Run in order. Prompts 1-4 build the funnel before you build the site — most founders do the reverse and end up with a beautiful page that no-one ever reaches. The funnel decides the site, not the other way around. Don’t write a single line of site copy until prompt 4 has produced a one-page funnel diagram with a conversion target at every step.

1. The customer’s actual path. “Here is my MVP from M5, my positioning from M6, and my segment from M2: [paste]. Map the path a stranger in my segment takes from never having heard of me to having paid. Don’t describe the funnel I wish existed — describe the funnel that exists in their experience. List every touchpoint in order: where they first see me, what they do next, what they see after that, what makes them decide, where they pay, what they receive. For each step, name the medium (post, ad, page, email, DM, call), the action required of them, and the friction that could stop them. Output a numbered list of 5-9 steps maximum. If the path has more than 9 steps, the funnel is broken — cut steps before adding any.”
2. The conversion target at every step. “Take the path from prompt 1. For each step, assign a realistic conversion target — the percentage of people from the previous step who continue to this one — based on benchmarks for my channel and segment. Be honest. Reach to landing page: 1-3%. Landing page to email signup: 10-30%. Email signup to first purchase on a $497 product: 1-5%. Multiply across the full funnel: if I want 10 paying customers in launch week, how many strangers must enter the top of the funnel? Output the funnel as a table — step, conversion %, people entering, people exiting — and the top-of-funnel number I need to hit. If the top-of-funnel number is implausible for the channels I have, the funnel needs fewer steps or a different offer, not more traffic.”
3. The friction inventory. “For each step in the funnel from prompts 1-2, identify the one specific point of friction most likely to lose the customer. Friction examples: a form with too many fields, a price that appears too early, a price that appears too late, an unclear next action, a payment method missing, an email arriving at the wrong moment, a calendar booking step that adds a day of delay, a ‘contact us’ where a ‘buy now’ belongs. For each friction point, write the one specific fix and rank by leverage. Output the top 3 friction fixes I do this week — the ones that move the funnel from leaking to holding. Ignore the rest until after launch.”
4. The funnel diagram on one page. “Combine prompts 1-3 into a single diagram I can draw on one page by hand. Top of page: source (where strangers enter). Middle of page: the 5-9 steps with conversion target between each. Bottom of page: paying customer. Side margin: top 3 friction fixes, in order. This is the page I pin above my desk. Every decision about site, copy, email, ad, post must serve a step on this diagram or it doesn’t ship. Output the diagram in plain text or markdown table, and a one-sentence rule for what gets cut: ‘if it doesn’t serve a step on this page, it waits.’”
5. The minimum viable site checklist. “Based on the funnel from prompt 4 and the positioning from M6, list the absolute minimum pages I need live on launch day. For a $497 digital product or a $1,200 physical product, that’s typically 3-5 pages, not 10. For each page, name: (a) the one job the page does in the funnel, (b) the one CTA on the page, (c) the proof element that page carries (testimonial / number / named partner / specific result), (d) what’s on the page on launch day, (e) what’s deliberately not on the page until v2. Output as a build checklist with each page as a row. The MVS rule: if a page doesn’t appear in the funnel diagram, it doesn’t get built for launch.”
6. The post-purchase loop. “The funnel doesn’t end at payment — it ends at delivery and the first follow-up. Map what happens in the first 7 days after a stranger pays: the immediate confirmation, the access / delivery moment, the first-use prompt, the request for the first signal (testimonial, NPS, photo, review), the second touchpoint that opens the door to a repeat or a referral. For each, name the channel, the timing, and the one sentence that goes out. The post-purchase loop is where the funnel either compounds or stops dead. Output as a 7-day timeline. The launch isn’t complete until this loop is built — most founders ship the buy-flow and forget the after.”
7. The site copy in the voice from M6. “Using the voice rulebook from M6 [paste] and the never-say list [paste], draft the copy for the 3 most critical pages from prompt 5 — landing page, product / offer page, and post-purchase confirmation. Each page: a headline that names what the buyer gets in plain English, a 30-word subline that earns the right to that claim, the proof element, the one CTA, the one sentence that handles the most common objection. Reject anything that sounds like a launch announcement — write as if the page has been live for 12 months and a stranger has just found it. Output the draft copy for the 3 pages, plus 5 alternative headlines for each in case the first is too close to the AI default.”
8. Soft launch versus hard launch. “Given my funnel, my channel reach today, and my tolerance for things going wrong in front of strangers — should I soft launch (warm circle first for 1-2 weeks, learn what breaks, fix, then expand) or hard launch (public from day 1, more visibility, more risk, less time to repair)? Score me honestly on 5 dimensions: (a) is my warm circle big enough to produce 5-10 first transactions without paid traffic, (b) have I tested the buy-flow with at least one stranger end-to-end, (c) is my post-purchase delivery actually working today not just on paper, (d) can I respond to a customer issue within 4 hours during launch week, (e) is the offer’s price defensible without discounting under pressure. If I score 3+ no, soft launch. If I score 4+ yes, hard launch is open. Output the recommendation and the reasoning.”
9. The launch sequence calendar. “Based on the soft / hard decision from prompt 8, draft the launch sequence on a dated calendar. For a soft launch over 14 days, that’s typically: day -7 to day -1 — warm-circle invitation + small list warm-up; day 0 — soft open to warm circle only with a private link; day 1-3 — fix what breaks, document what works; day 4-7 — open invitation to second-ring (waitlist, segment-specific list); day 8-14 — public open with the first paid traffic. For each day, name: the action, the audience, the message, the channel, the success metric. Output as a table. The calendar is the launch — not the announcement. The announcement, if any, sits on day 8 or later.”
10. The first message to the warm circle. “Write the one message that goes to my warm circle on day 0 of the soft launch. The audience: 20-100 people who already know me — past customers, peers, segment-specific contacts, the list M4 helped me build. The job: invite them to be among the first to buy or to share with one specific person they know. Voice: from the M6 rulebook. Tone: peer to peer, not announcement to audience. Length: under 200 words. Includes: what the offer is in one line, who it’s specifically for (named segment), the price, the private link, the one ask (buy if it’s for you, forward if it’s for someone you know, ignore if neither). Reject any sentence that sounds like a press release. Output the message, plus the 3 versions tailored for the 3 highest-leverage sub-segments of my warm circle.”