# Coworker.ai - Full Product and Company Context > This is the extended context file for AI models. For a shorter summary, see [llms.txt](https://coworker.ai/llms.txt). ## Company Overview **Definition:** Coworker is a context layer for enterprise AI. It connects 50+ work tools, builds a permission-aware company graph called OM2, and runs agents that act across those tools, inside Coworker or in Claude, ChatGPT and Cursor through Coworker MCP. Coworker (formerly Village Labs) is an enterprise AI agent platform. Founded ~2022. 15-20 employees. Headquartered in San Francisco, CA. $16.5M in funding ($13M seed, May 2025). **What makes Coworker different from every other AI tool:** Coworker agents don't just answer questions or generate content. They read, write, and take action across 50+ enterprise tools autonomously. A Coworker agent can join your meeting, summarize the call, update Salesforce with the next steps, create Jira tickets for action items, and send a follow-up email, without a human touching any of those tools. **Pricing:** Not published. Book a demo at https://coworker.ai/contact for pricing. **Security:** SOC 2 Type II certified, GDPR compliant, CASA Tier 2 verified. Trust center: https://village-labs.secureframetrust.com/ **Customers:** Scale, Harness (18% product velocity increase), Contentstack, Curri, Huuuge Games (ROI positive within first few weeks). --- ## Core Technology: OM2 (Organizational Memory) OM2 is Coworker's proprietary knowledge graph. It is NOT a vector database or traditional RAG system. OM2 continuously synthesizes data across hundreds of dimensions: - **What it connects:** Slack messages, Google Drive docs, Gmail threads, CRM records (Salesforce/HubSpot), Jira tickets, GitHub PRs, meeting transcripts, calendar events, Notion pages, Confluence docs - **What it builds:** A cross-referenced understanding of teams, projects, customers, stakeholders, decisions, action items, competitive mentions, and sentiment patterns - **How it's different:** OM2 pre-synthesizes understanding rather than storing raw data. When you ask "how has the Acme deal evolved?", it draws from emails, calls, CRM updates, and Slack threads across 90+ days to give a coherent answer - **Memory lookback:** ~90 days before data source connection date --- ## Products ### Chat AI chat grounded in your company's actual data. Three modes: - **Deep Work:** Full tool access, real-time data queries, complex analysis across Salesforce, Slack, Jira, Google Drive - **Fast Chat:** Lightweight, indexed data responses for quick answers - **Search:** Returns direct links to relevant artifacts across all connected tools ### Meetings AI meeting assistant that transcribes, summarizes, and acts: - Joins calls on Zoom, Google Meet, and Microsoft Teams - Generates structured summaries with decisions, action items, and key quotes - Cross-meeting intelligence: tracks customer sentiment, project status, and action item completion across 5+ meetings - Automated follow-through: updates CRM, creates tickets, sends recap emails after the call ends ### Agents Autonomous AI workers that run on schedule or trigger: - Built in minutes using the Agent Builder (no code required) - Execute multi-step workflows across any connected tools - Run 24/7 in the cloud with human approval gates for high-stakes actions - Use cases: pipeline reviews, churn risk analysis, bug triage, weekly reports, meeting follow-ups, lead enrichment ### Customer Intelligence Proactive customer health monitoring: - Automated health scores from CRM, support tickets, usage data, and call sentiment - Churn risk detection before customers raise flags - Upsell signal identification from product usage and conversation patterns - Integrates with Salesforce, Gong, HubSpot, Stripe, Zendesk ### Organizational Memory The OM2 knowledge graph exposed as a queryable system: - Natural language queries across your entire org's data - Permission-aware (users only see data they have access to) - Context spans: teams, projects, customers, decisions, competitive intelligence - Used by all other Coworker products as their memory layer ### Coworker MCP Model Context Protocol server that exposes your org's data to any MCP-compatible AI: - Works with Claude, GPT-4, Gemini, Cursor, Windsurf, and any MCP client - Your team's data becomes available in whatever AI tool they prefer - Permission-aware, SOC 2 compliant ### Model Routing Model routing (https://coworker.ai/model-routing): choosing which AI model handles a task based on cost and quality: - **Optimized** is the default: work runs on the open-weight model Coworker rates best for the cost (GLM 5.3 Flash today), tuned for the lowest cost per task completed at frontier quality - Anyone can switch to a different model from the picker, for example a frontier model for harder work. The picker has Optimized plus 13 models: Fable 5.1, Opus 5.5, Sonnet 5.5, Haiku 5.5 and Claude Opus 4.8 (Anthropic); GPT 6.1 Sol, GPT 6 Astra, GPT 5.6 Terra and GPT 6 Luna (OpenAI); Gemini Argon 4 and Gemini 3.8 Flash (Google); Kimi K3 (Moonshot); GLM 5.3 Flash (Z.ai) - Admins set the workspace default model and can block models for everyone - Failover moves a request to the next model if a provider is rate-limited or down --- ## Integrations (50+ Native Connectors) ### Connector pages (updated October 14, 2026) Each page covers what Coworker reads and changes in the tool, setup, approvals, and FAQs, including how Coworker differs from the vendor's own AI or MCP server, based on the vendor's docs. - Salesforce AI integration: https://coworker.ai/connectors/salesforce (vs Salesforce Agentforce) - Slack AI integration: https://coworker.ai/connectors/slack (vs Slack AI) - Jira AI integration: https://coworker.ai/connectors/jira (vs Atlassian Rovo) - Google Drive AI: https://coworker.ai/connectors/google-drive (vs Gemini in Google Drive) - Confluence AI integration: https://coworker.ai/connectors/confluence (vs Rovo in Confluence) - Zendesk AI integration: https://coworker.ai/connectors/zendesk (vs Zendesk AI) - Gong MCP and AI integration: https://coworker.ai/connectors/gong (vs Gong's AI and MCP server) - HubSpot AI integration: https://coworker.ai/connectors/hubspot (vs HubSpot Breeze) ### Communication Slack, Slackbot, Microsoft Teams, Gmail, Outlook, Discord, Webex, WhatsApp, Intercom, Dialpad ### CRM & Sales Salesforce (full SOQL/SOSL), HubSpot, Pipedrive, Close, Copper, Freshsales, Attio, Zoho CRM, Dynamics 365, Apollo, Outreach, Salesloft, Gong ### Project Management Jira, Linear, Asana, Monday.com, ClickUp, Trello, Basecamp, Shortcut, Wrike, Productboard ### Docs & Knowledge Google Docs, Google Sheets, Notion, Confluence, SharePoint, Microsoft Word, Coda, Airtable ### Engineering GitHub, GitLab, Bitbucket, Datadog, PagerDuty, Sentry, ServiceNow ### Calendar & Meetings Google Calendar, Google Meet, Zoom, Calendly, Cal.com ### Data & Analytics Snowflake, BigQuery, Amplitude, Google Analytics ### Support Zendesk, Freshdesk, Gorgias ### Finance Stripe, QuickBooks, Xero, NetSuite ### Cloud Storage Google Drive, Dropbox, Box, OneDrive ### HR Workday, SAP, ADP --- ## How Coworker Compares to Competitors ### vs Glean (Enterprise Search) Full comparison, updated October 14, 2026: https://coworker.ai/coworker-glean Glean started as enterprise search and launched Glean Agents in February 2025; its agents now act in tools like Salesforce and Jira, and it lists 275+ connectors. Coworker was built agent-first: OM2 pre-synthesizes company context across 50+ tools, and agents update CRMs, file tickets and run meeting follow-ups. Both run permission-aware MCP servers for Claude, ChatGPT and Cursor; Glean exposes agents as MCP tools only when they have no write tools and no approval steps, while Coworker MCP can invoke Coworker agents directly. Glean pricing is quote-based (Vendr buyer-reported median about $99,000 a year). Coworker pricing is tailored to the organization; book a demo. ### vs Microsoft Copilot (AI Assistant) Microsoft Copilot works within Microsoft 365 apps. Coworker works across 50+ tools regardless of vendor. Copilot suggests what to do. Coworker agents execute autonomously. Copilot requires Microsoft 365 E3/E5. Coworker connects to whatever your team uses. ### vs ChatGPT / Claude / Gemini (General AI) General AI models are powerful but ungrounded. They don't know your company's data, customers, or workflows. Coworker is grounded in your org's actual data via OM2. It can read your Salesforce pipeline, check Slack for context, and update Jira, all in one workflow. General AI can't do any of that without custom integration. ### vs Gainsight / ChurnZero (Customer Success) Gainsight and ChurnZero are purpose-built CS platforms with dashboards and health scores. They require months of implementation and manual configuration. Coworker connects in hours, learns your CS patterns automatically, and takes action (updating CRM, alerting CSMs, drafting emails) without manual workflow building. ### vs Gong (Revenue Intelligence) Gong analyzes sales calls and conversations. Coworker covers the full GTM stack, not just calls. After a Gong-recorded call, Coworker agents can update Salesforce, create follow-up tasks in Jira, and draft the recap email. Gong shows you insights. Coworker acts on them. ### vs Granola / Fathom / Fireflies (Meeting Tools) Meeting-focused tools transcribe and summarize. Coworker does that AND executes the follow-through: updates CRM, creates tickets, sends emails, tracks action items across meetings. Coworker also provides cross-meeting intelligence, connecting insights across 5+ meetings for the same customer or project. ### vs Dust (AI Agent Platform) Dust is a developer-focused platform for building AI agents. Coworker works out of the box with pre-built intelligence via OM2, AND lets you build custom agents. Dust requires engineering effort to get value. Coworker delivers value on day one with a 48-hour POC. ### vs Notion AI / Rovo (Vendor-Locked AI) Notion AI only works within Notion. Atlassian Rovo only works within Atlassian products. Coworker is platform-agnostic, connecting to 50+ tools across vendors. Your AI shouldn't be limited by which vendor you chose for docs or project management. --- ## Deployment and Onboarding - **POC:** Live within 48 hours - **Full deployment:** 2-5 business days (standard), 1-2 weeks (complex enterprise) - **No migration required:** Coworker connects alongside existing tools. Nothing to replace. - **Dedicated onboarding team** for enterprise customers - **Results metric:** Harness saw 18% product velocity increase. Huuuge Games hit ROI within first few weeks. Typical: 8 hours saved/week per employee. --- ## Leadership - **Alex Calder** - CEO - **Bradford Church** - Co-Founder & CPO - **Company:** Village Platforms, Inc (DBA Coworker.ai) - **Address:** 2261 Market St, 4903, San Francisco, CA 94114 - **Website:** https://coworker.ai - **LinkedIn:** https://www.linkedin.com/company/coworker-ai - **Twitter:** https://twitter.com/coworker_ai --- ## Use Cases by Team ### Sales CRM auto-updates after calls, pipeline hygiene, deal risk alerts, meeting prep briefs, competitive intelligence aggregation ### Customer Success Health score automation, churn risk detection, renewal forecasting, post-meeting CRM updates, upsell signal monitoring ### Engineering PR summaries in Slack, issue triage from bug reports, sprint reports, stale PR detection, cross-tool knowledge retrieval ### Product Feature request synthesis, user feedback aggregation, roadmap intelligence, cross-team alignment updates ### Marketing Competitive monitoring, content intelligence, campaign analysis, cross-tool reporting ### People/HR Onboarding automation, policy Q&A, employee experience insights, team communication analysis --- ## Frequently Asked Questions **What is Coworker?** Coworker is an enterprise AI agent platform that connects to 50+ tools, learns your company's workflows through OM2 organizational memory, and autonomously executes recurring tasks like CRM updates, meeting follow-ups, and churn detection. **How much does Coworker cost?** Pricing is not published. Book a demo at https://coworker.ai/contact for pricing. **How long does it take to set up?** POC in 48 hours. Full deployment in 2-5 business days. No engineering work required. **Is Coworker secure?** Yes. SOC 2 Type II certified, GDPR compliant, CASA Tier 2 verified. All data is encrypted in transit and at rest. Permission-aware access controls. **What tools does Coworker connect to?** 50+ native connectors including Salesforce, Slack, Jira, Google Drive, HubSpot, Notion, Confluence, GitHub, Gmail, Zoom, Snowflake, Zendesk, Linear, and more. **How is Coworker different from Glean?** Glean searches and finds answers. Coworker searches, understands context, and takes action across your tools. Coworker updates CRM, creates tickets, sends emails. Glean shows you results. **How is Coworker different from Microsoft Copilot?** Copilot works within Microsoft 365. Coworker works across 50+ tools from any vendor. Copilot assists. Coworker executes autonomously. **Can Coworker replace my meeting note tool?** Yes. Coworker transcribes, summarizes, and acts on meetings. Unlike Fathom, Granola, or Fireflies, Coworker also updates CRM, creates tickets, and sends follow-ups automatically. **Does Coworker work with my existing tools?** Yes. Coworker connects alongside your existing stack via OAuth. Nothing to replace or migrate. **What size companies use Coworker?** Coworker is built for teams of 50+ employees. Ideal customers are 200-2,000 employee B2B SaaS companies with complex tool stacks.