Three live days with Kane, Alessia and Wes: find the paid problem inside your expertise, turn your method into a one-page product canvas, and build an ugly first version you can show a real human. Then 12 weeks to build the product, launch it and sell it. Useful, easy to run, and a bit silly on purpose.
The shape
Pre-work4 posts in Kane's AI PRODUCT CHALLENGE Facebook group before Day 1.
3 live daysFind it · Design it · Start building it. One thing made each day.
The offerTurn Your Expertise Into an AI Product: 12 weeks, cohort 1 starts Tue 1 Dec.
12 weeksBuild the Product → Launch It → Sell It. One visible thing shipped every week.
The promise, from today's email
No code, no existing business and no idea needed. Products: AI agents, intelligent apps, digital assistants, smart coaching systems and automated workflows. For entrepreneurs, business owners, consultants and first-timers.
Proof we can say: Inc. 5000 two years running, launched more than a dozen tech companies, co-founder, board member or advisor to 30+ tech companies. The core audience is 45–70: lots of experience, not much tech. They need "oh, I can actually do this" more than they need novelty.
Four rules that keep it easy
1. One room, two tracks. Everyone gets the same teaching and the same worksheet, then a 10-minute track twist at the end.
2. Every day ends with a thing. One artefact, a done-when test and a real-person check. More proof, not more content.
3. Small and real beats big and imaginary. One use case, no hardening, no code. Kane on 19 Oct: "we're not going to be teaching them how to actually hard code or harden products."
4. No rabbit holes. No new tools unless they help today's thing. "Don't worry what it's called yet. First, make it useful."
Every day runs the same way
Win or poll5 min
Homework check5 min
Short teach15 min max
Demo5 min
Build sprint20–40 min
Share-backs10 min: one insight, one stuck, one next step
Homework + link5 min
The same order every day, so nobody has to learn a new format and the presenters don't burn out. Each day closes with a one-line nudge toward the 12 weeks; the full invite only comes on Day 3.
Two tracks
A · The Curious
"I know my expertise is worth something, but I don't know what AI product to build yet."
Leaves with a named buyer, a product canvas and an ugly v1 they've shown to someone.
B · The Builders
"I've built or started something, but I need buyers, partners and sales conversations."
Twists stay light: a clear buyer, a pilot offer, one sales conversation and one partner DM.
Sorted by the track poll (pre-work post 4) and a one-letter reply in the Zoom chat: "A" or "B".
1
FIND IT
Find the paid problem inside your expertise.
Opener: the enemy
The enemy isn't scary tech. It's building clever things nobody pays for. "Most AI products don't fail because the tech is bad. They fail because nobody had a painful enough problem." Kane: "You don't need to become a software company. You need to package the thing people already trust you for." Wes: "And this is how that becomes something people can use."
Big idea
Build what people already pay to fix. Your best product is hiding in the questions you've answered a hundred times.
Funny bit
The Museum of Unnecessary AI (the Shark Tank of terrible ideas). Kane and Wes pitch awful AI products with total confidence: toast-done alerts, a life coach for houseplants, a chatbot that apologises for your emails. The chat hands out awards:
🏆 The "Nobody Asked For This" Award
🎩 The "That's Just ChatGPT With A Hat On" Award
👩👦 The "My Mum Said It Was Good" Award
🤖 The "Technically Impressive, Commercially Useless" Award
The lesson: "If the idea is funny because nobody would pay for it, don't build it." Then: "Now let's find yours."
Build sprint
The Money Map (25 min). List the problems people already come to you with. Take your best 3 and score each one 1–3 on:
💰 MoneyDo people already pay to solve this?
🔥 PainDoes it hurt often enough?
🛠 MethodCan I explain how I fix it?
Under 7 out of 9? Park it. Then the reach test: can you name, and actually reach, one real person who has this problem? If not, park it too.
Hey Genius, find my paid problemIt asks one question at a time: who asks you for help, when the problem shows up, what it costs them, what you do differently, and who you could contact. It doesn't suggest a product until you've confirmed its summary.
Artefact
"I help [specific person] solve [painful problem] using [my method], so they can [result]."
Done when: one named buyer · the problem in their words · the line above · one real person you'll contact. No real person? You stay on evidence on Day 2, not features.
Real-person check
One conversation where you're not selling anything. "I'm exploring a simple AI-supported way to help with [problem]. I'm not selling anything today. Could I ask you four quick questions?"
When did this last happen? · What did you do? · What was frustrating or costly? · What would a useful first result look like?
One conversation, not three. Never "would you pay?", because people will politely lie to your face.
Track twist
A · CuriousPick one problem and write the first useful answer your product would give.
B · BuildersOne stalled sale: name the real objection (problem, trust, timing, price or clarity), one buyer to re-contact and one partner. Just one, not three.
2
DESIGN IT
Turn your method into the product.
Opener: myth-bust
"You're not here to build a robot version of yourself. You're here to build a guided version of your best thinking."
Big idea
Your method is the product. AI is the delivery van. "AI shouldn't replace your judgement. It should remove the repeatable drag around it." The AI asks, sorts, drafts and spots gaps; you make the call, hold the nuance and lead.
The menu of 7 shapes: Diagnostic (questions, a score, next steps) · Advisor (tailored guidance) · Planner (goal to plan) · Coach (reflect, decide, practise) · Workflow (messy in, useful out) · Sales Helper (scripts, objections, proposals) · Training Companion (apply a course). Most people should start with a guided assistant or workflow, not an app.
Hey Genius, turn my method into a product canvas
Funny bit
The taste test: Generic vs Guided. The same client question answered by plain AI, then by AI guided by Wes's method. The chat votes, scoring on four things: does it sound like the expert, does it give a useful next step, does it know what it can't decide, does it follow the method. Kane explains why the winner wins commercially. Then the roast: we let AI name a method and Kane gently destroys "The Quantum Success Acceleration Matrix 360™".
Artefact
A one-page canvas and a promise: "This helps [buyer] go from [messy state] to [better state] by [method]."
Done when: all 8 boxes filled · one shape picked · a 60-second explanation: "When this happens, the user brings X and leaves with Y."
Real-person check
A 5-minute walkthrough with one real person, ending with: "If this existed, what would you want it to help you do first?" Bring back a quote, a confusion or an objection, not "sounds great!"
Track twist
A · CuriousNarrow it to one shape and write the first user journey, step by step.
B · BuildersThe pilot offer only: "My pilot helps [buyer] solve [problem] in [timeframe]. I'm looking for [number] testers at [pilot price, or free for feedback]."
3
START BUILDING IT
Build the ugly first version, then the next step.
Opener
"Today isn't the final product. It's the first proof that your expertise can become an experience."
Big idea
If v1 doesn't embarrass you a little, you waited too long. Build ugly, show early, polish later.
Build sprint
The Ugly V1 Build (40 min), from a template. Fill in the recipe: who you help · problem · method · rules · tone · questions to ask · the output to produce · what not to do. Paste in your promise and steps, add 5 intake questions, test it on one real client scenario, and save the link.
1 · SupportedThe one route we demo live, built in the supported tool.
2 · Template fallbackUse the web template and paste your skill into the build form.
3 · Spec-onlyFinish the build spec now, then bring it to the tech drop-in.
Hey Genius, build my ugly v1A producer handles login problems off stage. Nobody's progress waits on a password reset.
Funny bit
Break My Bot. The chat gets attack cards: 🃏 the vague user · 🃏 missing or contradictory information · 🃏 a request way outside the product's job. They try to break Wes's demo.
Wes explains what broke.
Kane says whether it matters commercially.
Wes fixes one thing live.
"Your bot doesn't need to survive the internet. It needs to help the right buyer take the next right step."
Artefact
An ugly v1 link (or a finished build spec) and an offer message.
Done when: link or spec ready · 3 test questions run · one thing changed because of testing · the invite message sent or scheduled. A screenshot alone doesn't count.
Real-person check
Invite one real person to a 15-minute test. "Send it before you go to bed. Do not edit it for 47 minutes."
Track twist
A · CuriousGet it working once. Pretty is optional. Book the 15-minute test.
B · Builders5-minute buyer role-play in pairs: why is this urgent now? What have you tried? What would make it worth paying for? Would you test v1? Then send one partner DM.
Self-score + invite
Five yes/no questions: named buyer? product promise? rough method? ugly v1? first message sent?
3 or more: "The 12 weeks turn this into a real product people can use and buy."
Fewer than 3: "The 12 weeks get you there without wandering around in AI tutorials."
Then the invite: "In three days you found a real problem, designed a first useful result and started a working version. The hard part now isn't more prompts. It's finishing, testing with real people and turning it into an offer." Three gates: Working v1 → Tested v1 → Sellable offer.
Running notes for the 3 days
3 core commands:Hey Genius, find my paid problemHey Genius, turn my method into a product canvasHey Genius, build my ugly v1. The other Genius Challenge commands go on a bonus command sheet.
The offer: it goes at the end of Day 3. The calendar's "Challenge Bonus day" (Mon 2 Nov, 22:00–00:00 SAST) could be the repitch and Q&A. That's still a proposal.
Never say: no income promises and no invented results. The harness and Moat get a preview only ("members hear first"), with no dates.
48-hour follow-up to confirm
When
What
End of Day 3
The live invite
Same day
Replay email
Next morning
"Look what you built" email: their artefacts from the 3 days
24 hours before close
Decision email
Final hours
Last reminder with a short FAQ
Based on Kane's webinar playbook: most of the follow-up work happens in the 48 hours after the event.
Facebook pre-work
Four core posts for Kane's AI PRODUCT CHALLENGE group. Each one is fun, and each one collects material we use live.
Post 1 · intro
"Introduce yourself in 3 emojis, then finish this sentence: People usually come to me when they need help with ______."
Starts the Day 1 Money Map before Day 1.
Post 2 · the 100-times question
"Drop the question you've answered 100 times. And who asks it most often?"
Real questions for the taste test and Break My Bot, plus a buyer for each one.
Post 3 · your AI clone
"Your AI clone runs things for a week while you're on the beach. What is it absolutely NOT allowed to do? And what would you gladly let it do every week if it saved you 3 hours?"
The "not allowed" list becomes the AI job and human job on the Day 2 canvas. The funniest answer gets read out live.
Post 4 · pick your track
"A: I know my expertise has value, but I don't know what AI product to build yet. B: I've built or started something, but I need buyers, partners and sales conversations."
Follow-up comment: "Can you name one person you could message this week who has this problem?"
Optional bonus poll, if there's a spare day: "What would you build?" Agent · Assistant · Workflow · Coaching system · App · "I don't care what it's called, I just want it to make money."
The 12-week program
$997 paid in full, or $297 deposit + 3 × $297 ($1,188)
Agreed with Kane on 19 Oct for "Turn Your Expertise Into an AI Product": 12 weeks, cohort 1 starts Tue 1 Dec 2026. TBC with Kane whether it still holds now the challenge is 3 days.
The pitch: "Over 12 weeks you'll turn your expertise into a working AI product, test it with real buyers, launch it by link and start selling it, without becoming technical." It's an AI product course. The brain is just the engine under the bonnet.
The weekly rhythm
Monday · Build Lab
90 min, recorded. Wins → gate check → short teach → demo → 25-min sprint → share-backs → one action.
Thursday · with Wes
Clinic or hot seat. No new frameworks: tech triage, test reviews, gate checks, commitments.
The rule: ship one visible thing every week and post it in the group. No new tools unless they help this week's thing.
1Weeks 1–4
Build the Product
W1 · Paid problem. Ship: buyer, problem in their words, promise, one real conversation. Nobody builds before this.
W2 · Method map. Ship: 3–7 steps, AI job vs human job, shape picked, a "does / doesn't do" list.
W3 · The engine. Ship: your method, rules, tone, questions and output format loaded from a strict template.
W4 · First useful skill. Ship: one working skill, 10 test questions, a failure log.
Gate · week 4: paid problem · named buyer · method map · first skill working.
2Weeks 5–8
Launch It
W5 · Guided experience. Ship: welcome message, intake flow, "what this can and can't do".
W6 · V1 live by link. Ship: the link, a 90-second demo video, 5 testers invited. No new features until a tester finishes the core path.
W7 · Test week. Ship: a scorecard of 20–25 real questions plus 3 trick ones. It passes if it helps the right buyer, asks instead of making things up, sounds like the expert and gives a next step people would use.
W8 · Fix, protect, package. Ship: fixes retested, one-page offer, an IP and usage checklist, a feedback request.
Gate · week 8: v1 live · tested with real questions · offer drafted · IP boundaries clear.
3Weeks 9–12
Sell It
W9 · Price + pilot. Ship: one of 3 options (free pilot for a testimonial, low paid pilot, or a bonus inside an existing offer) and a 10-name list.
W10 · Page + checkout. Ship: one simple page, a checkout link, onboarding instructions. Watch one person try to buy.
W11 · Conversations. Ship: 5 buyer conversations (across weeks 11–12), an objection log, follow-ups and a partner DM.
W12 · Demo Day. Ship: a live demo. Peers give one strength, one risk and one next experiment. Plus a 30-day plan.
Gate · week 12: checkout live · 5 conversations · partner plan · next 30 days clear.
Six outcomes by week 12
A narrow product with a named buyer and a first useful result
A working, shareable v1
Test evidence from real users and trick questions
A clear paid pilot or offer, with a way to pay
5 buyer conversations, or partner introductions instead
A 30-day launch plan
We grade the evidence and the next experiment, not revenue. The first paying users are a coached goal, never a promise. The proposed holiday break is 21 Dec – 3 Jan.
What's included
Item
Status
Monday Build Lab, recorded
Wes, 24 Oct replaces the "twice a week" agreed on 19 Oct
Thursday clinic or hot seat with Wes
Agreed 19 Oct
Lifetime access to the Facebook community
Agreed 19 Oct
IR × MyGentic Product Dev Suite for 12 weeks: product builder, progress dashboard, Kanban (MyGentic workspace or a Claude/ChatGPT plugin)
Agreed 19 Oct demo build still open
November pre-work: gather your method, materials and client questions
Suite price after week 12 · tokens, Moat and Stripe costs · partner terms · guarantee wording
TBC with Kane
Earlier bonus ideas (offer v3 kits, SATT, the objection bonuses) are parked until Kane confirms.
Open questions for Wes + Kane
Dates and times for the 3 days. The calendar still has the old 5-day slot (Mon 26 – Fri 30 Oct, 21:00–23:00 SAST) plus "Challenge Bonus day" on Mon 2 Nov. The 20 Oct plan proposed 9–13 Nov.
Does the program stay the same? 12 weeks, the price above and a 1 Dec start. And the name: "Turn Your Expertise Into an AI Product", or a "3-month sprint", which could be confused with the Sprint?
Entry: free, free plus a VIP upgrade, or paid?
The pitch: end of Day 3, with the Bonus day as the repitch? Who says the price? Who writes and sends the 48-hour emails?
Track B: the same offer, or a different next step later?
Day 3 build: which tool is the supported route? The Product Dev Suite demo go/no-go was set for 30 Oct.
Bonuses: confirm the 5 practical bonuses, and who builds them before 1 Dec.
After week 12: the suite price (it has to be published before enrolment), who pays for tokens, Moat and Stripe, and partner terms in writing.
Community: does Kane's AI PRODUCT CHALLENGE group become the lifetime cohort community, and who moderates it?
Sources
Email, "What if your next business was an AI product?" from Kane & Alessia (24 Oct): the 3-day format, day themes, audience, product types and proof points.
Fathom, Wes + Kane "Impromptu Zoom Meeting" (19 Oct): price, 12 weeks, 1 Dec start, cadence, Facebook community, Product Dev Suite, small MVPs.
Google Doc, "Turn Your Expertise Into an AI Product" new notes (19 Oct).
Fathom, MyGentic Founders Product Sync (20 Oct): the challenge needs working demos, full access 1 Dec.
Course Creator and Rockstar reviews (24 Oct): the daily rhythm, done-when tests, canvas, shapes, attack cards, gates, the 48-hour follow-up and the bonuses. Merged, picking the simpler option wherever they differed.