Handoff
Instant project context for new team members
Thesis
Onboarding new team members to projects is painful. Document synthesis can generate context from scattered materials instantly.
The Problem
Onboarding someone to an active project is chaos: Where's the brief? Which Slack channel? Who's the client contact? What decisions were already made? The new person asks questions for a week. The existing team gets interrupted constantly. All this information exists somewhere. It's just scattered across 47 places and nobody has time to compile it.
Implementation Approaches
Auto-Generated Context Doc (Recommended)
AI compiles project context from all sources into a single doc
Implementation
- →Pull from: brief, contract, Slack, email, tickets, drive
- →Generate: project overview, key people, decisions made, current status
- →Include: links to important docs and threads
- →Update automatically as project evolves
- →Role-specific views (designer vs developer vs PM)
Pros
- +Massive time saver, measured in days not hours
- +Always current, not a stale wiki page
- +Flywheel: learns what new team members actually need
- +Natural extension of Guildry's project model
Cons
- −Requires broad data access
- −Quality depends on information existing in connected tools
- −Complex to get right, lots of edge cases
Q&A Bot
Chatbot that answers questions about the project
Implementation
- →Index all project materials
- →New team member asks questions in natural language
- →Bot answers with sources and links
- →Learns what questions are common
Pros
- +Interactive, surfaces what's actually needed
- +Doesn't require perfect pre-generation
- +Can handle edge case questions
- +Usage data shows gaps in documentation
Cons
- −Requires new person to know what to ask
- −Doesn't provide proactive overview
- −RAG quality can be inconsistent
Guided Onboarding Flow
Step-by-step onboarding checklist with AI-populated content
Implementation
- →Standard onboarding checklist template
- →AI populates with project-specific info
- →New person works through steps
- →Tracks completion, flags gaps
Pros
- +Structured, ensures nothing missed
- +Simpler to build than full synthesis
- +Clear progress visibility
- +Can work with partial information
Cons
- −Less comprehensive than full context doc
- −Template might not fit all project types
- −Still some manual work required
Validation Plan
Hypothesis to Test
Teams will pay $49/mo per project to onboard new members in hours instead of days
Validation Phases
Manual Context Doc
1 week- •Pick 2 active projects with recent additions
- •Manually compile context doc using Claude
- •Have new team member review: 'What's missing?'
- •Time: how long to create, how much time saved
Automated Generation
2 weeks- •Build multi-source context generation
- •Test on 3 different project types
- •Compare: auto-generated vs manual quality
- •Iterate on what to include/exclude
Live Onboarding Test
2 weeks- •Use tool for actual new team member onboarding
- •Measure: time to first contribution, questions asked
- •Gather feedback from both new person and team
- •Validate $49/mo per project pricing
Kill Criteria
Stop and move on if any of these become true:
- ✕Generated context docs are too inaccurate or incomplete
- ✕Teams don't have enough in connected tools to work with
- ✕Problem isn't painful enough (onboarding is 'fine')
- ✕Per-project pricing doesn't work for teams with many projects