Before building your first agent, it helps to understand which type fits your goal. Zoho Zia Agents supports several distinct configurations, each designed for a different kind of work. Choosing the right one from the outset makes setup quicker and results more reliable. That said, agent type is not a permanent decision — you can adjust the configuration at any point.
When you create an agent in Zia Agents, its type is determined by two things you provide: a Knowledge Base and tools.
The combination of these two elements places your agent into one of four types:
| Agent Type | Knowledge Base | Tools | Best for |
|---|---|---|---|
| No tools or KB | No | No | Controlled AI assistant for drafting, brainstorming, or general Q&A |
| Knowledge Base only | Yes | No | Responses grounded in your own documents and content |
| Tools only | No | Yes | Performing actions in your systems via API |
| Tools and Knowledge Base | Yes | Yes | Agents that need both business context and the ability to act on it |
The simplest configuration. The agent relies entirely on the language model's built-in knowledge to respond. It can draft content, summarise topics, answer general questions, and generate ideas — but it has no awareness of your specific business data and cannot interact with your systems.
The advantage over using a standalone LLM is the controlled environment: you define the agent's role, set its behavioural instructions, and configure guardrails that limit what it can and cannot say. Those boundaries are applied consistently across every session, and you gain visibility into performance over time. Good use cases include writing assistants that help your team match a consistent tone, brainstorming tools that stay within brand guidelines, or general Q&A agents where you want to control how responses are framed.
The agent searches through your uploaded documents before generating a response, grounding its answers in your actual content rather than general training data. This is a good fit when the agent's primary job is to answer questions accurately based on information specific to your business. Typical examples include a support agent in your help centre that resolves common queries from your documentation, a product FAQ assistant customers can chat with, or an internal policy bot that helps employees find the right information without raising a ticket. The agent understands your world but does not take any actions within it.
This configuration gives the agent the ability to perform actions via API calls, without any uploaded documents to reference. It uses the language model's general reasoning to decide when and how to use the tools you have connected. This works well for operational agents whose job is primarily about executing tasks rather than answering questions: for example, an agent that monitors new CRM leads, scores them, assigns them to the right representative, and sends an introductory email; or an agent that watches a support queue, routes tickets by urgency, and updates statuses automatically.
The full configuration. The agent has both your business context from uploaded documents and the ability to act through connected tools. It can understand your processes, reference your documentation, and follow through with actions in your systems. Most teams building production-ready agents end up here. Examples include a sales assistant that understands your product catalogue and can update deal records, a support agent that references your troubleshooting guides and escalates tickets it cannot resolve, or an operations coordinator that reads your internal playbooks and triggers the right workflows.
Beyond the four types above, Zia Agents also supports connecting multiple specialised agents into a single workflow, where each handles a specific part of a larger process. Each agent completes its step and passes context to the next. A hiring workflow, for instance, might have one agent screening applications, another scheduling interviews, and a third generating offer letters — all coordinating end-to-end without manual handoffs.
This is not something you need to start with. Most teams begin with a single agent and expand into multi-agent setups as their processes become more complex.
Most teams start with a conversational agent backed by a Knowledge Base because it is the most intuitive to set up and test. From there, you can layer in tools or connect it to a larger workflow as your needs evolve.