Run: maxswarm #017 · Date: 2026-05-06 · Phase 1 author · Mission: 3-page About Us PDF + page-3 interactive demo
#StepOwnerDaysNotes
1Copy outline (3 pages, voice locked)draper + Ewing approval1Page 1 origin, Page 2 method, Page 3 demo teaser 2Copy v1 full draftwriter1Blocks design + image briefs 3Caricature briefs (3 founders + style ref)quarterback + draper0.5Parallel with copy v1 4aCaricature generation (Midjourney + Sora)AI iter w/ Ewing approval3-5LONGEST POLE — style consistency across 3 faces is the killer 4bLogo + AZ backdrop + golf scene assetsAI generation1-2Parallel to 4a 5Layout v1 (InDesign or Figma)contractor designer2Needs copy locked + assets in 6Demo build (page-3 interactive)agent (Vercel/Next) + Ewing review2-3Can start in parallel after spec lock day 1 7Demo hosting (Vercel subdomain)architect0.25Trivial once demo builds 8QA pass (audit-quality + 3 founder testers)audit-quality + Ewing's network1Gate 3 below 9PDF export + final proofcontractor designer0.5— 10Send v1Ewing0.25—
Critical path total: 1 (outline) → 2 (copy) → 4a (caricatures, longest pole) → 5 (layout) → 8 (QA) → 9 (proof) → 10 (send) ≈ 10-12 working days assuming Ewing has ~30 min/day on this and caricatures don't redo more than twice.
Copy outline (1) blocks everything downstream. must lock in day 1. Once outline is locked, three streams run in parallel: copy v1 (writer), caricature briefs (qb+draper), demo spec (agent). Caricatures (4a) and demo build (6) are independent. neither blocks the other. Demo can be hosted before PDF is done. Layout (5) is the choke point. needs copy locked AND all images in. If caricatures slip, layout slips 1:1. QA (8) needs the PDF mock + the live demo URL. Founder-tester recruiting can start day 1 (parallel slack to 5 founders, expect 3 to respond).
The real parallelism: demo build can finish + go live BEFORE the PDF is done. That gives us a working URL to paste into page 3 of the PDF and 1 week of soak time for tester feedback before send.
Today is 2026-05-06 (Wed). Three competing deals (HR.com, Design Precast, Capstone) eat ~80% of Ewing's bandwidth. Realistic v1 ship: 2026-05-22 (Fri) — 12 working days out. Stretch goal 2026-05-19 (Tue) if caricatures hit on first generation. Worst case slip-to-Memorial-Day-week: 2026-05-29 (Fri). Recommendation: commit to 2026-05-22 publicly, hold 2026-05-29 as buffer. Do NOT promise sooner. caricature consistency burns drafts.
AssetSourceExternal costRisk
Caricatures of 3 foundersMidjourney + Sora prompts (AI) — fallback to Fiverr illustrator $150 - 400$30 AI subs OR $400 humanHIGH. style match across 3 faces Next Chapter logo source filesInternal. confirm vector exists$0 if found, $200 redraw if notLOW. likely on file AZ desert / Sedona backdropAI generation (Midjourney) or Unsplash license$0 - 50LOW Golf sceneAI generation$0 - 30LOW. generic enough Layout designerExternal contractor (Upwork/existing)$300 - 800MEDIUM. find someone who can turn in 48hr Demo hostingInternal Vercel$0LOW Founder testers (3)Ewing's network$0MEDIUM. getting 3 responses in 48hr PDF font licensesInternal or Adobe Fonts via designer$0LOW
Total external $: ~$330–$1,280. Place caricature order day 1. that's the longest external pole.
Gate 1. Copy locked (target 2026-05-08 EOD). Ewing signs off on full 3-page copy. After this, no copy edits without restart penalty. If this slips past 2026-05-11, ship date moves. Gate 2. Demo URL live + tested with 3 founders (target 2026-05-15 EOD). Demo deployed to a real subdomain, working in Safari + Chrome on iPhone + desktop, 3 outside founders have clicked through and given written feedback. If <2 testers respond by 2026-05-18, send anyway. perfect is the enemy. Gate 3. audit-quality verdict ≥ CONCERNS (target 2026-05-21 EOD). Run audit-quality on the assembled PDF + demo combo. PASS or CONCERNS = ship. REWORK = one revision cycle, hard cap 48hr, then ship anyway. FAIL = stop and re-plan.
R1 (HIGH) — Caricature style drift. Three founders, one consistent illustration style. AI tools struggle with face consistency across separate generations. Mitigation: commission a single style-locked illustrator on Fiverr day 1 as a parallel hedge ($400 budget approved up front). If AI hits by day 5, cancel. R2 (HIGH) — HR.com / Capstone / Design Precast preempt Ewing's review windows. Three live deals each have at least one this-week deliverable. Mitigation: book 3x 30-min "About Us only" calendar blocks (Mon/Wed/Fri), treat them as untouchable. Have writer + draper queue questions in a single thread to batch Ewing's review. R3 (MEDIUM) — Demo scope creep. "Interactive" can mean 5 lines of JS or a full SaaS. Mitigation: spec is locked day 1 to ONE interaction (likely a deal-fit calculator or before/after slider). Any "wouldn't it be cool if" goes to a v2 backlog file. Quarterback enforces.
The 3-page About Us with hosted page-3 demo can ship in 12 working days (target 2026-05-22) if and only if the copy outline locks on day 1, caricatures get a parallel human-illustrator hedge alongside the AI attempt, and Ewing protects three 30-minute review windows from HR.com / Design Precast / Capstone preemption. Caricature style consistency across three founders is the single longest pole. everything else is downstream of it. Demo build can run fully in parallel and go live before the PDF is finalized, giving a free week of tester soak.
I do not actually know whether Next Chapter has a vector logo source file on hand or only a flattened PNG, and I assumed a 48-hour turnaround designer is findable on Upwork without having checked Ewing's existing contractor list. I also haven't verified whether Sora is on Ewing's active subscription list versus only Midjourney. if Sora isn't subscribed, the caricature plan loses one of its two AI legs. Finally, I assumed three of Ewing's network founders will respond to a tester request inside 48 hours; that conversion rate is unverified and the swarm has no historical data for it.
Novel run, no prior maxswarm match for "ship a marketing PDF with hosted demo." Closest pattern is run #015 (dash scrubber deploy) where a parallel critical-path/demo split worked because hosting was decoupled from the rest. same shape applies here. The caricature-consistency risk rhymes with run #002 image-generation experiments where AI face-consistency was unreliable across multi-shot generations; that prior signal is why I'm hedging with a paid illustrator on day 1 instead of trusting AI alone.
Going forward, when sequencing any deliverable that depends on AI image generation across multiple consistent subjects (faces, brand assets, characters), I will commission a paid human illustrator hedge on day 1 in parallel rather than waiting for AI to fail first. the cost is bounded, the schedule slip from finding-an-illustrator-after-AI-fails is not.
Generated from 017__quarterback.md — do not edit this HTML directly.