In this Cloud Wars conversation, Bob Evans sits down with Shub Bhowmick, CEO and Founder of Tredence, alongside Yasmeen Ahmad from Google Cloud to explore how enterprises are moving from AI applications to AI agents. Their discussion focuses on what it takes to turn intelligence into action — covering data foundations, semantic layers, agentic architectures, and the operational shifts required for businesses to scale AI successfully.
Turning AI Into Action
The Big Themes:
- AI Agents Redefine Applications: Traditional AI apps assist by querying data, generating recommendations, and supporting limited workflows. AI agents, however, represent a much deeper operational shift. As Ahmad explains, agents are multi-step reasoning engines that can access multiple systems, coordinate actions, and execute entire business processes autonomously. Instead of simply helping humans decide, they can perform work across ERP systems, supply chains, and customer interactions. This changes the role of the database itself — from a storage and query engine into a reasoning engine with vector search, graph RAG, and semantic understanding. Examples like Home Depot and Danfoss show how this creates massive efficiency gains
- Why Questions Require Agentic Intelligence: Shub Bhowmick draws a critical distinction between “what” questions and “why” questions. A conversational BI system can answer what happened — such as how much sales dropped — but a “why” question demands deeper reasoning. Why did sales decline? Was it pricing pressure, competitor behavior, inventory constraints, or macroeconomic events? These problems require hypothesis-driven exploration. Tredence addresses this through business semantic layers, knowledge graphs, and hypothesis banks that support open-ended reasoning.
- Closed Systems Create Long-Term Risk: Bhowmick warns against enterprises rushing toward closed, inflexible platforms simply because they promise faster short-term value. While packaged solutions may accelerate deployment, they often restrict ownership, adaptability, and future innovation. In contrast, open architectures built with hyperscalers like Google Cloud allow customers to own the IP, customize solutions, and evolve as the market changes.
The Big Quote: “Gone are the days when these migrations used to take 12 to 18 months. Nowadays, you have to complete these migrations in less than three to four months.”
More from Tredence and Google Cloud:
Learn about the partnership between Tredence and Google Cloud and AI agents on Gemini Enterprise.
