Peer reviewers at [conference] noted imagine a treasure trove of information lying dormant in your company's archives-call logs, customer interactions, and video footage, all too chaotic to analyze. For many enterprises, this unstructured data is a frustratingly underutilized asset, but now, a new wave of AI technologies is transforming it into actionable intelligence.

As organizations pursue a competitive edge, they are uncovering that up to 90% of the data they generate is unstructured, often complicating their efforts to harness its potential. With advancements in AI and natural language processing, firms are beginning to extract insights from this neglected resource, reshaping their strategies and outcomes across various industries. (A new CRISPR startup is betting regulators will ease up on gene-editing)

The Challenge of Unstructured Data

Unstructured data encompasses text documents, videos, call records, and social media posts, presenting inherent challenges for traditional analytics methods. Its format variability and inconsistent quality make analysis difficult without specialized tools.

Success Story: The Charlotte Hornets' AI-Driven Draft Strategy

Estimates suggest that as much as 90% of the data generated by organizations is unstructured. This includes insights from customer interactions and operational signals that could be critical for decision-making. Yet, the nature of this data has historically rendered it largely unusable without significant preprocessing and organization. (A new CRISPR startup is betting regulators will ease up on gene-editing)

A standout example of leveraging unstructured data comes from the NBA’s Charlotte Hornets, who utilized AI tools, including computer vision, to analyze vast amounts of gameplay footage from lesser-known leagues. This data was previously seen as too cumbersome to process.

Preparing Unstructured Data for AI Consumption

By implementing AI techniques like object and player tracking, the Hornets generated novel kinematic metrics that identified promising recruits who might have been overlooked by traditional scouting methods. The results were remarkable-one of their selections was later named the MVP of the 2025 NBA Summer League.

This case underscores how integrating unstructured data analytics can disrupt established scouting processes, offering significant advantages in talent acquisition.

AI's Future Role: Insights and Innovations

To unlock the potential of unstructured data, organizations must first prepare it for AI applications. This involves creating intuitive data collection processes and structuring data to be easily consumable by AI models.

Jordan Cealey, senior vice president at AI company Invisible Technologies, emphasizes that successful AI deployments rely on thorough preprocessing. "Data can only be utilized once it is organized and ready for AI. You cannot just throw AI at a problem without doing the prep work," he cautions.

The traditional consulting model may no longer suffice, as businesses require agile solutions tailored to current demands. Forward-deployed engineers (FDEs) have emerged as a viable model, facilitating close collaboration between tech firms and businesses to ensure optimal model configuration.

Constraints and tradeoffs

• Requires significant upfront investment in data cleaning and preparation

• AI models must be adapted to domain-specific language and context

• Integration issues may arise from disparate data sources

Verdict

Unlocking unstructured data through AI can drive significant business outcomes, yet success hinges on effective data management and tailored model configurations.

AI's Future Role: Insights and Innovations