Forum explores challenges in operations, HR, ERP and security

Moon Hong-ki, chief executive of PwC Consulting, delivers opening remarks at the "Building an AI Native Enterprise in Action" forum held Thursday at Amore Hall in Yongsan-gu, Seoul. [Samil PwC]
Moon Hong-ki, chief executive of PwC Consulting, delivers opening remarks at the "Building an AI Native Enterprise in Action" forum held Thursday at Amore Hall in Yongsan-gu, Seoul. [Samil PwC]

As AI evolves from a productivity tool into an active agent of business execution, companies that fail to make the AI-native transition risk being left behind, experts said Thursday.

PwC Consulting hosted the "Building an AI Native Enterprise in Action" forum Thursday at Amore Hall in Yongsan-gu, Seoul. More than 200 corporate representatives attended, covering key issues and action items for AI-native transformation — including business process redesign, human resources, enterprise resource planning, security and AI ecosystem strategy.

PwC Consulting Chief Executive Moon Hong-ki introduced the firm's new global consulting model, the Globally Integrated Consulting Model, or GICM. The GICM consolidates PwC's capabilities — previously dispersed across countries and business units — into a unified framework for client collaboration. It comprises five sectors, seven platforms and more than 160 service offerings, and draws on partnerships with global technology companies including AWS, Microsoft, Google, OpenAI, Anthropic, SAP and Palantir.

In the first session, Baek Jong-moon, leader of PwC Consulting's operations practice, outlined the concept of an AI-native enterprise — one that has redesigned its strategy, processes, data, technology architecture and overall organization around AI. He said executives must address six core areas: strategy, processes, data, technology and infrastructure, organizational governance, and culture and talent.

The second session featured partner Lee Jang-hyeok, who analyzed the causes of the AI performance gap and laid out an approach for achieving AI-native transformation. According to Lee, the top 20 percent of companies capture 74 percent of the economic value generated by AI. He said this gap stems not from differences in the scale of AI adoption or model performance, but from whether companies have redesigned their workflows to let AI operate effectively. "Workflows must be redesigned from the outset to end, leaving only the points where human intervention is truly necessary," Lee said.

In the third session, Ku Bon-jae, leader of the cloud and AI practice, introduced the "AI-Ready Knowledge" platform and the "PI (Process Innovation) Agent." The system enables autonomous AI to select and combine the skills needed for a given objective and carry out tasks in connection with enterprise systems. The AI-Ready Knowledge platform converts undocumented knowledge — such as the judgment criteria and exception-handling methods accumulated through frontline employees' experience — into knowledge that AI can understand and act on. The PI Agent engages in dialogue with frontline staff to map out workflows and decision criteria, then structures that information into minimum knowledge units, called AKUs, that AI can immediately use.

The fourth session, led by partner Kim Ju-su, examined what HR functions must prepare for as humans and digital workers increasingly operate side by side. He said HR departments should define AI agents as "digital employees" and establish a full lifecycle management framework covering onboarding, authorization, performance evaluation and decommissioning. Workforce planning criteria should also shift from headcount to task-based units, and new systems for performance management, career transition, and AI transformation competency training and assessment will need to be designed.

In the fifth session, partner Park Moon-jin presented an analysis of how ERP's role is changing in the AI era. He argued that ERP must evolve into the operational backbone that safely executes AI-driven decisions, ensuring corporate transactions are carried out under consistent rules. To that end, he proposed three approaches: designing business processes around application programming interfaces that AI can call directly; building data structures that allow tracing from management indicators down to individual transactions and documents; and connecting planning, execution and review into a single learning cycle.

Partner Jeon Sang-mi, who led the sixth session, outlined responses to three categories of AI security threats. The first involves AI being weaponized as an attack tool, which she said requires using AI itself to accelerate defense and vulnerability scanning. The second targets AI models directly and calls for preventive controls around the model — including authentication, access management and input-output guardrails. The third involves AI agents being exploited as attack vectors; she said agents must be managed under the same identity, authorization, approval and audit frameworks applied to human employees.

Jeon demonstrated PwC Consulting's AI-based hacking simulation service, in which AI probes and validates attack paths and proposes countermeasures. "Incidents involving AI agents may stem not from external intrusion but from the misuse of legitimately granted permissions," she said. "AI security must be treated not as a simple technical challenge but as a core management-level agenda item."

In the final session, partner Kim Beom-su identified four key technologies companies should watch: autonomous AI, physical AI, edge AI and AI trust and efficiency technology. He said companies should distinguish between building capabilities in-house and adopting external solutions based on the pace of change and the nature of the work, noting that rapidly evolving technologies are generally better served by external solutions. He also introduced PwC's "AI Agent Navigator," which helps companies identify and onboard AI agents suited to their needs, and demonstrated an AI internal audit platform that supports audit target selection, risk analysis, audit execution and reporting, and whistleblower investigation.

In his opening remarks, Moon said the time has come for companies to make the AI-native transition, given advances in technology and the need to demonstrate return on investment. "We hope to help companies draw the big picture of their AI transformation by presenting strategies and practical approaches for AI-centered business operations," he said.

Meanwhile, PwC Consulting posted sales of 463.7 billion won ($344 million) for fiscal year 2026 — covering July 2025 through June 2026 — up 4 percent from the same period a year earlier. The firm has been expanding customized services in areas including AI-driven system and process innovation, digital management advancement, risk response and business strategy. Over the same period, Samil PwC recorded total sales of 1.63 trillion won and operating profit of 19.8 billion won.


park.jiyeong@heraldcorp.com