more durable learning than re-reading.	Roediger & Karpicke (2006)
Distributing review across time substantially increases long-term retention.	Cepeda et al. (2006)
Learning conditions that feel harder in the moment often produce stronger long-term learning.	Bjork (1994)
Bounded, in-context grounding keeps an AI tutor anchored to curated course material.	Jabbour et al. (2025), SocratiQ (arXiv:2502.00341)

Built on a foundation of proven research, Classy AI scales high-impact instruction to reach every student in the classroom.

Classy AI serves as a force multiplier for the instructional team. While the professor remains the center of gravity for the course, the AI acts as an expert mentor available whenever students choose to study — democratizing high-level academic support.

Invitation

There is a third path

If you have read this far, you are likely one of the faculty we built Classy AI for: someone who refuses to treat the current moment as a forced choice between adopting tools that hollow out learning and rejecting AI altogether. There is a third path. Classy AI is our attempt to walk it.

We are particularly interested in partnerships with instructors willing to pilot Classy AI in a course and contribute to its evolution. A pilot can start at the scale of a single section, with the option to expand.

---

Get in touch

Website	stayclassy.ai
Incubated	Harvard Innovation Labs
Founder	Kuljit Takhar
Email	kt@stayclassy.ai
For faculty	Pilot programs and course partnerships available. Our privacy policy and terms of service are published on this site.

Classy AI is on a mission to democratize the power of personalized instruction for every student, everywhere.