Artificial intelligence is changing how companies operate but it’s only as effective as the data feeding it.Rohit Vyas, Director of Solution Engineering, Confluent India, explains that trusted, real-time data is essential for building accurate and reliable AI systems. Confluent, founded by the original creators of Apache Kafka, helps organisations stream data smoothly and securely from different sources. By making it easy to move and manage high-quality data, Confluent supports businesses in building smarter applications, improving decisions, and preparing for a future shaped by responsible and governed AI.
How is Confluent helping companies use real-time data to enhance their AI systems?
Confluent is uniquely positioned in this space, having pioneered data streaming and co-founded Apache Kafka. We understand the dynamics of fast, real-time data. In today’s AI-driven world, the phrase “no data, no AI” has evolved — we believe “no trusted data, no trusted AI.” And, if AI needs to sustain, real-time data is the key.
We provide high-quality, real-time data pipelines essential for AI and large language models (LLMs), including support for Retrieval-Augmented Generation (RAG). Our platform integrates effortlessly with a wide range of technologies at both the source and destination, enabling continuous, reliable data flow that powers smarter and more accurate AI.
How does Confluent support the development of AI-powered applications in real time?
Modern applications are distributed and API-led. Traditionally, point-to-point integrations have created tight coupling between systems, which limits flexibility and increases maintenance complexity. Confluent addresses this by decoupling services through an intelligent streaming layer, enabling scalable, resilient, and modular architecture. Building a real-time data mesh democratizes data across the enterprise. This enables AI teams to find data sets and tap into them in a frictionless manner without having to directly work with the owner and set up a point-to-point integration.
For AI systems, especially predictive AI, the volume and velocity of data are critical. Brittle infrastructure can lead to increased latency and cost. Confluent offers robust, industry-proven connectors, supports schema evolution, and provides strong data governance. With a single platform, organisations can meet the needs of applications, operations, cybersecurity, and AI—without having to maintain separate technology stacks.
How does Confluent manage and protect the data used by AI models, particularly around accuracy and security?
Data quality is fundamental. One of Confluent’s core strengths is the enforcement of data contracts through evolving schemas. If incoming data doesn’t conform to the schema, it can either be rejected or redirected to a dead-letter queue for manual review—ensuring only high-integrity data enters the system.
In multi-LLM environments (e.g. using Llama and Cohere), Confluent preserves context and enables validation across each handoff. It can check for inappropriate language, content quality, or compliance with specific rules. This level of centralised governance builds trust and consistency across AI pipelines without fragmenting the quality assurance process.
With many Kafka tools on the market, what differentiates Confluent for AI-focused organisations?
Confluent delivers far more than just Kafka—we offer an enterprise-grade data streaming experience. We set data in motion to power the world’s real-time operations and analytics. While others may use open-source Kafka, Confluent leads its development: over 80% of Apache Kafka’s code committers are employed with us. We also chair Kafka’s governance body and help define its roadmap.
Other vendors might offer temporary solutions by customising Kafka, but they can’t guarantee long-term compatibility or integration with future updates. Confluent ensures sustainability, reliability, and forward-compatibility—critical for enterprise-grade AI systems. The Data Streaming Platform (DSP) we designed serves as an intelligent connective tissue linking all of the applications, systems, and data layers within the company into a single central nervous system of real-time streams of data.
Moreover, we’ve extended this influence to Apache Flink, the de facto standard for stream processing. Following our acquisition of Immerok in early 2023, we now host the majority of Flink committers and have brought the technology to the cloud in a fully serverless format.
We now offer what we call the “holy trinity”: Kafka (for streaming), Flink (for processing), and TableFlow (for stream-to-data lake integration)—covering both operational and analytical workloads. Whether it’s applications, operations, cybersecurity, or AI—you’ll find Confluent playing a pivotal role. Additionally, Confluent Cloud accelerates AI initiatives by delivering managed services and ensuring compliance with data residency and privacy regulations—an essential requirement for Indian enterprises.
What does the future hold for AI and real-time data streaming, and how is Confluent preparing for it?
The future of AI hinges on responsible development and governance. It’s a powerful tool—but also a double-edged sword. If regulated properly, it can be transformative; without control, it can be harmful.
Confluent has always operated with caution and integrity, adhering to legal frameworks and ethical standards. We strive to use technology for the betterment of society and actively call out areas of risk or misuse. As a responsible corporate citizen, we remain committed to promoting safe, governed, and effective AI adoption.
Could you share examples of Indian organisations using Confluent for AI?
Certainly. Several leading organisations in India are leveraging Confluent for AI-driven use cases. Swiggy uses Confluent extensively across a range of generative and creative AI applications. The reliability of Confluent’s platform is integral to Swiggy’s operations, supporting everything from real-time order processing to dynamic user interactions and personalization.
Viacom18 makes use of real-time data pipelines for clickstream analysis, enabling precise and timely advertisement placements. Our real-time streaming infrastructure ensures their AI-driven analytics make intelligent, data-backed ad decisions.
Zomato relies on Confluent throughout the entire customer journey—from login to delivery—using real-time data to power personalised recommendations and optimise logistics, such as assigning the nearest delivery rider. Our real-time data streaming underpins both the user experience and operational efficiency.
Meanwhile, Zee5 harnesses Confluent to deliver near real-time personalisation of its homepage and content suggestions, ensuring that each user receives a tailored viewing experience upon logging in.
We also enable some of India’s most innovative digital platforms to derive business value from real-time, trustworthy data, supporting both operational excellence and AI-powered personalization. We’re privileged to work with a lot of the innovative brands and support their AI journeys.
