Fennie's Chat: A Tutor That Teaches Instead of Tells
The first time most students use Fennie's chat, there's a moment of friction. They paste in a problem, expect a solution, and instead get a question back. "What have you tried?" or "Where in the problem did you stop being sure?"
This isn't a bug. It's the entire design.
Fennie's chat is a tutor, not a chatbot. It's built to teach you the thing you're working on, not to do the work for you. That distinction shapes every part of how to use it well — what to ask, how to ask, when to lean on it, and when to do the thinking yourself.
What makes Fennie's chat different
Three things actually distinguish it from a generic AI chatbot:
It's connected to your courses. When you open chat, it knows what classes you're taking, which one you're working on right now, and what's coming up on your calendar. You don't have to set context. The system already has it.
It reads your notes and documents. Anything you've put in your Notes, any document you've uploaded — your textbook chapters, lecture handouts, problem sets — is available context. So you can say "based on my biology notes from this week, can you walk me through what I'm missing about Krebs cycle" and it actually has the notes.
It feeds Memory. Every concept that comes up in a chat — whether you nailed it, struggled with it, or asked the same question three different ways — informs Fennie's picture of what you know. That picture shapes tomorrow's plan and which flashcards surface next.
So a chat session in Fennie isn't a one-off. It's a teaching moment that becomes part of your study trajectory.
A real chat session
Here's roughly what a productive Fennie chat looks like, lifted from how students actually use it:
You: I'm working on my orgo problem set. Question 4 is asking me to predict the major product of an SN1 reaction with a tertiary alcohol and HCl. I drew the carbocation but I'm not sure if it rearranges.
Fennie: Good — drawing the carbocation is the right first move. Before we worry about rearrangement: where's the carbocation actually sitting on your structure? Tertiary, secondary, or primary?
You: Tertiary I think.
Fennie: Right. Now: rearrangement happens when there's a more stable cation nearby. Look at the carbons next to your tertiary carbon — any of them give you something more stable than tertiary if you shift?
You: No — they're all secondary or worse.
Fennie: So what does that tell you about whether the cation rearranges in this case?
That's the rhythm. It doesn't give you the answer. It surfaces the next decision point and lets you make it. Two more exchanges and you have the answer, and more importantly you understand why — which is what'll matter when a similar question shows up on the exam.
If you've used a generic AI chatbot, you'll feel the difference. Generic AI dumps the full reaction mechanism on you. You feel productive. You learned almost nothing.
What chat is genuinely good for
Working through homework problems. Bring your specific stuck point. The chat will help you over it without doing the rest for you.
Understanding a concept differently. "I read the textbook on Bayesian inference and it didn't click. Can you explain it like a coin-flipping game?" — this is a fantastic use of chat. Multiple framings of the same concept until one lands.
Reviewing notes you took in lecture. "Pull up my notes from yesterday and quiz me on the parts I sound confused about." Chat plus your notes plus the quiz feature is a powerful loop.
Pre-essay brainstorming. Talking through possible angles before you write. You don't get a draft — you get clarity.
Decoding a textbook chapter. Upload the PDF, ask "explain section 3.2 like I'm seeing this material for the first time."
Spot-checking your reasoning. "Here's my proof for why this limit converges. Is there a logical step I'm hand-waving?" Honest feedback fast.
What chat won't do (and why)
It won't write your essay. You can ask it to read your draft and tell you which paragraph is weak. You can't ask it to rewrite the paragraph for you. This is intentional and probably the most common frustration in the first week. It's also why faculty are starting to be less suspicious of Fennie users — the tool isn't optimized for laundering work.
It won't dump full homework solutions. It'll walk you through them.
It won't give you essay grades. It'll point out where the argument feels thin or the evidence weak. The grading is your professor's job.
It won't replace your textbook. Use the textbook. Then bring leftover questions to chat. The combination is much higher-leverage than either alone.
How to ask better questions
There's a real skill to using chat well. The students who get the most out of it are the ones who:
Show their work first. "Here's what I tried, here's where I got stuck" gets a 10x better response than "what's the answer."
Specify the kind of help they want. "Don't give me the answer. Just tell me whether my approach makes sense" is precise. The model will do exactly that.
Push back when the explanation doesn't land. "I don't follow step 3 — can you slow down" is fine. The model adjusts.
Use it for retrieval, not just exposition. "Quiz me on chapter 5" is a different (and often better) use than "explain chapter 5."
Don't restart the conversation every time. Each chat builds context. Ten messages in one chat is more useful than ten one-off prompts.
Chat plus the rest of the system
The thing that took me longest to internalize when I first used Fennie is that chat isn't a standalone feature. It's connected to everything.
A few flows worth knowing:
Chat → Note. Had a great explanation in chat? Save it as a note (one click). Now it's in your knowledge base for future sessions.
Note → Quiz → Chat. Take a quiz, get something wrong, click into chat to talk through the question.
Chat → Flashcard. "These two concepts keep tripping me up — can you make me flashcards?" Two clicks, deck created.
Chat ↔ Calendar. Friday's the test? Chat will gently steer toward review and practice problems on the relevant material in the days leading up.
Chat → Memory. Everything feeds back. Tomorrow's plan reflects what you struggled with today.
This is the part that actually matters and it's hard to convey in a tutorial. Chat in isolation is fine. Chat as part of a connected study system is genuinely different.
Common first-week stumbles
Asking for direct answers and getting frustrated. Reframe. "Walk me through it" works.
Pasting a giant block of text. Trim it. "Here's the relevant paragraph and my specific question."
Treating chat like Google. It's a tutor. The conversational depth is what makes it useful. One-shot questions are fine but you're missing 70% of the value.
Not connecting chat to notes and the calendar. This is the multiplier. If you only chat, you're using one piece of a system designed to work together.
A few more things worth knowing
The chat supports LaTeX rendering — drop equations in math mode and they render properly. It handles mermaid diagrams for flowcharts and process maps. Code blocks are syntax-highlighted. Long responses are streamed so you can start reading immediately rather than waiting for the whole reply.
There's a free tier (chat works without paying anything for basic models) and a premium tier ($4.99/mo) that unlocks the best models, the full memory system, and unlimited generation. Most students start free, and either stay or upgrade once they realize the memory system is doing real work.
Try Fennie's chat — the tutor that teaches instead of solves. Free to start, no card.