Finding the right person for the job is no walk in the park, as any recruiter will tell you. And it’s at least as challenging for job seekers to wade through the opportunities, apply for a job, and get that coveted first interview.
But iCIMS believes generative AI can help. Today, the company is announcing its own iCIMS Copilot, an AI assistant powered by the GPT-4 in Azure OpenAI Service introduced earlier this year by Microsoft that will live as a persistent, easily accessible chatbot inside the existing iCIMS Talent Cloud recruiting software, allowing recruiters to instantly ask it questions, prompt it to generate job descriptions and interview questions, draft candidate offer letters, and create marketing content around job opportunities.
“If you’re at a point in using our products where you need to create content over and over, let the Copilot give you a starting point,” said iCIMS CTO Al Smith in an exclusive video interview with VentureBeat.
Whatever data recruiters have introduced to their secure and private iCIMS Talent Cloud instances can be drawn upon and leveraged for use by the Copilot, but users can also prompt it with their own outside data.
iCIMS is no stranger to AI. The company headquartered in Holmdel, New Jersey, was an early leader at bringing software to the recruiting journey, having been founded in 2000, and has since grown to service more than 6,000 customers across sectors, specializing in those needing a high volume of new hires and those with high turnover, such as retail outlets and seasonal businesses.
For the last five years, even long before the current generative AI gold rush, it has pursued integrating what it calls “responsible AI” with its platform, and has “worked very, very hard around the world,” according to Smith, to fulfill the various compliance requirements where it and its enterprise customers do business.
So far, iCIMS has integrated responsible AI into its platform in the form of “AI-led search, match, and job fit recommendation capabilities to candidates,” including routing and sorting job candidates who are both the “best fit” and a “close match” to the job description. iCIMS also takes into account hard and “soft,” or derived skills when evaluating candidates to see who could grow into organizations to take on more roles.
Eliminating bias is especially critical for recruiting, where laws prohibit technology discrimination and bias in hiring. For example, Smith told VentureBeat that iCIMS is compliant with New York City’s recently enforced NYC 144 law.
“We have a big belief that AI is not this black box that says ‘bing, here’s your answer!’” explained Smith. “It should be helping you make better contextual decisions and be explainable on why it’s helping you provide that information.”
However, Smith acknowledges this explainability is a challenge with generative AI solutions like its own new iCIMS Copilot. Large language models (LLMs) such as OpenAI’s GPT-4, which powers the iCIMS Copilot, are great at delivering fast answers and content on demand, but not necessarily at explaining how they arrived at those responses.
That’s why iCIMS has worked to introduce fine-tuning and safeguards around its implementation of generative AI, and views it as a recruiting aid, especially for content creation, not as a replacement for human recruiters.
Generative AI assistants can help “eliminate like the undifferentiated work that anybody could do, and let you focus more on the outcomes that you’re trying to do,” Smith said, referring to his recruiter customers.
As VentureBeat has previously remarked, it is still early innings in the metaphorical baseball game of enterprises competing to figure out the best applications for generative AI.
iCIMS believes its Copilot is a valuable addition, but plans to offer users the chance to give feedback continuously and at multiple points through their usage.
“Honestly, I want to learn ‘what is the definition of high quality in this technology?’” said Smith. “I want feedback that when you use it, you can choose options that say ‘this was really helpful, somewhat helpful, or not helpful at all.’”
As iCIMS collects feedback from its enterprise users, it will use it to iterate and refine its Copilot, as well as add new capabilities and bring new generative AI products and tools to market.
As such, “we’re still trying to figure out the monetization side of it,” said Smith. Right now, the iCIMS Copilot will be bundled in with its regular Talent Cloud, which charges on an “enterprise pricing model,” not per user.
VentureBeat presents: AI Unleashed – An exclusive executive event for enterprise data leaders. Network and learn with industry peers. Learn More
Finding the right person for the job is no walk in the park, as any recruiter will tell you. And it’s at least as challenging for job seekers to wade through the opportunities, apply for a job, and get that coveted first interview.
But iCIMS believes generative AI can help. Today, the company is announcing its own iCIMS Copilot, an AI assistant powered by the GPT-4 in Azure OpenAI Service introduced earlier this year by Microsoft that will live as a persistent, easily accessible chatbot inside the existing iCIMS Talent Cloud recruiting software, allowing recruiters to instantly ask it questions, prompt it to generate job descriptions and interview questions, draft candidate offer letters, and create marketing content around job opportunities.
“If you’re at a point in using our products where you need to create content over and over, let the Copilot give you a starting point,” said iCIMS CTO Al Smith in an exclusive video interview with VentureBeat.
Whatever data recruiters have introduced to their secure and private iCIMS Talent Cloud instances can be drawn upon and leveraged for use by the Copilot, but users can also prompt it with their own outside data.
Event
AI Unleashed
An exclusive invite-only evening of insights and networking, designed for senior enterprise executives overseeing data stacks and strategies.
An AI evolution for recruiters
iCIMS is no stranger to AI. The company headquartered in Holmdel, New Jersey, was an early leader at bringing software to the recruiting journey, having been founded in 2000, and has since grown to service more than 6,000 customers across sectors, specializing in those needing a high volume of new hires and those with high turnover, such as retail outlets and seasonal businesses.
For the last five years, even long before the current generative AI gold rush, it has pursued integrating what it calls “responsible AI” with its platform, and has “worked very, very hard around the world,” according to Smith, to fulfill the various compliance requirements where it and its enterprise customers do business.
So far, iCIMS has integrated responsible AI into its platform in the form of “AI-led search, match, and job fit recommendation capabilities to candidates,” including routing and sorting job candidates who are both the “best fit” and a “close match” to the job description. iCIMS also takes into account hard and “soft,” or derived skills when evaluating candidates to see who could grow into organizations to take on more roles.
The importance of avoiding bias
Eliminating bias is especially critical for recruiting, where laws prohibit technology discrimination and bias in hiring. For example, Smith told VentureBeat that iCIMS is compliant with New York City’s recently enforced NYC 144 law.
“We have a big belief that AI is not this black box that says ‘bing, here’s your answer!’” explained Smith. “It should be helping you make better contextual decisions and be explainable on why it’s helping you provide that information.”
However, Smith acknowledges this explainability is a challenge with generative AI solutions like its own new iCIMS Copilot. Large language models (LLMs) such as OpenAI’s GPT-4, which powers the iCIMS Copilot, are great at delivering fast answers and content on demand, but not necessarily at explaining how they arrived at those responses.
A recruiting aid, not a replacement
That’s why iCIMS has worked to introduce fine-tuning and safeguards around its implementation of generative AI, and views it as a recruiting aid, especially for content creation, not as a replacement for human recruiters.
Generative AI assistants can help “eliminate like the undifferentiated work that anybody could do, and let you focus more on the outcomes that you’re trying to do,” Smith said, referring to his recruiter customers.
Early innings
As VentureBeat has previously remarked, it is still early innings in the metaphorical baseball game of enterprises competing to figure out the best applications for generative AI.
iCIMS believes its Copilot is a valuable addition, but plans to offer users the chance to give feedback continuously and at multiple points through their usage.
“Honestly, I want to learn ‘what is the definition of high quality in this technology?’” said Smith. “I want feedback that when you use it, you can choose options that say ‘this was really helpful, somewhat helpful, or not helpful at all.’”
As iCIMS collects feedback from its enterprise users, it will use it to iterate and refine its Copilot, as well as add new capabilities and bring new generative AI products and tools to market.
As such, “we’re still trying to figure out the monetization side of it,” said Smith. Right now, the iCIMS Copilot will be bundled in with its regular Talent Cloud, which charges on an “enterprise pricing model,” not per user.
VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.
Author: Carl Franzen
Source: Venturebeat
Reviewed By: Editorial Team
