More HR Teams Use AI, but Results Stall: Organizational Context Is Key

As AI use spreads across companies, more HR departments are moving to automate repetitive tasks and make HR decisions more objective. Human Consulting Group (HCG), however, finds that actual use still remains at a surface level, such as individual tasks. To embed AI in HR work and expand productivity, it says, companies need systems that capture the organization's own decision criteria and context.

Surveys in Korea and abroad show the same trend. In a survey of 259 HR professionals in Korea conducted by Daou Office HR in March this year, 83.4% of respondents said they use AI at work, but only 24.7% had gone as far as company-wide adoption or automation of repetitive tasks. The picture is similar overseas. In Culture Amp's '2026 AI in HR' study released in July, HR professionals' AI use was concentrated on individual tasks such as content writing (86%), while 39% had applied it to automating HR operations. And according to McKinsey's 'The State of AI in 2026,' released in August, 37% of companies said AI had affected their operating profit (EBIT), and only about 6% were seeing meaningful financial results. There is still a gap between individual AI use and work transformation at the organizational level.

HCG attributes this gap between AI adoption and actual results to the particular nature of HR work. HR decisions are shaped not only by each company's rules but also by its own organizational culture and the tacit knowledge its members share. A question such as 'By what criteria should we evaluate performance?' can only be answered by taking into account the data and cultural context accumulated inside the organization.

Defining competencies is no different: each organization expects different behaviors. Faced with the word 'leadership,' one organization values quick decisions while another values the ability to build consensus. That is because each company aims for a different talent profile and a different formula for success. Performance ratings and comments, too, take on different meanings depending on an organization's practices and culture. In finding and placing talent, factors beyond career history and performance records, such as trust among colleagues, ways of working, and fit with a leader, shape who is judged the right person. In one-on-one meetings, too, the context that matters most often stays outside the conversation notes. This is why general-purpose AI trained only on documented rules struggles to fully grasp the organizational context that HR decisions require. It is also why AI that sees only recorded data produces 'answers that aren't wrong, but don't fit our company.'

Organizational culture and tacit knowledge do not accumulate in a system on their own. To trust AI's recommendations in key HR decisions such as evaluation and placement, organizations need evidence data that shows the criteria they have used to make decisions.

As a solution, HCG proposes a system that builds up organizational context, together with the HR know-how to understand what that system should hold. First, the system must record not only performance ratings and placement outcomes but also the criteria and discussions that led to those conclusions. HCG explains that only by recording the feedback exchanged in meetings and the changes that followed, and by linking the reasons for placement decisions to their outcomes, can a foundation be laid for AI to understand how the organization makes HR decisions. A system alone, however, is not enough. Technology alone cannot decide at what behavioral level to define a competency, what evidence to keep alongside evaluation data, or what to record in placement decisions. Only with HR expertise that knows how policies actually work in real organizations does the data accumulating in the system finally gain context.

Demand in the HR field points in the same direction. The AI that companies want is closer to a partner that supports work on top of the organization's decision criteria and context. In payroll and attendance operations, typical roles include finding anomalous data after closing and pinpointing the cause, or flagging overtime before it exceeds the plan. For work to be processed accurately, rules that differ by company and by labor union must be reflected.

In decision support, AI is expected to forecast the results of changes to evaluation criteria in advance, and to compare, with supporting evidence, organization and workforce plans aligned with the business plan or candidates for a role. AI provides the information needed for analysis and comparison, while leaders and HR managers make the final decisions.

When handling employee inquiries and requests, too, the organization's rules and authority structure serve as the standard. AI must answer within the scope of each employee's own permissions and present the grounds in company policy and labor law alongside its answers. In talent and career development, the key is to reflect individual skill gaps and the growth paths the organization expects.

The common thread in field demand is clarity about AI's role and the grounds for its judgments. Organizations must separate the work they hand to AI from the points where people decide, and explain results and their grounds according to the organization's own standards. This is why companies need a system that accumulates HR data and organizational context, and AI that uses it to get work done.

For 25 years, HCG has carried out both HR consulting that designs companies' HR policies and HR system implementation that runs those policies. Drawing on this experience across both policy design and the accumulation of operational data, HCG is setting out the conditions for using AI in a way that reflects organizational context. HCG expects that HR's competitiveness in AI will be decided by structures that accumulate an organization's decision criteria and experience together with their context. Companies' concerns, too, are moving beyond simply whether to adopt AI, toward what difference a dedicated AI that deeply understands HR can make. HCG will host its HR AI conference 2026, 'GROUND ONE,' on October 28 at the Harmony Ballroom of The Westin Seoul Parnas in Samseong-dong, Seoul, where it plans to present new standards for AI x HR operations.

"HR data does not gain its full meaning from numbers alone. Even the same performance rating has to be read within that organization's standards and culture," said Soonwon Hong, Head of the AX Consulting Division at Human Consulting Group. "Only when HR expertise that knows what context to keep as data comes together with a system designed to accumulate that context can AI properly take on the work of HR."