The HR software company Phenom announced over 20 new AI agents available on its platform at its annual conference in Philadelphia last week. It released Ontologies, which is a category of strategic HR agents, and Agent Studio, which contains a variety of verticalized agents.
The Agent Studio offerings are calibrated for candidate sourcing, assessment and scheduling across industries such as health care, financial services, manufacturing and retail and for other HR strategy goals such as coaching and team management or onboarding. They are named for their industry and task area, such as Healthcare Sourcing Agent or Hospitality Skills Assessment Agent.
"I don't think there are any general agents that will work, in my opinion—verticalization is key," Mahe Bayireddi, CEO and co-founder of Phenom, said in a media and analyst session at the IAMPHENOM conference last week. "Our thought process is a nurse recruiter agent is entirely different from a financial analyst recruiter agent, which will be entirely different from a technical analyst recruiting agent."
The Ontologies offering is meant to be a bridge between HR and the broader business by collecting enterprise data for workforce planning insights around skills, productivity and employee personas, and it does financial modeling. Phenom said the goal of these insights is to reduce the lag in collection and communication of feedback informing HR and overall business strategy.
The Agent Studio also allows users to develop their own custom agents. They can start with existing templates, and the internal data from Ontologies will help configure other agents for internal use. The goal is to save time for recruiters and other HR staff and leaders. Recruiting has seen the most maturity in AI adoption within HR, according to industry experts.
Bayireddi emphasized that the underlying datasets used by general agentic AI offerings cannot adequately meet the needs of specific industries and functions given the roles' nuances as well as how those roles can be different at different organizations.
For example, recruiting agents are configured by data based on their location, such as unemployment rates, or industry information such as the talent supply and distribution for their industry or certain roles, Bayireddi said.
"If we don't really optimize for an individual persona [regarding] what impact it can really make, you cannot measure the outcome, and if that's not working for them, these products will fail," Bayireddi said, adding that the cost of compute power to deliver this customization has decreased rapidly, aiding in the development of this slew of new agents.
Phenom's vice president of product management, Kumar Ananthanarayana, shared that the company plans to launch more agent templates in April, and some clients have developed their own custom agents to support other functions such as customer service.
"We looked at use cases that drive the most effectiveness from an agentic workflow perspective," he said. "But there's tons of use cases."