We use cookies to understand how you use our site and improve your experience. Cookie policy
Last updated:
The Palette context layer builds and maintains a living, structured, plain-English map of how your organization actually works. This is your company's context: who's here, what they're working on, how decisions get made, what's in focus right now.
Your org chart shows reporting lines. Your wiki describes how things should work. Neither tells you what's actually happening. Your context does. This is the core of how the context layer works.
It covers three scopes: organization, team, and person. Each builds on the one above.
Everything about how your company works at the highest level. The context layer infers this from your connected tools and confirms it with your team.
ORGANIZATION
├── Company ..................... who you are, stage, size, location
│ ├── Name
│ ├── Stage .................. funding round, maturity
│ ├── Headcount
│ ├── Location
│ └── Business model
├── Product .................... what you build and who it's for
│ ├── Description
│ ├── Key features
│ ├── Tech stack
│ └── Target users
├── Strategy ................... where you're going, updates as signals shift
│ ├── Current priorities
│ ├── Key initiatives
│ └── Go-to-market phase
├── How we work ................ the stuff that's never written down
│ ├── Rituals ................ standups, demos, retros, planning
│ ├── Communication norms .... Slack-first? Async? Meeting-heavy?
│ ├── Decision making ........ who decides what, and how
│ └── Tool landscape ......... what tools the team runs on
├── Focus ...................... what matters this week
│ ├── Priorities
│ └── Direction
├── People ..................... who's here
│ └── Directory .............. name, role, email
└── Teams ...................... what teams exist and what they own
├── List
└── Workstreams
One per team. Each inherits the org-level context and adds team-specific detail.
TEAM (one per team, inherits org context)
├── Team ....................... the basics
│ ├── Name
│ ├── Purpose
│ ├── Members
│ └── Lead
├── Projects ................... what the team is building
│ ├── Tracked ................ explicitly managed in Linear or similar
│ └── Discovered ............. noticed from signals, not formally tracked yet
├── Rituals .................... standups, retros, planning cadence
├── Dynamics ................... how the team actually collaborates
│ ├── Collaboration style .... who works closely with whom
│ └── Dependencies ........... which other teams they depend on
├── Working patterns ........... learned from signals, not self-reporting
│ ├── Velocity ............... how fast the team ships
│ ├── Communication .......... where and how they talk
│ └── Async vs sync .......... do they lean into meetings or messages
└── Focus ...................... what the team cares about right now
├── Priorities
└── Active workstreams
One per person. Inherits org and team context, adds individual detail.
PERSON (one per person, inherits org + team context)
├── Person ..................... the basics
│ ├── Name
│ ├── Role
│ └── Teams
├── Working style .............. learned from behavior, not a questionnaire
│ ├── Communication .......... Slack or docs? DMs or channels?
│ ├── Async vs sync .......... meetings or messages
│ └── Pace ................... fast iteration or deep focus
├── Focus ...................... what they're working on right now
│ ├── Active projects
│ ├── Priorities
│ └── Blockers
├── Strengths .................. skills, expertise, contributions
└── Relationships .............. who they work with most
├── Frequent collaborators
└── Cross-team connections
Three inputs keep your context fresh:
Every statement carries a confidence score and a source. "Team uses 2-week sprints" (0.88, from Linear cycle data) means trust it. "Hiring plan: 3 roles open" (0.62, from a Notion doc last updated 6 weeks ago) means it probably needs confirmation.
The context layer doesn't pretend to be certain when it's not. It tells you what it knows, how it knows it, and how confident it is.
Your wiki says what you wish were true. Your context reflects what's actually happening, and tells you when it's not sure.
It's plain English. A new teammate can read it in a few minutes and understand how the company works. It's also structured for AI. Any AI tool connected via MCP reads your context automatically, starting every session with real organizational context instead of a blank slate.