Kaleido Life shenAI MVP — Implementation Briefing v2

From: Jonathan Gudger, Chief Architect To: Craig Du Bruyn, CEO Date: March 9, 2026
8/8 Decisions Confirmed — Implementation Scope Locked

All 8 client decisions were confirmed on the same day. 6 matched our recommendations exactly; 2 diverged with clear, actionable direction. Zero ambiguity remains. The implementation scope is locked and Phase 1 can begin immediately.

1 Progress Dashboard

Decision outcomes from v1 input — all 8 resolved in a single session.

6
Resolved — Matched Recommendation
2
Diverged — Clear Direction
0
Ambiguous — Needs Follow-up

Decision Outcome Distribution

Loading diagram...
What this means: No further input is required. The v1 briefing asked 8 questions; all 8 are answered. This document converts those answers into an implementation-ready build spec. Phase 1 can start immediately.

2 Confirmed Decisions

All 8 decisions presented as implementation facts. Two divergences marked with direction-change badges.

Resolved
D1: Social Auth Provider Priority
Email First, then Apple + Google
Phase 2 auth starts with email-based flow. Apple/Google OAuth layered on as enhancement after core flow is working. Simplest path to demo-ready auth.
Resolved
D2: EligibilityQuiz Screen Fate
Keep + Restyle to Match New Theme
4-question quiz remains in the user flow (life insurance, income, health, cash preference). Restyled with teal/aqua theme. Data feeds the Life Liquidity Score calculation.
Resolved
D3: AI Avatar Integration Depth
Scripted Responses at Key Moments
Avatar component with personalized messages at intent screen and score reveal. Personalized by name + intent selection. No LLM backend needed — lightweight, reliable, on-brand.
Resolved
D4: Processing Screen Handling
Merge into Biometric Scan
Separate Processing screen removed. "LIFE CASH LOADING.." header shown during active scan with live readouts (HR, BP) as overlay. One fewer screen transition. Cleaner UX.
Diverged — Clear
Direction Changed
D5: Splash Animation Format
Optimized WebM/MP4 (not CSS/JS)
We recommended CSS/JS animation for smallest footprint. Craig chose video for visual fidelity — the original MP4 has specific effects that matter to brand identity. Compress 4.4 MB MP4 to <500 KB WebM with fallback MP4. Autoplay, muted, 2-3s, auto-advance.
Resolved
D6: Background Implementation
Static Image (WebP) + CSS Particles
Convert POC background PNG (374 KB) to optimized WebP. Remove WebGL2 kaleidoscope entirely. CSS particle animation overlay for subtle movement. Lighter, faster, matches mockup exactly.
Diverged — Clear
Direction Changed
D7: DNA Behavior Integration Timeline
Full Defer Until Partnership Terms
We recommended starting API research now. Craig chose full deferral — no DNA Behavior work until business terms are locked. This simplifies the roadmap: focus entirely on visual MVP + auth + ShenAI. DNA moves from "research in parallel" to "wait for green light."
Resolved
D8: Assuretrac Demo Scope
Theme + Auth + Intent Redesign
Assuretrac demo includes: teal/aqua theme overhaul, email auth, icon-grid intent framing, explainer screen, enhanced biometric scan UX, teal score reveal. Does NOT include social OAuth, DNA, Didit, Plaid, or Terra.

3 Implementation Roadmap

Three phases with clear gates. Phase 1 is immediately actionable.

Phase 1
Visual MVP
10 tasks · Assuretrac-ready
Ready to Build

Theme overhaul, screen redesigns, UX enhancements, splash video, background optimization. All scope locked by D1-D8.

Phase 2
Authentication
3 tasks · Post visual approval
Queued

Email auth first (D1), then Apple + Google OAuth. Gate: Phase 1 complete + Craig visual approval.

Phase 3
Integrations
4 tasks · Post-demo + terms
Deferred

Didit + Plaid when ready. DNA Behavior fully deferred per D7. Terra awaiting SDK access.

Phase 1: Implementation Gantt

Loading diagram...

4 Phase 1 — Visual MVP Tasks

10 implementation tasks with source decisions, priority levels, and scope descriptions.

1 Theme Overhaul: Purple/Gold to Teal/Aqua
Highest CP-1 + D6
Extract design tokens from mockups. Replace Tailwind palette, main.css theme, and all screen color references. Foundation for every other task.
2 IR Redirect Fix
Highest CP-7 (Bug)
Fix kaleido-life.chem.dev routing so root path serves IR Copilot, not shenAI app. Investor confidence issue — someone already saw the dev environment.
3 Intent Screen: Icon Grid + Avatar
High CP-2 + D3
Redesign IntentFraming.ts with 2x2 icon card grid (Dream Vacation, Pay Off Debt, Down Payment, Get Married). AI avatar tooltip: "This helps me personalize your Life Liquidity Score."
4 "What Happens Next" Explainer
High CP-3
Replace FountainOfTruth orb animation with 3-step explainer: Quick Scan, Life Liquidity Score, Cash Unlock Roadmap. Tagline + dot progress + "Start Biometric Scan" CTA.
5 Biometric Scan UX + "LIFE CASH LOADING.."
High CP-4 + D4
"LIFE CASH LOADING.." branding during active scan. Live readout overlays (BPM, BP progress bar). Star rating for scan quality. Teal-themed face overlay replacing gold.
6 Processing Screen Removal
High D4
Remove separate Processing screen entirely. Merge its functionality into BiometricScan active phase. Score calculation happens inline during/after scan completion.
7 Score Reveal Retheme + Avatar Chat Bubble
High CP-5 + D3
Teal glow rings on circular gauge, hexagon logo at center. AI avatar chat bubble above score. Consolidate to 3 CTAs: Wellness Matrix, View Offer, Start Over.
8 Splash Animation (WebM/MP4)
Medium CP-6 + D5
Compress 4.4 MB MP4 to <500 KB WebM with fallback MP4. Autoplay, muted, no controls. 2-3s duration, auto-advance to Landing. Per D5 divergence: video fidelity over CSS/JS.
9 Quiz Screen Restyle
Medium D2
Keep all 4 questions (life insurance, income, health, cash preference). Apply teal/aqua theme, frosted glass card styling. Data flow to Life Liquidity Score unchanged.
10 Background: WebP + CSS Particles
Medium D6
Convert POC background PNG to optimized WebP. Remove WebGL2 kaleidoscope component. Add CSS particle animation overlay for subtle movement. Significant performance improvement.

Confirmed User Flow — Post-Implementation

Loading diagram...

5 Deployment Strategy

Parallel folder approach — safe iteration with instant rollback.

Parallel Folder Build

The v2 implementation will be built in a parallel folder alongside the existing v1 codebase. This means the current working demo remains live and untouched throughout the entire Phase 1 build. If any issue arises during development, or if a stakeholder demo is needed mid-build, the v1 app is always available at its current URL.

Once Phase 1 is complete and Craig approves the visual direction, we swap the deployment target to the v2 folder. If at any point we need to revert, it is a single-line configuration change back to v1.

shenai-app/  ← current live deployment (v1, unchanged)
shenai-app-v2/  ← new build target (Phase 1 implementation)

Go-live: Caddy proxy swap from shenai-app/ shenai-app-v2/
Rollback: Caddy proxy swap from shenai-app-v2/ shenai-app/
Why parallel folders? Craig's original splash animation and mockups represent a significant visual departure from v1. Building in parallel means we can iterate freely without risk to the live demo. Both versions can even run simultaneously on different subdomains for A/B comparison if needed.

6 API Architecture Update

Scope narrowed per confirmed decisions. ShenAI only for Assuretrac demo; all other partners deferred.

Assuretrac Demo Architecture — Confirmed Scope

Loading diagram...

In Scope for Assuretrac Demo

  • ShenAI biometric scan (60s face scan)
  • Life Liquidity Score calculation
  • Email authentication
  • Full visual theme overhaul
  • Redesigned user flow

Deferred — Post-Demo / Post-Terms

  • ○ Apple / Google OAuth (Phase 2)
  • ○ Didit KYC (when needed)
  • ○ Plaid Financial (when needed)
  • ○ DNA Behavior (deferred per D7 — awaiting terms)
  • ○ Terra Wearables (awaiting SDK)
Impact of D7 (DNA Behavior full defer): The 5-partner API flow diagram from v1 is aspirational architecture, not current sprint scope. For the Assuretrac demo, the data pipeline is streamlined: User → Email Auth → Quiz → ShenAI Scan → Score. This reduces integration complexity significantly and lets us focus entirely on delivering a polished visual experience.