# ZooWork ZooWork lets domain experts build and run AI agents grounded in business knowledge, tools, and workflows, with approval gates for actions and execution history for review. - Findings24: - Website: - Pricing: Freemium - Categories: AI Agent Builder, AI Workflow Automation ## What is ZooWork? ZooWork is an AI agent platform for domain experts, teams, and product developers who want agents to carry out business workflows rather than only answer prompts. Users define an agent’s role, goals, company knowledge, working procedures, delivery standards, and tool permissions, then connect it to relevant systems. Agents can plan multi-step tasks, use tools, produce work such as drafts or replenishment recommendations, and pause for human approval before consequential actions. ZooWork also offers an API and SDK for embedding managed agents in applications. ## What can ZooWork do? ### Conversational agent building Describe a task in Agent Builder and refine the resulting agent through conversation. Templates for industry and role workflows provide starting points for configuring an agent. ### Business knowledge and skill configuration Set role goals, provide company context and policies, document working procedures, and define delivery standards so the agent can follow a specified playbook. ### Tool use with approval gates Agents can read from connected systems and prepare actions; approval policies can require a person to review before the agent writes data or commits an action. ### Managed execution and event history The managed runtime executes agent sessions in a sandbox. The API exposes durable, sequence-numbered events so developers can follow execution and retrieve results. ### Model and service connections The platform lists models from providers including OpenAI, Claude, Gemini, DeepSeek, Kimi, and GLM, plus connections such as Google Workspace, Microsoft 365, Slack, Jira, and GitHub. It also supports MCP-compatible internal tools. ### ZooData structured extraction ZooData converts URLs into structured JSON for agent use, with examples spanning e-commerce, finance, food delivery, and social platforms. ## Use Cases ### Retail inventory and replenishment Combine inventory, sales, and supplier information to identify stores at stockout risk, draft replenishment actions, and route proposed system updates for approval. ### Agency proposal preparation Apply a firm’s case library, brand voice, and proposal process to turn a client brief into a first draft for a team member to review. ### Sales pipeline review Use CRM opportunity data to summarize open deals, stages, amounts, close dates, and potential risks for sales teams. ### Product-embedded agent workflows Developers can create and start agents through the TypeScript SDK or Python client, open sessions, send tasks, and consume execution events in their own applications. ### Recurring operational follow-up Configure agents to run scheduled tasks or re-check work, such as verifying store actions after an initial review and escalating unresolved issues. ## Frequently Asked Questions ### Can an agent take actions in connected business systems? Yes. Agents can use connected tools for reads and writes. Approval gates can be configured for actions that require human review before committing changes. ### Can developers embed ZooWork agents in an application? Yes. The Managed Agent API and SDK documentation describes creating agents and sessions, sending events, and following execution. TypeScript and Python examples are provided. ### How are agent instructions and capabilities configured? An agent can be configured with a role or persona, skills, model, instructions, tool policy, and permissions. Business context, procedures, and delivery requirements can also inform its work. ### Does ZooWork support different model providers? The platform names OpenAI, Claude, Gemini, DeepSeek, Kimi, and GLM among its available models. The pricing page describes access to more than 20 leading models. ## Quick Facts - **Product category**: AI agent platform and managed agent runtime - **Primary users**: Domain experts, business teams, and developers embedding agents - **Agent inputs**: Business knowledge, instructions, tools, and workflow context - **Agent outputs**: Task plans, drafts, structured data, recommendations, and tool actions - **Developer interfaces**: Managed Agent API, TypeScript SDK, and Python client - **Website**: zoowork.ai