IBM’s New Acquisition Highlights Organizations Aren’t Ready for Real-Time

IBM’s recent acquisition of data streaming company Confluent gained headlines worldwide over its promise of access to live, operational data to transform global business procedures. However, they are missing one critical point – enterprise data isn’t ready for real-time. I’ll explain why, but first let’s look at what IBM is aiming to achieve with this multi-billion-dollar buyout.

Through this acquisition, IBM promises to provide clients with Confluent’s real-time data streaming platform, enabling them to deploy generative and agentic AI at a faster rate. The partnership boasts a future where AI agents seamlessly process live, operational data and transform business operations through a single platform. In theory, this would create a streamlined opportunity for AI agents to process data more efficiently and accurately. In practice, the data that they are aiming to streamline is unstructured and unavailable for real-time processing. Here’s why: imagine a funnel designed to transport water, but instead it’s filled with mud — messy, unstructured, and impossible to process efficiently. Herein lies the problem with IBM’s thinking: making the AI agents faster does not solve the core issue.

While the tech world is fixated on real-time data streaming, the more pressing issue is the overwhelming volume of unstructured documents – contracts, invoices, and forms, which hold the critical context that real-time AI often overlooks. With Gartner’s findings of 80% of enterprise data being unstructured, and a predicted 60% failure rate of AI projects without a strong data foundation, ignoring data readiness is not an option. While AI agents promise autonomy and speed, data is often hidden, hard to access, and needs context. Increasing the speed at which unstructured data is accessed is not only counterproductive but also a threat to enterprises whose data is not ready before processing.

See also: The Convergence of AI and Real-Time: IBM Acquires Confluent

Bridging the Gap: From Data Silos to Intelligent Workflows

Here’s what analysts are seeing on the ground: data silos are clogging the AI funnel. According to IDC MarketScape’s recent report, “Buyers want data privacy assurances and data encryption, along with robust auditing capabilities and traceability to build and extend user trust”. This insight emphasizes the relationship between compliance, trust, and successful AI implementation – and the importance of bridging the gap between the lofty promises of AI and the reality of enterprise readiness. A clear example of this is the Know Your Customer (KYC) processes to provide financial institutions and insurance organizations with a compliance-first intelligent workflow. The challenges of KYC mirror those of enterprise data streams: manual data entry is slow, error-prone, costly, and falls short in terms of document validation and audit readiness. While the specifics may differ, the overarching goal remains the same: bridge the gap between unstructured and structured data.

Now that we understand why implementing faster data streams alone will only amplify the problem, let’s explore how to solve it. The complexities of Know Your Customer (KYC) will serve as a process example for how to properly implement AI. The approach involves three structured steps, all logged, explainable, and auditable. The process begins with sophisticated, purpose-built extraction of data from documents, even if they are poorly scanned or vary by country. This foundational step, missing from the IBM and Confluent partnership, ensures that unstructured data is transformed into clean, usable formats.

The next step in the KYC integration is IBM watsonx.ai Orchestrate, which routes clean data through the KYC workflow. This is where validation, errors, and follow-ups get flagged and integrated directly into internal compliance platforms.

The third and final step is process intelligence, which monitors each case’s operational data in real time. When exceptions arise, instead of acting on that information at high speed without the proper understanding of what action to take, the process intelligence alerts the right team and triggers the corrective action. This is a perfect example of Agentic AI implementation done correctly.

Even with a well-structured process, there are still challenges that tie back to our core issue. While document AI can understand complex, unstructured data with up to 90% accuracy, there are still large amounts of data that remain locked up and unreachable from AI agents. According to the Fivetran AI and Data Readiness Survey, 42% of enterprises reported that more than half of their AI projects have been delayed, underperformed, or failed due to data readiness issues. Additionally, 68% of organizations with less than half of their data centralized report lost revenue tied to failed or delayed AI projects. This is especially striking considering that in IBM CEO Arvind Krishna’s own words, their “data is spread across public and private clouds, data centers, and countless technology providers”. These statistics illustrate the need for enterprises to shift their focus from deploying as many AI tools as they can to preparing their data prior to AI implementation.

See also: How Kafka and Edge Processing Enable Real-Time Decisions

Compliance as a Competitive Advantage in AI Adoption

When preparing enterprise data for AI systems, business leaders should prioritize a proactive approach to compliance. As we look ahead to 2026, Forbes predicts that “Governance and auditability will become competitive assets”, with surveys finding that “59% of enterprises say regulatory compliance is their top challenge in managing data for AI”. This reinforces that the enterprises that succeed will be those that prioritize data readiness as the foundation for AI adoption. The broader trend is that AI is no longer about speed or autonomy alone; it’s about context, compliance, and trust.

With the volume of unstructured enterprise data and risk of failing AI projects mounting, the focus will shift from deploying AI tools to building the infrastructure that supports them. By addressing these data challenges and implementing ethical, purpose-built AI systems, businesses can unlock AI’s transformative potential while mitigating risks. The defining trend for 2026 will not be the speed of AI, but the preparation that powers it. Enterprises that embrace this shift will lead the way in shaping a future where AI delivers on its promises.