The Architecture Shift Behind Reliable Enterprise AI - with Ravi Marwaha of Arango

2026-05-14 · Source: The AI in Business Podcast · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Robotics & Autonomous Systems · Depth: Intermediate, extended

Summary

Ravi Marwaha, COO & Chief Technology Product Officer at Arango, discusses why AI pilots often fail to scale across enterprises, attributing the issue not to model performance but to fragmented, inconsistent data context. He highlights that while models are highly capable, their effectiveness is hampered when they cannot access unified, relevant, and timely information required for reliable decision-making at scale. Marwaha emphasizes that companies must redefine what specific information each AI use case truly needs, how that data is accessed, and how agent-generated decisions will be stored and governed. This necessitates a fundamental re-architecture of data practices, moving beyond traditional data warehousing to a context-focused approach that supports agentic AI operations and auditability, rather than solely human consumption.

Key takeaway

For CTOs and VPs of Engineering struggling to scale AI pilots, your focus should shift from optimizing models to re-architecting your data strategy around "context." Define the precise, timely, and relevant information each AI agent needs, and establish new systems for storing and governing agent-generated decisions and actions. This paradigm shift will enable reliable, auditable AI deployments and prepare your enterprise for agent-driven operations, moving beyond fragmented data practices.

Key insights

AI scaling failures stem from fragmented data context, not model limitations, requiring a re-architecture for agentic systems.

Principles

• Models are not the problem, context is.
• Focus on business problems, not AI trends.
• Agents need context, not just data.

Method

Define specific, timely, and relevant information for each AI use case; establish new architectures for agent memory, decision storage, and governance; and prepare data for AI consumption via an AI services layer.

In practice

• Prioritize relevant, timely, and accurate data for AI agents.
• Rethink data architecture for agent-driven operations.
• Implement governance for agent-generated decisions and actions.

Topics

• Enterprise AI Adoption
• Data Context Management
• Agentic AI Architecture
• Multimodal Data Platforms
• AI Governance

Best for: CTO, VP of Engineering/Data, Executive, Director of AI/ML, AI Architect, Consultant

Related on AIssential

• Tackle Enterprise AI’s Hardest Question At Forrester’s AI Forums — Enterprise AI success requires a cognitive operating model, skills-oriented architecture, and robust context engineering to deliver business value.
• Understanding the data core: From legacy debt to enterprise acceleration — AI scalability and reliability depend on reinventing the data core, not just upgrading technology.
• The End of Software’s Monopoly on Work — AI agents shift enterprise work from applications to a trusted data context layer, demanding unified, governed data for reliable action.
• What it takes to scale agentic AI in the enterprise — Scaling agentic AI demands robust governance, data infrastructure, and system integration, not just advanced technology.
• Jedify raises $24 million Series A to build context layer for enterprise AI - CTech — Fragmented enterprise data impedes AI agent deployment, necessitating a unified "context graph" for reliable operation.
• From AI promise to business impact: building future-ready enterprise AI — Scaling enterprise AI requires addressing data quality, architectural complexity, and realistic expectation management.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The AI in Business Podcast.