Zoho unveiled Zia Chat, a unified conversational interface that spans its apps and connected external systems, and expanded into vertical SaaS across seven industries including retail, automotive, BFSI, healthcare, education, real estate, and restaurants.

The platform integrates into existing workflows with role-based profiles, delegates multi-step tasks to Zia Agents, and maintains data provenance, ensuring data accessed isn’t used to train Zoho’s models.

The AI interface will launch in English with early access planned in about a month, designed to pull together information, workflows, and industry-specific rules across the organization.

Zoho’s FY25 results show revenue of ₹12,313 crore with 17.7% operating revenue growth and ₹3,191 crore in net profit, while FY26 revenue figures were not disclosed.

Zoho remains privately held, profitable, with more than 19,000 employees and over 150 million users served globally from its own data centers.

The company faces rising data center and token-related AI costs and is focusing on cost discipline and integrated AI value rather than relying on standalone AI revenue or price hikes.

Zoho emphasizes responsible AI use, warning against over-reliance and stressing that employees must understand and take responsibility for AI outputs.

The announcements come amid broader enterprise AI adoption and geopolitical tensions, with leaders highlighting India’s need to balance protecting domestic interests while pursuing global opportunities.

Founder Sridhar Vembu argues software is becoming a commodity as AI lowers app-building costs, but Zoho’s edge lies in integrating apps, staying regulation-ready, and anchoring AI in structured business data.

Vembu stressed reducing token costs and creating a single interface to minimize fragmentation and boost efficiency.

The strategy combines AI’s broad knowledge with Zoho’s structured data, enabling a context-rich AI layer while upholding privacy, permissions, and regulatory controls, especially in sensitive sectors like healthcare.

The system enforces enterprise permissions, with approvals kept with the appropriate approvers, auditable user actions, and responses linked to underlying records for source verification, while reaffirming that accessed data isn’t used to train Zoho’s AI models.