J Welcome from Jennie Soft Tech Kit no. 002

The score is in the questions you got wrong.

Not another course. A system that makes Khan, Magoosh, UWorld — or your $80/hr tutor — about 3× more efficient. Drill only the questions you actually missed. Stop studying the stuff you already know.

$27 built piece by piece · $420

One-time · Lifetime · 14-day refund · Claude + ChatGPT + NotebookLM

get the kit — $27 → same system I built for my own kid — j.

Test in 4 weeks? Crunch Mode is built for you.

The Mistake Engine

Upload the practice test. AI builds a podcast on what you missed.

You already know rereading the textbook doesn’t work. The two things that do: testing yourself + spacing it out. The Mistake Engine runs both — automatically, from the questions you missed this week.

You upload your wrong answers to NotebookLM. It builds a study guide, an audio overview, a quiz, and flashcards — grounded only in your errors. Listen on the way to school. Drill the weak patterns. Retest. The error log gets shorter.

Studio · generating

Why you missed 11 grammar questions.

Audio Overview · Brief format · building…

“Subject-verb agreement when there’s a prepositional phrase between them.”

0:00 ~ 5 min

← podcast from your mistakes

live preview · the workflow you actually run

A torn spiral-notebook page with handwritten error log entries Q4 through Q33, coffee ring stains, and handwritten margin notes in red pencil

← pattern showed up on Q4, Q11, Q18.

The error log

A real artifact, not a marketing graphic.

This is what the Mistake Engine builds. Pencil. Spiral page. Question numbers down the left. Cryptic notes on what each one tested. Margin annotations when the same trap shows up twice.

Three weeks in, the page looks like this. By week five, three patterns dominate. By week eight, those three patterns are mostly gone. That’s how the score moves.

What changes

Same hour. Same AI. Mistake Engine in front.

A real ACT prep session — the exact same hour, the only difference is whether you opened a chat cold or ran the Kit’s workflow.

Before · generic AI

Generic plan. No memory of what the student missed. No drilling on actual weak spots. By next practice test, the same 11 grammar questions are still wrong.

After · the Mistake Engine

Specific pattern. Targeted drill. The next practice test shows the gain.

Same hour. Better hour.

By Sunday of week 1

You’ll know it’s working. By the end of week one.

Not in 60 days. Not after the test. By the first Sunday you sit down with this Kit, you’ll have specific, measurable things in hand.

Day 1

A real baseline score from one official full-length test, taken under strict timing.

Day 2

The 3 AI tools (Claude + ChatGPT + NotebookLM) set up — copy-paste primer + project knowledge filled.

Day 3

Your top 3 named weak patterns — surfaced by NotebookLM from your scored test.

Day 4–6

Three 25-min Mistake Engine drills on those exact patterns — logged + retested.

Day 7

One timed section retake. A measurable shift on the patterns you drilled.

Sunday review

An audio overview from your mistakes. Listen on the way to school the rest of the week.

If by Sunday night you don’t have all six — email me the word refund. Same-day money back. The Kit only works if it actually moves you in seven days.

The 3-tool stack

Three tools. Three jobs. Each one has a superpower.

Stop using AI tools interchangeably. The Kit shows you exactly when to use each one — and what to never use it for.

Claude

the tutor

Long, careful reasoning. Error analysis. A persistent ACT Tutor Project that remembers your student across chats (Claude Memory, free). Best feedback on writing.

ChatGPT

the drill partner

Voice tutoring in the car. Study Mode for Socratic walk-throughs (free). Scheduled Tasks for Sunday review. Canvas for essay editing.

NotebookLM

the blender

One uploaded test becomes an audio podcast (Brief, Critique, or Debate), a study guide, a quiz, a mind map, and flashcards — grounded in your sources.

Inside the Kit

Eight named traps. Drill them until automatic.

Every ACT question is engineered to make at least one of these look right. Once you can name the trap, you stop fooling yourself.

Eye-candy answer

The longest, fanciest-sounding choice. Almost always wrong on rhetoric questions.

Partial-solve

You solved for x. The question asked for 2x+1. Always re-read the last line.

Off-by-one decimal

Answer is 0.0042; trap is 0.042. Watch units. Watch placeholder zeros.

Inverted relationship

Chart shows X up = Y up; trap says X up = Y down. Always check direction.

Right answer, wrong Q

Factually correct in the passage — but doesn’t answer the actual question.

Half-truth in Reading

First half correct, second half slips in a claim the passage didn’t make.

Tempting extreme

Choices using always / never / all / none. Almost always wrong.

Outside-knowledge

True in the real world — but not in the passage. Point at the sentence or skip.

Built for three people

A real chapter for each — not one generic guide with a “parents corner.”

Whoever opens this Kit gets instructions written for them, in plain language, with prompts they can copy.

Students

25 minutes a day, then you’re done. Voice-mode drills in the car so you don’t have to sit at a desk. A reset for the night before a test. One rule: try the problem yourself first — AI is the tutor, not the answer key.

Teachers & tutors

How to run this Kit in a prep class. Differentiate by baseline score. Give individualized feedback at scale using NotebookLM. The academic-integrity conversation, scripted.

Parents

The five things that actually help — and the five that backfire. What to spend money on. The weekly AI prompt that tells you if your kid is on track without nagging.

Built on real learning science

Two study techniques actually work. The Kit runs on both.

Testing yourself + spacing it out. Everything else — highlighting, rereading, watching videos — is low-yield. AI doesn’t change the science. It just makes the diagnosis fast. The actual work of missing and re-attempting stays effortful, on purpose. That’s the part that builds the score.

Sources: Dunlosky 2013 (practice testing + distributed practice rank high-utility) · Roediger & Karpicke 2006 (testing effect) · Bjork 2011 (desirable difficulty)

“mom’s notebook actually worked” — a.

What’s inside

Seven working parts. Built piece by piece.

If you bought each piece as its own course, it adds up. Today, you get all of it for the cost of one ACT tutoring half-hour.

The Mistake Engine workflow (the spine of the Kit) $67

3-tool AI setup (Claude Projects + ChatGPT Study Mode + NotebookLM) $47

25 copy-paste prompts — ACT-specific, not generic $95

Section playbooks (English, Math, Reading, Science, Writing) $76

Crunch Mode · 30, 60, 90 day plans + 4-week scenario $47

Trap library + pacing math + calculator code (TI-84) $50

Teacher chapter + parent chapter (run it w/ a class or without nagging) $50

Today, all of it $27

unlock all 7 parts — $27 → 14-day refund · email the word “refund.”

Includes a 4-page printable checklist — weekly rhythm, trap library, test-week countdown, morning-of pacing card. preview the checklist →

Compared honestly

vs. every other ACT prep option.

The Kit isn’t built to replace official materials — it’s built to make them 3× more efficient.

Option	Price	Personalized to your mistakes	Works on free AI tools	Updated for 2026 Enhanced ACT
Khan Academy ACT	free	no	n/a	partial
Magoosh Premium	$129/yr	no — generic Qs	no	partial
UWorld QBank	$99–249	no — same Qs for all	no	yes
Princeton Review Online	$299+	no — pre-recorded	no	yes
Private ACT tutor	$80–200/hr	yes	n/a	yes
Acely AI ACT tutor	$49–149/mo	yes	paid only	yes
The ACT Prep Kit	$27 once	yes	yes (Claude + ChatGPT + NotebookLM)	yes

The honest pitch: use this alongside Khan + an official practice book. The Kit makes whichever course you already own work 3× harder.

if you’re the student reading this

Your parent didn’t buy you another course. They bought you 25 minutes a day instead of 3 hours.

No worksheets. No videos to half-watch. You upload the test you just took, the AI tells you the exact pattern you keep missing, you drill it for 25 minutes, you go live your life. The score moves because you’re drilling the 11 questions you actually get wrong — not the 200 you already know.

open the Kit. start with section 2. — j.

5 kids. 2 currently in the test years.

Why I built this

I built this before I built any of it.

I’m a mom of five. Two of mine are in the test years right now. I built the Mistake Engine, the prompts, the printable checklist, the whole system for my own kids first — then realized every parent I knew was asking me the same question.

I don’t teach test prep for a living. I built the thing I needed for the kids in my own house. That’s the entire pitch.

— Jennie

the promise

Try it. If it doesn’t change how you study, I send your $27 back.

Build the three tools. Run the Mistake Engine for two weeks. If by then it hasn’t changed how you study — email me the word refund and I’ll send your $27 back same-day. No form. No “are you sure.” No back-and-forth. I want it to work. If it doesn’t, I don’t want your money.

— Jennie

Questions

The honest answers.

Do I need to pay for Claude, ChatGPT, or NotebookLM?

No. Every core workflow runs on the free tier of all three tools. If you already pay for Claude Pro or ChatGPT Plus, the Kit flags where that helps — but free is genuinely enough to raise the score. (Updates included if Anthropic/OpenAI/Google rename features — the workflow still applies.)

How is this different from Khan Academy, Magoosh, UWorld, or Princeton Review?

Those teach you the test. The Kit teaches you a meta-system for using AI to study what you got wrong — on top of whatever course or question bank you’re already using. Khan is free + official; UWorld is the gold standard for question banks. The Kit makes both about 3× more efficient because it converts every wrong answer into a targeted drill instead of letting them pile up in a notebook. Includes a teacher chapter for running the Kit in a class.

Test is in 4 weeks. Too late?

No. The Kit has a 30-day “Crunch Mode” plan specifically for four-week timelines, with an expanding spacing schedule on your top 3 weakest patterns. The error log and Mistake Engine produce measurable gains even over 4 weeks.

Does this work for the 2026 Enhanced ACT?

Yes. Written specifically for the Enhanced ACT — correct section counts (7 Science passages including 1 unscored field-test), correct scoring (composite = English + Math + Reading averaged; Science feeds the STEM score, not composite). Optional Science + Writing addressed with guidance on whether to sit for them based on your college list.

How does access work?

Password-protected web guide. Works on phone, laptop, tablet. The 4-page printable checklist is a separate one-click print/PDF download. You’ll get the access link the second you check out — bookmark it.

one more thing

$27. One tutor session is $80.
The session ends in an hour.
This Kit doesn’t end until you do.

Buy it. Read it in one sitting. Build the three tools. Watch what happens on the next practice test.

start the next practice test smarter →

One-time · Lifetime · 14-day refund

no app. no account. no monthly.

Stripe-secured checkout Instant access after purchase Lifetime updates included 14-day no-questions refund