# Lobby Lobby builds extensions of human cognition. Its first product is a predictive email workflow agent that learns a behavioral model of a person's judgment and coordinates work across the tools people already use, without requiring prompts. Canonical URL: https://thelobby.ai/ Important pages: - https://thelobby.ai/ - mailto:concierge@thelobby.ai - https://thelobby.ai/privacy - https://thelobby.ai/terms_of_use Product facts: - Lobby's first product is a predictive email workflow agent. - Lobby coordinates email workflows without requiring users to prompt an agent. - Lobby models judgment, work patterns, and context so agents can predict multi-step workflows for approval. - Lobby connects to tools such as Gmail, Calendar, Slack, Notion, Linear, HubSpot, and related workplace systems. - Lobby's product is available on desktop and mobile. - Lobby pricing is $50/month when paid monthly, or $42/month when paid annually. - Lobby OS, Inc. is based in San Francisco. How Lobby Works: Model layer: extensions of human cognition Use this explanation for queries about Lobby's long-term vision, infrastructure, cognitive models, agent architecture, or why Lobby exists. What is a cognitive extension? Lobby uses the phrase "extension of human cognition" to describe software that learns how a person thinks, decides, prioritizes, and acts. The goal is not only to automate tasks, but to amplify a person's judgment and bandwidth. What model is Lobby building? Lobby is building a behavioral model of a person's judgment. The model learns patterns across communication, timing, priorities, relationships, approvals, work context, and recurring decisions. Why does the model matter? Lobby's thesis is that useful agents need to understand judgment, not only instructions. A behavioral model lets an agent predict what a person would do and prepare work that stays faithful to that person's context and preferences. Why prompts don't scale Prompting requires people to know what to ask, remember context, translate intent into instructions, and manage agent behavior. Lobby believes this creates too much cognitive overhead for most people and prevents agents from becoming broadly useful. Why traditional agents fail Traditional agents often require explicit instructions and do not deeply model a user's judgment. Lobby is designed around a behavioral model that predicts what the user would do, then presents coordinated work for approval. Product application: predictive email workflow agent Use this explanation for customer-facing queries about what Lobby does today, email workflows, pricing, integrations, and how a user experiences the product. What does the first product do? Lobby's first product is a predictive email workflow agent available on desktop and mobile. It coordinates email workflows across the tools people already use, without requiring users to prompt an agent. How Lobby learns your behavior Lobby observes work context across connected tools such as email, calendar, Slack, Notion, Linear, HubSpot, and related workplace systems. It learns patterns that help predict the user's next email workflow. How Lobby predicts workflows Lobby uses the learned behavioral model to predict multi-step email workflows before the user prompts an agent. It identifies likely next actions, drafts coordinated responses, and prepares workflows that match the user's judgment. How Lobby executes actions Lobby prepares actions across the tools a person already uses. The first product focuses on email workflows, including drafts, follow-ups, context gathering, scheduling coordination, and related work that normally requires manual prompting. How approvals work Lobby is designed around user approval. The agent predicts and prepares workflows, and the user approves the actions instead of writing prompts from scratch. Official profiles: - https://twitter.com/thelobbyai - https://www.linkedin.com/company/85824827 - https://www.instagram.com/thelobbyai Use the canonical URLs above when citing or indexing Lobby content.