This is a fictional walkthrough, but the interactions are real. Everything Maya does in this post is something Fennie actually supports today, in roughly the way I describe it.
Maya is a sophomore studying mechanical engineering. She's taking five courses this term: differential equations, statics, materials science, thermodynamics, and a writing seminar. Her diff eq midterm is in eleven days. She's behind on materials science notes. Her writing seminar paper draft is due Friday.
She has Fennie open on her phone and her laptop. Here's what Tuesday looks like.
7:42 AM — alarm, shower, walking to coffee
Phone in pocket. Fennie not open yet.
She's not thinking about studying. This is intentional. The plan will be there when she sits down — she doesn't need to be planning while she walks.
8:10 AM — sitting down with coffee, opens the daily plan
She taps Fennie. The morning plan is already there. Four items today.
Tuesday, April 15
- Review notes: second-order linear ODEs (8 min)
- Flashcards: diff eq Laplace transform deck — 9 cards due (10 min)
- Lesson: variation of parameters walkthrough (22 min)
- Quiz: ODE methods mixed (12 min)
Total: ~52 minutes
She notices the diff eq weighting. The midterm is on the calendar in eleven days. The plan started ramping toward it three days ago.
The materials science note backlog isn't on the plan today. She knows why — she manually pushed materials to lower priority over the weekend because of the diff eq midterm and the writing paper. Fennie respects that choice but quietly notes it. By Saturday, materials science is going to start showing up again whether she wants it to or not.
She does the first item before her second cup of coffee. Eight minutes of reading her own clean notes from last Wednesday's lecture. The notes are short, in her words, with one worked example at the bottom. She remembers about 80% of it. The 20% she fuzzed on, she makes a mental note to revisit.
8:25 AM — flashcards on the train
Train ride to campus is 18 minutes. She runs the flashcards on the way.
Nine cards on Laplace transforms. Six she nails. Two she fumbles — partial fractions in the inverse transform context — and one she just fails. The system marks them for re-surfacing in two days.
She doesn't think about it as "studying." It feels like a reflex now. Eight months in.
9:00 AM — class: thermodynamics
Phone away. Eyes up.
She takes handwritten notes during class. Her preference. She'll transcribe the important parts later. Fennie isn't involved here.
10:15 AM — between classes, the variation of parameters lesson
Library, 45 minutes before her next class.
She opens the lesson item from the morning plan. It's a 22-minute walkthrough on variation of parameters — the method for solving non-homogeneous ODEs when undetermined coefficients doesn't work.
The lesson runs in chat-explanation form. It starts with a worked example. Then a small twist. Then asks her to set up the next problem on her own before showing the solution.
She gets stuck on building the Wronskian. She types in chat: "I don't see why we need this determinant. Why not just guess and check?"
Chat explains. Not by answering "because the algorithm requires it" but by walking her through what the Wronskian is actually doing — checking that the two solutions you're using are linearly independent, which is necessary for the variation method to produce the right particular solution. It does this in three short paragraphs with a tiny example.
She gets it. Continues the lesson. Finishes in 24 minutes — slightly over the estimate. Fine.
11:00 AM — class: differential equations
Phone away. Two hours of class, including a problem session at the end.
During the problem session she gets stuck on a problem the professor set. It's a fourth-order ODE with a forcing term that doesn't fit any standard form. She writes it down. She'll bring it to Fennie later.
1:00 PM — lunch and a 12-minute quiz
She picks at a sandwich and pulls up the morning plan's last item — the ODE methods mixed quiz.
Twelve minutes. Six questions. She gets four right, two wrong. One of the wrong ones is on undetermined coefficients (she knew it was rusty), and the other is on the same Laplace partial fractions that tripped her on the morning flashcards.
The system notices. Memory updates: undetermined coefficients moves from "shaky" to "weak." Laplace partial fractions stays "weak" — third strike this week.
Tomorrow's plan is going to surface both. She doesn't have to think about that.
2:00 PM — class: statics
Phone away.
3:30 PM — open chat to ask about the problem from this morning
Coffee shop near campus. She types out the problem from the morning class — the fourth-order ODE — and asks how to set it up.
Chat doesn't solve it. It asks: "What method would you try first?"
She types: "I don't know. It's higher order than anything we've done."
Chat: "Right. So the move is to recognize that for any constant-coefficient linear ODE — even higher order — the characteristic equation approach still works. What's the characteristic equation here?"
She writes it out. Gets the polynomial. Has to factor it. Stuck again.
Chat: "Try the rational root theorem. What are the rational roots of this polynomial?"
She tries. Finds two. The other two are complex.
Chat walks her through what complex roots mean for the homogeneous solution. Then asks her to set up the form of the particular solution given the forcing term — which is something she does know how to do.
20 minutes total. She has the problem solved. She didn't get the answer handed to her. She built it.
She drops a note: "Higher-order constant coefficient ODEs use char equation just like 2nd order. Complex roots → sin/cos terms. RRT is useful for factoring." Three sentences. She'll regenerate this into a flashcard later.
4:30 PM — quick flashcard run, materials science
She has 15 minutes before a club meeting.
She remembers materials science is on her backlog. She opens her materials notes from last week and skims them. Bad notes. Bullet points without context. She doesn't have time to clean them up now, but she generates flashcards from them anyway, low confidence — the cards are going to be rough.
This is fine. Tomorrow's plan will surface them, she'll get half wrong, the system will notice, and the bad-card flag will push her back to the source notes. She'll clean those up Wednesday evening.
5:00 PM — club meeting
Phone away. Engineering club. Existence outside studying.
7:30 PM — dinner, then writing seminar work
The paper draft is due Friday. She sets a 50-minute timer.
She's not using Fennie's chat for the writing — her seminar professor is strict about AI involvement on essays, and she respects that. What she does use Fennie for is essay outlining: she puts the prompt in, asks chat to ask her questions about her thesis, and uses the questions to find weak spots in her own argument.
It works because chat asks instead of answers. "What's the strongest counterargument to your thesis?" "If your thesis is true, what should we expect to see in the data?" "Where in your draft is the warrant for that claim?"
She rewrites her introduction. The body paragraphs are next, but tomorrow.
8:30 PM — note cleanup → quiz spawn
This is her favorite half-hour of the day.
She opens her ODE notes from this morning's class. Cleans them up — turning the lecture shorthand into a proper note, adding the worked example from the variation of parameters lesson, and writing a one-paragraph "what I learned today" at the bottom.
Then, two clicks: generate quiz from this note.
She does the quiz immediately. Five questions. Gets four right. The wrong one is — yes — partial fractions. Again.
Memory has now seen partial fractions fail four times in two weeks. Tomorrow's plan is going to have a real lesson on partial fractions, not just flashcards. She doesn't have to ask.
9:15 PM — looks at tomorrow's preview
She glances at the morning preview in Fennie. It's there because she enabled it.
Wednesday, April 16 — preview
- Lesson: partial fractions for inverse Laplace (25 min)
- Flashcards: ODE methods deck (12 cards, 12 min)
- Review notes: variation of parameters (6 min)
- Quiz: undetermined coefficients targeted (10 min)
Materials science: due to surface Thursday or Friday.
The partial fractions lesson is the win. She's been getting that wrong for two weeks. The system finally has enough signal to elevate it from "card I keep failing" to "topic I need to actually relearn."
She closes the app. Reads a novel. Goes to bed at 10:45.
What this day demonstrates
The whole loop showed up here. I didn't engineer the day to hit every feature; this is roughly what a Tuesday looks like for someone using Fennie consistently.
The morning plan was four items, totaled under an hour. The plan was responsive to her calendar (midterm in 11 days), her recent performance (Laplace partial fractions kept failing), and her own priority decisions (materials science on hold for a week).
Chat showed up twice. Once for a lesson she was working through. Once for a problem she was stuck on. In neither case did chat hand her an answer. In both cases it asked questions, gave hints, and let her find the path.
Notes connected to everything else. Her morning ODE notes spawned a quiz. Her bad materials notes spawned bad flashcards that she'll have to clean up tomorrow. Her writing seminar didn't use chat for the writing itself, but did use it for outlining.
Memory was running quietly the whole day. Tracking what she nailed, what she fumbled, what kept failing. The fact that tomorrow's plan automatically elevated partial fractions to a real lesson — that's Memory doing its job.
The thing I want to point out: Maya did not spend any time deciding what to study. She woke up, the plan was there, she did it across the day, the system updated. She spent her decision-making energy on the actual material, not on the question of what to do next.
That's the whole point.
See what your Tuesday looks like with a plan that builds itself. Start with Fennie free — no card, no commitment.