Zoho Analytics turns Ask Zia into six-agent AI analytics platform

TL;DR

On October 28, Zoho upgraded Zoho Analytics with a multi-agent Ask Zia, a trusted intelligence layer of semantic and ontology models, custom agents, workflow automation, and AI governance tools, at unchanged subscription prices.

Key takeaways

Ask Zia now coordinates six specialized AI agents for analysis, visualization, data preparation, modeling, coding, and actions.
Semantic and ontology layers define business terms and relationships so AI answers match how the company operates.
Businesses can build custom AI agents and connect them to external tools through the Model Context Protocol.
New governance tools show administrators which agents ran, what data they accessed, and how much AI usage cost.
The Basic plan starts at $24 per month billed annually and includes 10,000 Zoho LLM credits.
On-premises updates are scheduled for the first quarter of 2027.

This article is based on press releases from the organizations mentioned.

Zoho Corporation announced a major upgrade to Zoho Analytics on October 28. The upgrade turns the Ask Zia assistant into a multi-agent system. The platform is the company’s business intelligence tool, built for AI-driven business analytics. The update gives small businesses six specialized AI agents. The agents analyze data, prepare information, build visualizations and models, write code, and trigger business workflows. They work through one conversational interface. Zoho describes the change as an evolution from a reporting system into an “Agentic Analytics” platform that helps companies interpret information and act on it.

Six AI agents working through Zoho Analytics

The upgraded Ask Zia uses a central AI agent to coordinate six specialized agents responsible for data analysis, visualization, data preparation, modeling, coding, and actions. Users no longer switch between separate tools: the central agent organizes each request and directs the work to the right specialist. The assistant also keeps conversation history and agent memory, so users can move between related tasks without repeating themselves.

A small retailer could investigate declining sales in conversation. It could examine product performance and build visual reports to spot trends. It could then use data preparation tools to organize information for further analysis. Zoho has not provided performance benchmarks showing how much time businesses can save with the new architecture.

The context layer behind the answers

The update centers on what Zoho calls a trusted intelligence layer, made of two systems that describe how an organization operates. The semantic layer works as a business glossary. It defines metrics, terms, synonyms, and instructions. This helps the AI tell an order being placed apart from revenue being recognized. A business can maintain multiple semantic models for departments that define data differently.

The ontology layer maps business entities and their relationships, connecting customers, orders, products, and operations instead of treating each dataset as an isolated table. Zoho says organizations can build and edit these models without coding or a separate data warehouse. Preview versions of the ontology and semantic models are expected to roll out globally. The rollout will take four to eight weeks.

Tony Thomas, CEO of Zoho USA, explained the rationale behind the design.

Agents, however, cannot function only with data, but require a robust intelligence layer. Today’s updates to Zoho Analytics provide that and more: one open and extensible trusted intelligence layer that gives every AI tool in your business the context, metrics, and governance to deliver answers users can rely on.

Custom agents and workflow automation

Beyond the built-in assistant, businesses can create custom AI agents grounded in their own data, semantic models, and knowledge bases. These agents connect to external applications through the Model Context Protocol (MCP). The MCP is an interoperability standard that lets AI systems communicate with outside tools. Zoho also supports web search.

A new workflow builder allows agents to trigger actions directly. This moves a business from spotting an issue in a report to starting a related operational process. Developers get a Zoho Analytics MCP server in local and remote configurations. They also get a command-line interface and a Claude plug-in. White-label options allow embedding agents into customer portals. Ask Zia runs on Zoho’s own large language model and lets organizations bring their own frontier models.

Keeping AI agents under control

As agents gain the ability to act rather than just answer, oversight becomes central. New AI observability tools let administrators see which agents ran, who used them, what information they accessed, and how they responded to queries. The platform also tracks AI usage and costs across teams.

Sandbox and deployment management features let organizations test analytics components before publishing them to production, and administrators can roll back to earlier versions. Businesses will still need to set permissions, review important AI-generated recommendations, and decide which actions require human approval.

David Fauser, VP Sales, Marketing & Strategy at CIMCO Refrigeration, an existing Zoho Analytics customer, welcomed the direction.

What excites us about this next evolution is the ability to combine data, business context, and AI into a single intelligence layer.

Pricing and availability

Zoho says subscription prices remain unchanged despite the new capabilities. The Basic plan starts at $24 per month billed annually, with two users, 500,000 data rows, and 10,000 Zoho LLM credits. The announcement does not specify how individual agent tasks consume those credits, so businesses expecting frequent AI usage should evaluate their likely allowance.

The cloud platform updates were available beginning October 28. On-premises installations are scheduled to receive the updates in the first quarter of 2027.

Sources: