← Back to feature backlog
🛠 Build Sprints — path of least resistance
Beyond the 🔴 Now / 🟡 Next / 🟢 Later priority view, this is the technical sequencing view. Features grouped into 12 sprints by what they share underneath: auth, data model, integrations, UI patterns. Build the right things at the right time and Otto's pace stays high. Skip the order and he'll re-do work three times.
Dev sequencing
12 sprints · ~9 months total
For Otto
The principle
Build the foundation. Then the data model. Then the integrations. Then the polish.
Each sprint is a coherent unit of work where every feature touches related code. Auth changes once, schema migrations happen once per area, integrations get wired once. Sprints 1–3 are FOUNDATIONS — without them, every later sprint costs more. Sprints 4–9 are CORE FEATURES — biggest user value. Sprints 10–12 are EXPANSION — high upside but assumes the core is solid.
~6 mo
With UI rebuild parallel
3
Foundation sprints (must come first)
Visual timeline
How sprints stack across ~9 months (sequential view)
W1
W2
W3
W4
W5
W6
W7
W8
W9
W10
W11
W12
W13
W14
W15
W16
W17
W18
W19
W20
S1
Foundation · Auth + Onboarding
S2
Data Model · Roles + Permissions
S3
Foundation · Brand Profile + Credits
S4
Project Flow Rebuild
S5
Cloud Storage as Drive
S6
Notifications Framework
S7
AI Creative Studio
S8
AI Chatbot + Script Funnel
S9
Dashboard + News + SEO
PARALLEL
UI Rebuild · runs alongside everything
Sprints 10–12 (Course/Live + Referrals + AI Avatar + Internal CRM + Tripod Kit + Own API) extend past W20. Most of those can be parallelized with junior help once the foundation is solid.
SPRINT 1
🔐 Foundation · Auth + Onboarding
~3 weeks
Why these belong together: Every single feature in the platform assumes a working signup → password → first-login flow. Build this wrong and every later sprint inherits the bugs. All of these touch the same auth state machine, the same Stripe webhook handler, the same email-queue service, the same wizard-state model. Build them as ONE coherent unit, not as 6 separate tickets.
Blocks: all other sprints. Without working signup, nothing else gets used.
Mockup: first-login.html
SPRINT 2
🎭 Data Model · Roles + Permissions
~2 weeks
Why these belong together: The user table needs new role columns (super_admin / manager / content_creator / client / assistant). Every later feature that's "manager-only" or "assistant-can-approve" depends on this layer existing. Build the role system once, then every authorization check downstream is one line of code.
Blocks: S6 notifications, S9 dashboard, S11 backend tools.
SPRINT 3
🎨 Foundation · Brand Profile + Credits Infrastructure
~4 weeks
Why these belong together: These are the data foundations that EVERYTHING else reads. Brand Profile feeds Studio + Avatar + Script Generator. Credits feed Studio + Avatar + every AI feature. Get the schemas right NOW or you re-migrate 5 times later. Credit infrastructure also includes the Stripe top-up flow (F-180–F-186) so micro-transactions work from day one.
SPRINT 4
📋 Project Flow Rebuild
~4 weeks
Why these belong together: Per-platform captions, custom thumbnails, edit-only mode, external share, project tracker, easier downloads — all touch the project schema and the project-detail UI. Migrate the schema once, redesign the project page once. Most are 🔴 Now bugs that block real client use cases (Sean Konja flow, the Darcy download asks).
Blocks: S7 Studio output destinations · S8 Chatbot project filing · S11 Bulk Upload.
Mockup: approval-flow.html
SPRINT 5
☁ Cloud Storage as Drive Replacement
~2 weeks
Why these belong together: Folder tree, file rename, soft delete, external share with tokenized links, capacity gauge, upsell tier. All touch the Cloud Storage data model + S3 backing. Standalone unit — depends on S3 infra (already wired) but doesn't block anything else for v1. Run in parallel with S4 if you have help.
SPRINT 6
📨 Notifications Framework + SMS
~3 weeks
Why these belong together: Twilio + SendGrid + opt-in framework + drip queue. Build the FRAMEWORK once (notification service with template + audience filter + opt-in check + delay queue). Then add specific notifications as instances. Adoption flow (S1 already kicked off uses this), inactive-30-day, pending-approval reminder, celebration emails — all just config on top of the framework.
SPRINT 7
✨ AI Creative Studio (client + backend)
~6 weeks
Why these belong together: The biggest single feature on the roadmap. Both the client-facing chat surface AND the backend autonomous queue depend on the same AI engine, the same per-client memory layer, the same prompt templates. Build them together, not separately. Heavy use of Anthropic Claude API + Higgsfield + Canva. Caches must be designed up front to control cost.
Depends on: S3 (Brand Profile + Credits).
Blocks: S8 Chatbot (handoff destination), S11 AI Avatar.
Mockup: creative-studio.html
SPRINT 8
🤖 AI Chatbot + Script Generator Funnel
~4 weeks
Why these belong together: Same AI brain (Anthropic SDK), same tool-use layer for project filing, same routing logic (AI Studio vs Human team). The Chatbot reuses the Script Generator's voice picker for "what voice do you want this in?". Teleprompter + Record reuse the script handoff. Build the funnel and the chatbot in parallel — they share the project-filing API and the Studio output destinations from S4.
SPRINT 9
📊 Dashboard + News Feed + Local Leader (SEO)
~4 weeks
Why these belong together: All widget-grid features. Build the dashboard widget framework once (grid, drag-to-reorder, persistence) and every widget below is just a component that plugs in. Counter, recap, leaderboard, roadmap, analytics, property reports, news feed, locked Local Leader — same pattern. Surfaces the value of every prior sprint.
SPRINT 10
📺 Course/Live + Referrals
~3 weeks
Why these belong together: Both are dashboard-adjacent retention layers. Course/Live reuses Mux + Pusher (already wired in S6 for Bini chat). Referrals reuses Stripe (S3) + dashboard widget framework (S9). Both standalone units once foundations are in place; safe to parallelize.
SPRINT 11
🤖 AI Avatar + Backend Staff Tools + CRM
~4 weeks
Why these belong together: Mostly independent features but all depend on the foundations being solid. AI Avatar is a Heygen API integration (~2 weeks once Brand Profile + Credits exist). Internal CRM is a build-from-scratch (~3 weeks). Staff tools are a focused unit. Group them as the "completion sprint" before tackling the long tail.
SPRINT 12
🛒 Product / Tripod Kit + Own API
~6 weeks (long tail)
Why these belong together: Both are long-running, hardware/integration-heavy projects that ship later. Product section needs e-commerce + BLE pairing for the AI Tripod. Own API replacement needs platform-by-platform application processes (months). Both can run with junior help once foundations are solid. Don't let these block earlier sprints.
Action ASAP (independent of build): Start the platform application processes for Own API NOW (FB / IG / LinkedIn / TikTok / etc — each takes months).
Mockup: product-tripod-kit
PARALLEL TRACK
🎨 UI Rebuild — runs alongside everything (F-017)
9 months ongoing
Why this is parallel: The UI rebuild can't be a single sprint — it touches every screen. Otto rebuilds each section to match the mockup aesthetic AS HE SHIPS that section. Sprint 3's Brand Profile = rebuilt to mockup. Sprint 7's Studio = rebuilt to mockup. By W20, the whole platform looks consistent. Same code time, no extra timeline.
Source of truth: the 19 mockup files in /mockups/. Use /styles/mockup.css as the design system starter.
Otto's golden rule for sequencing
If a feature depends on a foundation that isn't built yet, don't start it. You'll hit walls. Refactor 3x. Lose weeks.
If a foundation is built and a feature on top is ready, batch them. Two features that touch the same code = build them in one sprint to amortize the schema migration / API integration / UI pattern work. Save weeks.
← Back to feature backlog