Context Graph

Written By Manos Mastorakis

Last updated About 18 hours ago

The Context Graph is the foundation that makes Routine's AI genuinely useful. It's the connected network of objects, relations, and integrations that gives the AI assistant, agents, and automations a deep understanding of your work — who you work with, what you're working on, and how everything relates.

✍️ Functionality

The Context Graph:

  • Connects all objects in your workspace (tasks, events, pages, contacts, projects, custom databases) through relations and references

  • Enriches connections through external integrations (CRM, email, GitHub, Slack, etc.)

  • Gives the AI assistant and automations full context to act accurately and autonomously

  • Updates automatically as new objects are created, edited, or linked

  • Allows AI to traverse relations to find relevant information without being explicitly told where to look

  • Makes every integration a node in the graph — bidirectionally synced and linked to related objects

⁉️ Why the Context Graph matters

Most AI tools operate on isolated data — they see what you paste or describe, nothing more. Routine's Context Graph:

  • Gives AI the full picture of your work without you having to explain it

  • Connects a customer in your CRM to their emails, meetings, support tickets, and related tasks automatically

  • Lets the AI make intelligent decisions based on real relationships between objects

  • Means updating one object (e.g. a strategy document) automatically updates the context for every AI agent and automation that references it

  • Enables truly autonomous AI action — the AI knows what's relevant because the graph tells it

The richer your graph, the more powerfully the AI can act on your behalf.

🎙️ How the Context Graph works

Building the graph automatically:

The Context Graph builds itself as you work:

  • Every task, event, page, contact, and project you create becomes a node

  • Relations between objects (e.g. a task linked to a project, a contact linked to an event) become edges

  • Connected integrations (Gmail, Slack, GitHub, CRM tools) sync bidirectionally and add their data as nodes

  • The AI identifies relationships between new data and existing nodes automatically

Enriching the graph through integrations:

  1. Go to Settings → Integrations

  2. Connect your external services (Gmail, GitHub, Slack, CRM, etc.)

  3. Data from those services syncs into Routine as database objects

  4. The AI automatically links related objects across integrations:

    • A customer in your CRM linked to their email thread

    • A bug in your dev database linked to the affected customer

    • A meeting linked to the contact who attended

How AI uses the graph:

When you give the AI an instruction, it:

  1. Identifies the relevant nodes in the graph (e.g. "Dropbox" → finds the customer record)

  2. Traverses connected edges to gather full context (customer → emails → meetings → related bugs)

  3. Acts on the right objects with full awareness of their relationships

  4. Creates new nodes and edges as needed (e.g. links a new bug to the impacted customer automatically)

Keeping the graph up to date:

  • Update a shared document (strategy, ICP, preferences) and every agent and automation that references it updates automatically

  • Add a new integration and its data is immediately available to the AI across all agents and automations

  • Create new relations between types in Settings → Types to enrich the graph further

🗣️ Use cases

The Context Graph is especially powerful for:

  • CRM automation — link customers to their emails, meetings, and tasks so the AI always has full account context

  • Bug tracking — automatically connect bugs to affected customers, assignees, and related GitHub issues

  • Meeting intelligence — link meeting notes to contacts, projects, and follow-up tasks automatically

  • Strategy alignment — update a single strategy document and have every AI agent reflect the change instantly

  • Sales workflows — connect pipeline stages, communications, and tasks so AI can recommend next steps with full context

  • Team coordination — give AI agents access to the full graph so they can act across tasks, people, and projects without gaps

  • Autonomous AI — the richer the graph, the more accurately AI can act without asking for clarification

The Context Graph is what separates Routine's AI from a chatbot — it acts on real, connected data, not just what you tell it in the moment.