Someday/Maybe Projects

Strategic project ideas staged for future execution.

Strategic Backlog

Ideas

Raw captures

12PENDING

"This Moment Is Forever" — SBF Brand Slogan

Engine 1 — Weddings
  • Born from a conversation about impermanence — 'in this moment, this moment is forever'
  • Potential tagline for Stephan Boeker Films
  • Captures the philosophy: wedding films preserve what's fleeting
  • Evaluate against current brand positioning
  • Test with existing clients / social media

Content Creation Machine

Growth Engine
  • Agentic system: research + content creation
  • Text content + AI avatar video (Stephan talking)
  • Wedding side: Instagram/TikTok for Stephan Boeker Films
  • YouTube strategy for weddings
  • PST commercial side: content strategy

Post-Production Story Agent

Editorial Ops
  • Open OpenClaw to human editors (Jana, others)
  • Each wedding gets its own Discord channel
  • Channel pre-loaded with pre-sales intelligence, client dossier, narrative context
  • Editors use DaVinci Resolve IntelliScript + channel context
  • Limited human user accounts (e.g., Zette sees Sloane but can't modify system files)

Resellable Turnkey System

Productization
  • OpenClaw Wedding Filmmakers turnkey system - spawn off businesses
  • Same for commercial filmmaking
  • Revenue opportunity from productizing the architecture

Nonprofit OpenClaw Product

Impact
  • Prototype OpenClaw system for nonprofits
  • Service/product for agentic storytelling

Self-Improving Research Loop (Scholar Agent)

Agentic Architecture
  • Dedicated research agent ('Scholar') that proactively learns and broadcasts learnings to all agents
  • Runs on cron — hourly research tracks per domain (wedding enrollment, commercial production, storytelling)
  • Auto-broadcast: when Scholar learns something, it routes to the specific agent who needs it — no human routing
  • Shared LEARNINGS.md at workspace level — all agents read before writing
  • Cross-agent memory updates: Sloane's OODA learning automatically available to Koa's writer
  • Inspired by Atlas/Nate's production implementation — 200 lines of Python for the broadcast system
  • Prerequisite: reliable email chain protocol must be working first
  • Source: OpenClaw community post by Alex Finn / Atlas (March 2026)

Financial/Trading Advisor & Executor

Wealth Ops
  • AI agent that reads crypto market signals and understands market dynamics
  • Portfolio dropped from $55K → $26K (late 2025 → early 2026) — no bandwidth to watch
  • Goal: automated signal analysis, strategic trade execution, risk management
  • Stop-loss automation, portfolio rebalancing, trend detection
  • Prevent bag-holding through disciplined AI-driven exits

Design Agent — Creative Director

Creative Ops
  • Dedicated agent for all visual design: PDF formatting, layout, typography
  • Generative AI for graphics, mood boards, brand assets
  • Problem: Amergin produces great narrative content but poor visual output
  • Owns: creative briefs, pitch decks, proposal PDFs, client-facing documents
  • Could handle: social media graphics, thumbnails, brand consistency
  • Separates craft (Amergin = story) from presentation (Design Agent = look)

CFO — Chief Financial Officer

Finance
  • Full handle on the books: bookkeeping, accounting, tax strategy
  • Tax expert — keeps PST on the straight and narrow year-round
  • QBO integration already exists (Realm 193514716593489) — foundation is there
  • MUST run on local model — financial data cannot leak to cloud frontier models
  • Requires M5 hardware with enough memory for capable local model (Qwen 3.5+)
  • Sovereignty requirement: all financial reasoning stays on-device

Engine 2 Cold Outreach Tool

Engine 2 — Commercial
  • Inspired by Patrick Moreau's cold outreach tool (Loom walkthrough Apr 8 2026)
  • Input: prospect name + website → Perplexity research → insight/offer suggestions
  • Human selects insight + offer → generates cold email, follow-up, and LinkedIn message
  • Key formula: intent over demographics, offer + insight, science-backed storytelling gaps
  • Lightweight mode of the Prospect Intel pipeline (~$2-5 vs $15-30 for full report)
  • Full brief: briefs/ad-intel-tool-project-brief.md

Engine 2 Prospect Intelligence Tool

Engine 2 — Commercial
  • Inspired by Patrick Moreau's 'Ad Intel Spy Tool' — automated prospect research before discovery calls
  • Estimates ad spend, audits creative, runs MUSE storytelling diagnostic, identifies top 3 opportunities
  • Revenue-as-a-service offer modeling — tie fees to measurable CPA/ROAS improvement
  • Agency/in-house analysis + outreach email draft
  • Koa orchestrates pipeline; data agents on gpt-5.4-mini, synthesis on gpt-5.2/sonnet
  • Estimated ~$15-30/report vs Patrick's $50 with cost-optimized models
  • Full brief: briefs/ad-intel-tool-project-brief.md

Real-Time Interview Intelligence

Creative Ops
  • AI listens to live on-camera interviews and provides real-time coaching on a second screen
  • Suggests follow-up questions, flags dropped threads, identifies areas to push deeper
  • Architecture: BlackHole audio capture → Deepgram streaming transcription → Claude API analysis (every 30-45s) → browser overlay
  • Pre-interview brief (guest bio, topics) loaded into system prompt for gap detection
  • Display: color-coded suggestions — green (new thread), yellow (dropped thread), red (fact to verify)
  • MVP: ~300 lines Python + 50 lines HTML/JS — fully hands-off during interview
  • Cost: ~$5-10 per hour-long interview (transcription + LLM calls)
  • Nothing on market does real-time question suggestions — novel capability
  • PDF brief: briefs/real-time-interview-ai-architecture.pdf

In Consideration

Being evaluated

0PENDING

No projects in this stage yet.

In Production

Actively being built

1RUNNING

OpenClaw User Access & Permissions

Platform
  • Role-based access control for human team members (EA, editors, contractors)
  • First use case: Zette gets read/chat access to Sloane — no system config, no file edits
  • Tiered permissions: chat-only, read workspace, edit workspace, full admin
  • Separate from agent-to-agent communication — this is human-to-agent access
  • Protects system integrity while letting team members leverage agent intelligence
  • Could extend to clients (limited portal access) down the road

Alpha

Functional but rough — internal testing

0RUNNING

No projects in this stage yet.

Beta

Stable — limited real-world use

0RUNNING

No projects in this stage yet.

Operational

Vetted, tested, running reliably

0OK

No projects in this stage yet.