As AI agents are taking on an increasing share of tasks once performed by human workers, organizations are facing a new reality.
That shift raises a fundamental question: Should AI agents be considered only as data-driven software, or should they be viewed as AI-powered workers, namely AI resources, analogous to human resources?
AI is a socio-technical system that extends beyond code and data, and reflects complicated social interactions between people and machines. AI agents operating in an agentic AI environment may introduce risks such as misinterpreting tasks, agent competition, error stacking and abuse of credentials. As a result, organizations need clearer ways to govern how AI-powered workers operate, and in particular, how they interact with people, specifically, human workers and customers.
AIRP (AI Resource Practices) is a conceptual framework that is based on the idea that AI-powered workers should be managed as an integral part of the organizational workforce. This reflects a new mindset of a human and AI-powered workforce.
AIRP adopts Human Resource Practices (HRP) as an organizing idea to define practices for governing and managing AI-powered workers, defined as AI-Resources (AIR). It outlines three areas of practice: workforce turnover, appraisal and reward and employee relations.
In practical terms, AIRP defines, for example, practices to answer:
- How do organizations count and manage their AI agents’ population and their overall workforce capacity?
- How are AI agents assigned or relocated within the organization?
- What guardrails are put in place to define agency and set the autonomy boundaries? To what extent are AI agents permitted to compete with one another or with human workers?
- What are the rules for AI agents’ coordination and collaboration?
- How does the organization respond if these guardrails are being breached?
- How are AI agents evaluated and appraised for their performance?
- What are mandatory steps to decommission an AI agent?
As a start, organizations should establish an integrated HR and AI-Resources governance function. By applying a combined HRP and AIRP approach, this function can effectively capture the intricacies of operating a hybrid human and AI-powered workforce. In doing so, organizations can responsibly realize the full potential of AI, while supporting strategy and alignment with organizational goals.
The question is no longer whether AI will be part of the workforce, but are we ready?
Are we ready to steer safely and confidently through the non-deterministic waters of a workforce made up of both humans and AI-powered workers? Do we have the right governance and practices in place?
AIRP is proposed as a practical approach for tackling this challenge.
For a deeper understanding of the AIRP framework read my “Future-Ready: An Agentic AI Governance Framework ISACA Journal article.
About the author: Noam Koriat is the director of information systems audit at Discount Bank. He previously served as global chief information officer (CIO) of the Israeli Ministry of Tourism. In addition to his executive roles, Koriat is an adjunct professor at the Graduate School of Business Administration, Bar Ilan University (Ramat Gan, Israel), where he teaches digital transformation and innovation, knowledge management, and the information systems practicum. Koriat can be contacted on LinkedIn at https://www.linkedin.com/in/noamkor/.
