One card per product. Three questions each: what has to be built (engineering, mostly Mattia), what it takes to run (people and process, mostly Dylan's team), and what makes it sellable (the demo, the proof, the pitch).
Everything here exists to create worlds and book calls. It is the top of the funnel and the data engine for every deal.
Five step intake that becomes the client's world, their deal sizing, and their call prep. The qualification form wearing a great costume.
Backend form + world data model on the platform; wire answers to org record. The mockup is the spec.
Nothing manual; review queue for worlds that book calls.
It is the free product itself. Demo exists; proof metric is worlds built per week.
Booking that arrives with the world attached, so the call starts at the plan, not introductions.
Calendar integration (or Cal.com embed) + world summary auto attached to the booking.
A person on the calls. Call script from the world template.
"We arrive knowing your numbers." Conversion rate call to client is the proof.
The cartoon diorama: their building, their logo, their avatar, their hero product on the podium.
Template renderer fed by world data + uploaded assets; avatar generation with approval gate.
Asset collection per client (logo, product cutouts, photo approval).
The wow moment. Aurora demo with real logo and GlowDevice already exists; show it on the call.
Growth, Studio, Launch, Partner all sell the same spine: Know, Brand, Make, Push, Catch. These are the functions that spine is made of.
Every paid relationship starts with the foundation everything else points at.
Template family + per client theming; hosting pipeline. Storefront exists in the platform repo.
Designer hours per build; content gathering checklist.
Fixed price line every client understands. Before/after site screenshots as proof.
The engine studies how the brand already speaks and locks every output to it.
Voice engine exists (voice.py). Needs: ingestion connectors (site, socials, drive) + a voice approval screen.
One approval session with the client; periodic re tune.
Side by side: generic AI copy vs their voice. Instantly legible on a call.
Produced, approved, scheduled, posted. The visible weekly heartbeat.
Engine + scheduler exist. Needs: approval queue UI, publisher hardening, official APIs where available (honest note: browser automation sits against some platforms' terms; bridge only).
Editorial calendar per client; approver named per org; compliance queue for medical.
The roadmap view in the world demo. Proof: pieces shipped per month, on schedule.
After every publish, the system reads back what actually went live and confirms it matches the plan.
Verification pass in the publisher layer; Playwright makes read back nearly free. Log result to the org.
Exception queue only; humans see failures, not successes.
One line that closes: "the machine checks its work; a human checks the machine."
Spend, return, next move, in plain numbers, monthly. Bad weeks included.
Metrics pipeline from platforms + report generator into the world.
Account lead writes the "next move" line monthly; numbers are automatic.
Demo exists (scoreboard section + world stats). Honesty is the differentiator.
The chat presence that knows the world, learns as you talk, and hands to real humans with full context.
Claude powered agent scoped to the org's data; memory per world; escalation into team inbox. The chat mockup is the interface spec.
Team answers escalations same day; guardrails per compliance profile.
Live on every screen of the demo already. "Someone to talk to" is a tier feature from day one.
Each installs into a client's business as its own measurable revenue line. These are the Deal Desk add-ons made concrete.
Leverage targets: the short list of people whose attention multiplies the client's business, hit with precise personal campaigns.
Leverage target list on the client profile; custom audience push to ad platforms; personalized landing pages from world data.
Target research per client (sales team); creative sign off per gesture.
Tell the billboard story, then show their own target list. Aurora's 20 clinic owners is the worked example.
Every enquiry answered in minutes, booked where possible, followed up on a schedule.
Inbox connectors (DMs, email, forms) + reply guardrails + booking hook. Human in loop for regulated topics.
Escalation rota; reply templates approved per client.
Response time before vs after. Minutes vs days is a one slide close.
The client's own list worked back to life. Aurora's 11,000 Mailchimp companies is the launch case.
Mailchimp connector (official API, first integration to ship), dedupe + live/possible/dead segmentation, campaign templates. Spam Act consent check is part of the build, not an afterthought.
Segment review with client; send cadence approvals.
"Revenue from names you already own." Reactivated revenue as its own report line.
Every completed sale asks for the review; referrals tracked and rewarded instead of hoped for.
Post sale triggers, review link routing, referral codes + attribution.
Reward fulfilment rules per client.
Review count and star trend on the scoreboard.
Designed repeat purchase: enrolment, reminders, redemptions, and the numbers proving it pays for itself.
Points/visits model, client branded comms, redemption tracking.
Program rules set with client; staff briefing at the client's business.
Visit two rate before vs after. Already scoped in Aurora's next actions.
Content moved and made with prompts constructed by the Dyli team: Canva autofill for static, edit briefs for video.
Canva Connect API: brand templates + autofill from world data. Video: no public CapCut API, so the build is the edit brief generator (cut list, captions, hooks in client voice) + templated assembly where possible.
Design team executes briefs; editor runs CapCut from the brief.
"A design team on your brand" add on; turnaround time is the proof.
The Deal Desk pattern sold outward: buyer profiles, financial reads, talk tracks for clients like Aurora selling to medical professionals.
Client facing version of the ops profile: their buyers, their cheat sheets, their filters.
Quarterly refresh with client sales input.
Ops mockup exists; showing Aurora their own buyer sheet is the pitch.
Clients never buy these directly, but every deal leans on them.
Size by grid, build the deal, legal check, sign, register. Every deal different, every framework the same.
Deal record + schedule generator; e signature integration; write once contract storage.
Lawyer engaged (pending, on Dylan); real price bands entered (pending, on Dylan).
Transparency is the pitch: "every number on the table before you sign."
Per org isolation, encrypted credentials, role based access, audit logs, revocable connections, deletion on exit.
Architecture exists in the platform (RLS, encrypted sessions, per org keys). Needs: independent security review before first real data import.
Access reviews; breach response plan on paper.
The Data & Access Schedule + the FAQ answer. Trust is a feature; sell it explicitly.
Status honestly assessed: "Demoed" means a clickable mockup exists in the artifact set; "Partly real" means Mattia's platform has working foundations for it; "New build" means it exists only as design. Nothing here claims to be live in production. Companion documents: the agreements pack, the Deal Desk, and the seven interface mockups.