The ability of AI models to interact with external systems, APIs, and software tools, making decisions about which tools to use and when, then executing those decisions to accomplish complex multi-step tasks.
Also known as Tool calling, API access, external integrations
Tool use enables AI models to interact with the outside world beyond text. Instead of only generating words, a model can decide to call an API, execute a command, query a database, or trigger an action in another system. The model reasons about what needs to happen, selects the appropriate tool, and uses it correctly. The results come back, and the model incorporates them into its thinking.
This is the foundation of AI agents. Without tool use, AI is a text-in, text-out system. With tool use, AI becomes a system that can accomplish real work: schedule meetings, pull data, post content, modify files, update databases. The key is that the model decides when and how to use tools, not humans instructing it step-by-step.
Agency work involves moving information between systems: logging data in Airtable, posting content to platforms, updating spreadsheets, scheduling messages, pulling reports. Tool use turns AI from a thought partner into a hands-on team member that can actually do the work.
Automation with guardrails. Tool use enables you to give AI access to specific actions while maintaining control. You can configure which tools an AI has access to, which actions it can take, and which decisions require human approval. An AI can auto-post routine content but requires human approval for client-sensitive messages.
Workflow integration. Instead of copying outputs from AI and pasting them into other systems, tool use enables seamless integration. An AI drafts copy, approves it, and posts it directly. An AI reviews client feedback, categorizes it, logs it in the CRM, and flags urgent issues for the team.
Reduced manual data movement. Your team spends significant time moving information between systems. Tool use eliminates that friction by letting AI connect the systems directly.
Your social media AI receives 150 customer comments daily across Instagram, TikTok, and Twitter. Without tool use, it would generate text analysis of each comment. With tool use, it categorizes each comment (compliment, question, complaint, irrelevant), then uses the appropriate tools: for compliments it posts a grateful reply; for questions it pulls relevant product information and drafts an answer; for complaints it escalates to a human agent and logs the issue in your support queue; for irrelevant comments it ignores them. All this happens automatically, and your team only sees the escalations that need human attention. The AI isn’t just thinking about what should happen; it’s making it happen by using the tools at its disposal.
The automations and agents module of the workshop teaches you how to build AI workflows that compress the busywork without taking the craft out of the studio.