How It Works
From a professor's life work to a student's midnight breakthrough
Four steps. No IT department required. The first student question can happen the same day.
See the steps ↓Fifteen years of teaching. Thirty minutes to upload.
Professor Kowalski has been teaching Corporate Finance for fifteen years. She has 340 lecture slides, a custom textbook chapter, twelve case studies, and a semester's worth of recorded Q&A sessions. She drags them into Grasperly's upload panel. The system reads everything -- PDFs, PowerPoints, even handwritten notes she photographed with her phone. Thirty minutes later, her Digital Twin understands net present value the way she teaches it. Not the way a textbook explains it. Not the way a generic chatbot guesses. The way Professor Kowalski has refined over a decade and a half of watching students struggle, adapt, and finally understand.
Average setup time: 30 minutes per course. Most professors are done before their coffee gets cold.
The technology is invisible. The teaching isn't.
Behind every answer is what we call the Precision Stack. When a student asks a question, Grasperly doesn't search the internet. It searches Professor Kowalski's materials using Retrieval-Augmented Generation -- a technique that ensures every response is grounded in actual course content. The system finds the most relevant passages from her lectures, synthesizes an explanation in her pedagogical style, and attaches citations so students can go deeper. Before any Twin goes live, the professor reviews sample interactions. She can adjust tone, add guardrails, flag topics that need special handling. The technology disappears. What remains is her teaching, available at scale.
Every answer can be traced back to a specific lecture, slide, or paper. No black boxes.
At 2am on a Tuesday, Maria asks a question about NPV.
Maria is in her dorm room. The exam is Thursday. She opens Moodle, finds the Grasperly chat widget embedded in her Corporate Finance course, and types: 'I don't understand why we discount future cash flows. If the money is coming anyway, why is it worth less now?' Within seconds, she gets a response -- not a Wikipedia summary, but an explanation drawn from Slide 47 of Professor Kowalski's Week 3 lecture, using the coffee shop analogy that Professor Kowalski developed specifically because students found traditional explanations confusing. Maria asks a follow-up. The Twin walks her through a worked example. Citations appear in the margin: 'Source: Lecture 3, Slide 47-49.' She clicks through, reviews the original slides, and something clicks.
90% of student questions happen outside office hours. The ones at 2am are often the most important.
By Friday, Professor Kowalski knows something she didn't know on Monday.
The dashboard shows her that 23 students asked about NPV discounting this week -- and 18 of them used the same confused phrasing about 'money coming anyway.' She realizes her Week 3 lecture might need a different opening example. She also sees that students who used the Twin before the midterm scored an average of 12 points higher. This is the flywheel. Students use the Twin, generating data about where they struggle. That data helps the professor improve her materials. Better materials make a better Twin. A better Twin drives more usage. The cycle compounds. Every semester, the teaching gets better.
More usage -> better data -> better materials -> better Twin -> more usage. The flywheel that makes AI in education actually work.
The questions professors and administrators actually ask
About 30 minutes per course. You upload your syllabus, lecture notes, slides, and any other materials. Our system processes them automatically. Most professors finish before their coffee gets cold. No IT department involvement required.
Yes. Grasperly integrates natively with Moodle and Canvas via LTI 1.3. Students access their Digital Twin directly inside the LMS they already use -- no new logins, no new apps, no friction. We also support standalone deployment if needed.
Grasperly uses Socratic guardrails by default. Instead of giving answers directly, it guides students through the reasoning process -- asking follow-up questions, pointing to relevant lecture material, and encouraging critical thinking. Every response is professor-authorized and grounded in course materials only.
All data is stored on EU servers. We are fully GDPR-compliant and provide Data Processing Agreements (DPAs) for every institution. Student conversations are encrypted, anonymized for analytics, and never used to train third-party models.
Every answer includes citations linking back to the specific lecture, slide, or paper it draws from. Professors can review responses, flag errors, and add corrections that immediately improve the Twin. The system learns from corrections without retraining.
Between EUR 8 and EUR 15 per student per year, depending on volume and features. Pilot programs start from EUR 10,000 for a full semester. We offer ROI guarantees for institutional deployments.