Temus’ two primary current AI use cases involve talent acquisition and talent development. “For talent acquisition, we wanted to transform our hiring process and initially we focused on using AI to score job candidates and create a ranking list,” said Ang. “However, we realized this did not fit how hiring managers were making decisions, which was based on the favoring of particular characteristics, such as skills or seniority levels, at certain points in time.”

The business’s second iteration moved away from a quantitative approach to lean into generative AI’s unique capability of sorting and making sense of unstructured text information in resumes. “We began asking AI to summarize a candidate’s experience in three sentences, and the outcomes helped hiring managers prioritize as required,” explained Ang. “Now those managers drop resumes straight into our talent acquisition system to extract key information. Across 50-odd resumes, we have been able to reduce a review process that took three to four hours to just 10 minutes.”  

This process has also helped Temus target the best candidates more effectively and drive up the quality of hires.

“With Gemini and Vertex AI, we have the monetary value of the time saved, better quality recruitment and the improved engagement of hiring managers who, instead of going through the motions, are searching for the right candidate and going in to bat for them,” said Ang.

He believes the system will initially reduce manpower costs by $1 million-plus, based on a 90% reduction in resume processing time. This dollar figure is expected to rise as the business grows. The new system will also deliver considerable benefits to applicants for roles at Temus. Ang explains: “In the past, we would open a role, look at resumes, fill the role, and forget about the others who applied. A similar role may open up perhaps four months later and we would start the process again.”  

“With our new system, the information provided by everyone who applies for a role is already captured, analyzed and organized. The next time I need to fill a similar role, I have an existing pool of candidates. We are now putting in place processes whereby we stay in touch with candidates because they may have skills more relevant to another role, or upgraded their certifications or accreditations.” 

“Our talent development engine uses these two data sources to recommend material our people can use to move to the next level,” said Ang. “Where we ingest the data, we summarize this material and notes provided by people who have gone through the same process to deliver a living knowledge repository”. With an LLM layered on top and the right prompts, Temus’ people are guided through the process, with the system recording all of these actions. “Our managers can pick up this information and form teams from there, while our employees can step up their skills development,” shared Ang.

Ang has enjoyed working with the Google Cloud team on Temus’ AI projects. “Google Cloud is not afraid to be on the bleeding edge of technology and enables businesses like ours to play with new products and services with support from the pre-sales engineers,” he said. “The AI wave has barely gotten started and we’re looking forward to what is to come on the Google Cloud.”