Construction companies building bridges, towers, and factories depend on employees accessing correct information at critical moments. One major construction company understood this imperative and poured extensive information—policies, technical reports, training materials, workshops—into their company intranet. The concept was sound. The execution wasn't. Some workers began calling search functionality their "random document generator." Eventually the system was re-engineered and now functions effectively.
HR professionals might question why intranet projects fall within their domain. But when internal communication constitutes part of your responsibilities—as it does for many HR departments—intranet systems become crucial infrastructure. Too often, HR remains excluded from these decisions, producing results that leave employees less productive and unable to locate needed information. Contributing meaningfully requires understanding knowledge management fundamentals.
The Productivity Cost of Poor Information Access
Without effective information retrieval, organizations become sluggish and operations slow down. Yet most companies approach this problem haphazardly. At any moment, employees might search corporate wikis, shared directories, email, or conversational systems. They learn through workshops, company meetings, or conversations with experienced colleagues—until those colleagues retire. HR is intimately involved in some systems while excluded from others. But fragmented, overlapping, disconnected environments cannot solve information access problems. Adding more tools proves easy but increases complexity rather than reducing it.
When knowledge workers complain about "information overload," they're actually experiencing "filter failure." They'd welcome correct answers, but poor information organization instead overwhelms them with irrelevant content that erodes productivity and increases costs.
Understanding Information Diversity
Fixing this requires first understanding breadth and diversity of information at large companies. Unstructured knowledge like internal blogs coexists with rigidly structured systems like managed digital asset collections. Low-value, unfiltered content collections like chat posts exist alongside highly curated, vetted repositories like approved methodologies and best practices. Narrowly focused content like engineering troubleshooting guides shares space with broadly applicable content like vision statements. Aggregating all this without proper access engineering produces "random document generators."
The Classification Challenge
Solutions require careful, deliberate tagging systems indicating which structured, high-value content applies to given problems. But tagging everything represents mammoth, endless work.
AI makes critical difference here. Using text analytics and starting from carefully tagged training data, machines identify patterns in documents. Resulting "auto-classifiers" create models representing concepts contained in documents, based on master classification structures called ontologies. When documents contain organizational legal names, terms like "policy," employee instruction lists, links, and HR staff authors, text analytics can confidently classify them as HR policies.
Machines sometimes err in classification. But including humans spot-checking classifications improves algorithms. Improved algorithms then quickly, accurately tag huge document collections, making them accessible in ways most helpful to employees seeking answers.
Organizations can fine-tune performance further by having machines observe how employees react to information they receive. Do they find it useful? Keep searching because it fell short? Give up and create their own answers? Call support lines? Analytics on intranet pages reveals where algorithms surface high-value content and where they need improvement.
By substantially and continuously improving classification and retrieval, AI can elevate company productivity to far higher levels. When HR engages with these improvements, it makes meaningful, measurable contributions to organizational performance.
This article was originally published on HR.com and has been revised for Earley.com.