Everything students, professors, and administrators ask us most.
A tool that records how a human works with an AI model and produces verified provenance of their contribution: what they asked and how their thinking developed. The result is a Contribution Score and a timeline of the conversation for an instructor to review.
No. We don't analyze finished text to guess whether AI wrote it. We document the working process, and we say nothing about work done outside it.
A 0-10 number summarizing the human contribution observed in a conversation. It is a record of observed contribution, not a grade. The same number can mean different things on different assignments, which is why instructors see the conversation timeline alongside it.
The score is interaction-based and weighted, and it's tied to specific messages in the conversation, so an instructor can see why a score is what it is. We don't publish the weights during the beta, because testers' conversations are how we check that the score is doing its job.
No. It means the recorded conversation showed less observed contribution. That could be a quick task, a different working style, or work done elsewhere. It's a reason to read the conversation, not a verdict.
Yes, but only when you send it. The instructor you're matched with sees the dashboard you send them, including your conversation and timeline.
Not during the beta. Your instructor sees it.
No. The AI is there to help you work through problems. Nothing is graded automatically. Your instructor may review the conversation, and we measure your interactions so that review is possible.
It's built to ask questions and point you back to your assigned materials. The answers have to be yours as you learn the material.
Authentifi only documents what happens inside it, and it's evidence for a conversation, not a lock. Instructors can still ask you to explain your work in person.
Your conversation and its timeline. You can click any point to jump to that part of the chat.
We built it not to be a grade, and there is no auto-grading. We recommend against using the number as one. There are no fixed bands like “investigate below X”; we'll consider guidance once there is enough pilot data.
The score, plus a timeline of the conversation. Click any point on the timeline to jump to that part of the chat, or search the chat by word.
Not in this beta. We're using two pre-loaded assignments: one computer science and one business.
Yes. Data handling is designed around FERPA from the start. We don't claim certification or compliance status.
The conversation, timestamps, the score and what it's based on, and the course and instructor it belongs to.
Each submission is hashed when it's submitted, so later changes to the record can be detected.
Claude, from Anthropic.
Nothing. The beta is free for student and professor testers.
Use the Join the Beta button to request access, or Schedule a Demo to book a call.