Many AI Projects Fail Due to Problematic Data Layers
A large proportion of artificial intelligence (AI) projects fail before reaching even a single customer due to problematic data layers. Companies typically have models and mandates in place, but security risks and fragmented data prevent them from launching products.
‘We address this by making streaming layers the foundation for secure and production-ready AI,’ said Sean Falconer, Head of AI at Confluent, in a written statement on Monday (25 May). ‘In the APAC region, AI implementation faces harsh realities where most projects never progress beyond testing due to data layers that lack sufficient security or scalability for production use.’
Vice President and General Manager for APAC at Confluent, Greg Taylor, added that the company understands this. With automatic privacy features and AI-specific tools, Confluent provides organisations with a clear and secure pathway to create real-world use cases that deliver significant business impact.
The issue is widespread, according to a McKinsey report stating eight out of ten companies cite data limitations as a barrier to developing autonomous AI. The primary causes often involve security teams blocking data from entering AI workflows due to exposure risks, and developers spending hours switching tools to check and manage the data streams vital to their AI systems.
This slow manual process turns what should be a rapid iteration cycle into a bottleneck. Therefore, IBM and data streaming pioneer Confluent have released new features in Confluent Intelligence and Confluent Cloud to simplify the development and security of real-time AI applications.
These updates remove security barriers and complexities that prevent organisations from deploying AI workloads in real-world scenarios. Confluent unifies the AI lifecycle with tools developers already use, integrating Apache Flink pipelines with dbt, and introducing a fully managed Model Context Protocol (MCP) server and Agent Skills that enable AI to manage streaming operations.
With automatic PII masking and private connectivity to external models via Azure Private Link, Confluent integrates enterprise-grade governance directly into data streams.
Enrich, an AI-powered HR software, is designed to transform how companies build and manage their workforce.
GenDigital Academy successfully hosted the Code the Future AI Summit 2026 in Jakarta, showcasing AI innovations by students and university students for Indonesia’s digital future.
A recent survey in Japan revealed that 47.8% of elderly women prefer AI for relationship consultations, citing comfort and lack of judgment.
Seoul police revealed that voice recordings and text messages from Kim Sae-ron implicating Kim Soo-hyun were AI-manipulated by YouTube HoverLab.
Global crypto exchange OKX has officially announced Agent TradeKit for AI developers and ecosystem builders in Southeast Asia, including Indonesia.
Companies aiming to compete in the digital age require AI systems capable of collaborative work and real-time data response.
The Sell with Confluent programme aims to simplify partner offerings by providing easier tools and support to accelerate the sale of data streaming solutions.
In today’s fast-paced world, data streaming delivers instant insights into every aspect of business.
Data streaming infrastructure is now a prerequisite for modern AI application development, especially as business competition increasingly relies on speed and data accuracy.
Amid accelerated digital transformation and increasing regulatory pressure on data management, companies now require cloud-speed solutions.