When you ask Jim Dobinski how many employees work for his digital marketing agency StellarBlue.ai, he happily responds with, “15 humans and 63 AI agents.”
It’s a revealing answer, if not a slightly dystopian one, that reflects the evolution of the 21-year-old agency. Last year, Stellar Blue Technologies rebranded as StellarBlue.ai, the public-facing culmination of the agency’s three-year internal shift to leverage artificial intelligence not only for its clients, but in its own operations.
Nowhere can this shift be seen more clearly than in StellarBlue.ai’s hiring process, which over the last six months has been almost entirely automated using AI agents — performing tasks that previously required humans.
“We have AI do the majority of the processing for us now,” says Dobinski, StellarBlue.ai founder and managing director. “It’s changed the way we hire.”
What used to take three people three days to process is now managed solely by Dobinski, with the help of several AI agents, in a fraction of the time.
“We shed 24 hours per hire just in the processing,” says Dobinski, who has hired five people since implementing AI agents. “The time savings, you can’t compete with that.”
These days, AI agents create StellarBlue’s job postings based on previous listings the agency created, posts them to LinkedIn, monitors applications, downloads and analyzes resumes, and whittles down the top few candidates from more than 200 applicants in as little as four minutes. An AI agent then sends personalized emails to those selected to interview with Dobinski, who still handles the interviews — for now.
“Ultimately it’s going to be completely automated. We’re going to have avatars do the actual interviewing, record the sessions and then other avatars analyze the sessions and [select whom] we should hire,” he says. “That’s just the next step of the full workflow that will be built out by the end of this year.”
Dobinski says the quality of candidates has also increased as AI agents can more efficiently identify top talent.
Despite these benefits, not every business is comfortable implementing that much automation into its hiring process.
Kristen Jonas, owner of Kinesiology HR in Green Bay, has been in human resources for 25 years and remembers when job seekers had to make the effort to submit resumes by mail. With automated application systems, applying for jobs has become so easy that it often results in a high volume of low quality applicants that businesses struggle to manage.
Jonas says this creates a two-sided problem — employers can’t identify genuine candidates; candidates often don’t hear back from employers, but still see the “help wanted” sign hanging in the proverbial window.
“When we talk about building relationships in a pipeline, it starts with getting to know candidates — not just volumes of applications,” says Jonas, who sees over-reaching automation as a threat to businesses long term. “The relationship that you’re trying to build with a candidate pipeline is just completely absent.”
Jonas acknowledges it may sound “old school,” but she believes the path to better hiring outcomes is through investments in relationships, community presence and thoughtful candidate communication.
“If you’re a community-focused organization, you definitely want to think about your reputation in terms of how you have treated your candidates,” she says. “You’ve got to put forth a little bit more effort on the communication with the candidates.”
Jonas says that time spent on relationships yields better long-term hiring results, so even if you can’t respond to all applicants immediately, simply acknowledge receipt of their application and provide rejection emails once candidate screening begins.
For businesses seeking to automate some (or all) of their hiring process, Dobinski recommends starting small. Begin by automating simple tasks like writing job descriptions or creating emails, and build confidence through incremental steps.
Patience is key, Dobinski says. It can take three to six months to train custom AI agents who will make mistakes initially but, unlike humans, never forget their training.
“They’re absolutely going to make mistakes — don’t expect perfection right away. But once they’re trained,” he says, “now you’ve got employees for life.”
