Wauwatosa-based talent acquisition firm Cielo has launched a new Agentic AI tool aimed at streamlining the hiring process for professionals, allowing them to focus on more complicated tasks.
The newest offering within Cielo’s artificial intelligence-powered technology suite, the Agentic AI tool works almost like its own agent, completing a predetermined set of tasks without needing constant prompts from humans.
About five years ago, Cielo’s product team made the decision to start developing its own software. When ChatGPT launched in late 2022, Cielo began leveraging the technology to develop its own generative AI models.
Cielo had been using its generative AI agents for about a year and a half when the next iteration of the technology was released: Agentic AI. Agentic AI is more of a proactive technology, where generative AI is reactive.
“While we didn’t set out to become the next enterprise business platform, we recognized that we’re the proven global specialists on talent acquisition and if we started to build some of our own software, we could really take control of our own roadmap, while also creating compelling solutions for other businesses,” said Matt Jones, chief product officer at Cielo.
Cielo’s Agentic AI agents act independently, learning from past performance while combining artificial intelligence with human expertise to enhance worker productivity. The agents can help with tasks ranging from employer branding and talent marketing to gathering workforce insights and supporting recruiting efforts.
For example, an agent might be used to help create hyper-personalized advertising or develop inclusive language for copywriting.
“It really allows us as humans to give a set of predetermined tasks to AI agents to go out and complete,” said Jones. “They’re able to actually do things within the digital world compared to the ‘call and response’ point solutions with some AI (platforms).”
While Agentic AI does not require constant prompts from a person, the technology still reports back to its user and is not used for any final decision-making. Agentic AI operates “within a set of guardrails,” Jones explained.
“We’re elevating humans from doing research, preparation or administrative tasks; things that are foundational but not value-add, to allow them to spend more time on higher-value, higher-thinking work,” he said.
Cielo’s Agentic AI agents can also complete tasks while human employees are off the clock. This could be useful for businesses in the health care industry or consumer brand companies, which often receive feedback and customer concerns all day.
“In a time when we’re constantly hearing ‘do more with less,’ this is a refreshing opportunity to ‘do more with more,’” said Rebecca Volpano, vice president of product marketing at Cielo. “Agentic AI goes beyond reducing administrative burden and improving efficiency; it truly serves as a much-needed addition to any recruiting team.”
Cielo is currently working on the next iteration of its Agentic AI tool, which will involve the creation of a concierge agent. Eventually, a Cielo client using the company’s Agentic AI tool might end up having dozens of agents working behind the scenes to complete research and other tasks.
“The concierge agent manages a workflow while talking to other agents to carry out specific activities, later reporting back and queuing humans at critical milestones,” said Jones.
Cielo has nine employees on its data science and technical product teams who are continually working to upgrade the company’s suite of AI products.
Preventing bias and possible legal challenges is always a concern when AI is used in the talent acquisition or hiring process. To prevent bias within its Agentic AI tool, Cielo has a team regularly come in to audit its security and privacy policies. All of the company’s software is also ISO certified. Jones emphasized the fact that the Agentic AI tool is never used to make any final decision regarding a candidate’s employment.
Any company using AI as part of its talent acquisition process must also disclose that fact to candidates. Typically, companies use privacy policies to do so.
“We’re using the major model providers such as GPT-4, Claude 3.5 and 3.7,” said Jones. “While they can’t guarantee there’s no inherent bias, because it’s trained on public data, we know that we’re taking models that have some level of safety and then bringing our own ethical use of AI standards as well through our configuration and testing.”
As AI agents evolve, Cielo will continue to provide clients with new iterations of its technology. Cielo routinely tests more than 30 AI models in its lab, including those from Meta, OpenAI, Anthropic and Mistral, and plans to deploy multi-agent systems to tackle and audit even more complex recruiting challenges.
“The analogy I use is, when was the last time you called Amazon? I don’t call Amazon. I use the technology. I love it. It makes my life better. That’s how we think about deploying agentic AI,” said Jones. “We’re not trying to hide the fact that we’re using AI from people. We’re deploying AI that makes better human experiences.”
