Zoho, the enterprise software solution company provided a view of how Zoho is approaching AI adoption in enterprises from a complete customisation and fit-for purposes perspective, rather than a one size fits all approach. The presentation by Andrew Bourne, Head of Zoho EMEA, outlined the development and strategy behind Zoho’s Large Language Model (LLM) AI, called Zoho ZIA.
Their core goal in developing their own proprietary AI system was to provide the market with an adaptable, privacy-first, highly integrated AI solution tailored specifically for Zoho’s ecosystem of enterprise business applications.
Summary of Zoho’s LLM AI Development and Approach
- Zoho has built a connected ecosystem of over 55 enterprise applications including sales, marketing, finance, HR, analytics, IT, and collaboration tools, all unified under a common technical stack and data platform.
- Their Zia AI is deeply embedded within Zoho’s apps to enhance functionality contextually and meaningfully, improving customer experience, automation, and decision-making.
- Zoho has always approached the LLM developments with proprietary AI models and infrastructure (including homegrown LLMs with billions of parameters) to avoid reliance on external AI vendors like ChatGPT. This ensures cost-effectiveness, data privacy, and better model tuning for real business workflows.
- Localization and global reach were also important factors taken into consideration, with local support, pricing in local currencies, and integration with local third-party services, tailored to different markets, including Africa.
- The AI platform includes tools like Co-Creator (to generate apps from prompts), ZioSearch (advanced AI search), and AI-powered analytics and customer interaction management.
- Strong emphasis on privacy-first design ensures no customer data is used to train external models, building customer trust through data security and compliance.
- Zoho also focuses on right-sizing AI models (using smaller, specialized models for specific tasks to save resources and costs) and multi-model approaches to optimize performance and innovation.
Advantages of Zoho’s Customised LLM AI (as outlined by Andrew Bourne)
- Privacy-first: Customer data stays within Zoho’s ecosystem; no data is leaked or used externally, which ensures compliance and builds trust.
- Cost Efficiency: Owning and maintaining their own models avoids high costs and pricing volatility associated with third-party AI providers.
- Contextual Intelligence: AI models are integrated directly with business data, enabling smarter, context-aware insights, automations, and decision support.
- Flexibility and Customisation: Businesses can build and configure AI tools tailored to their unique workflows using low-code platforms like Zoho Creator and the AI agent builder.
- Localization: Local offices, languages, and integrations make the AI practical and effective across diverse markets.
- Infrastructure Control: Zoho controls the full stack (data centres, infrastructure, AI models), ensuring reliability and performance at scale.
- Embedded AI Focus: The AI is designed as an embedded feature to enhance business operations rather than general consumer use.
- Bespoke AI Models: Combination of various smaller and specialist AI models instead of relying solely on large generic models.
Rationale for Businesses to Use Customised AI Instead of General ChatGPT-Type Solutions
- Data Privacy and Security: Zoho’s AI ensures business data never leaves the secure Zoho environment, unlike general AI services that may reuse customer data for training.
- Cost Control: Zoho avoids the cost unpredictability and potential price hikes from third-party AI providers, offering more predictable and affordable AI integration.
- Tailored Intelligence: AI is designed specifically for the types of tasks and workflows businesses run in Zoho apps, delivering more relevant, actionable insights.
- Operational Integration: The AI’s deep integration enables automation and augmentation across sales, finance, HR, support, and more, not possible with standalone models.
- Scalability and Performance: Control over infrastructure allows Zoho to optimize AI performance for large enterprises with complex data needs.
- Customisation and Extendibility: Businesses can create their own AI agents via low-code prompting, including tailoring AI functionality uniquely suited to their needs.
- Compliance and Localisation: Increased assurance of compliance with local data laws and better alignment with local business realities.
In essence, Zoho’s customised AI solution offers enterprises a secure, cost-effective, highly relevant, and configurable AI experience embedded directly into their business software, providing significant advantages over generalized AI platforms like ChatGPT, which are primarily consumer-focused and not tailored to the nuanced needs of complex enterprise environments. This makes Zoho’s LLM AI a compelling choice for businesses prioritising privacy, integration, and operational effectiveness.
