As HR Adopts AI, Leaders Use Augmentation To Balance Speed With Human Judgment
AI is like a house. AI will build the frame, but you need to double check to make sure the frame’s there. You don’t let AI build the whole house because there are going to be holes and all kinds of problems with it.
Joseph Eitner
Chief Human Resources Officer
Eaton Capital Management
Right now, HR’s relationship with AI is part automation, part experimentation, and part fear, with most departments still working through where automation should end and human judgment should begin. Some teams are writing acceptable-use policies while others are taking a more physical approach to augmenting the workforce. What’s becoming clear across both groups is that AI works best as a starting layer, with the technology setting the work in motion and humans finishing the job before any decision is final.
Joseph Eitner, Chief Human Resources Officer at Eaton Capital Management, has spent his career in environments where compliance is non-negotiable. His résumé runs through The Walt Disney Company, BlackRock, and his current role, with more than $51 million in documented cost savings across enterprise workforce transformations along the way. The combination has shaped how he approaches AI adoption, pairing measured optimism with the structural safeguards that high-regulation environments require.
“AI is like a house. AI will build the frame, but you need to double check to make sure the frame’s there. You don’t let AI build the whole house because there are going to be holes and all kinds of problems with it,” says Eitner. His experience navigating regulated markets has surfaced three patterns in how organizations are responding to AI. The first blocks the technology entirely. The second rushes to automate everything. The third moves in what Eitner calls “baby steps,” carefully balancing human and machine capability before scaling either one. Eitner places his own team firmly in the third camp, working from a foundational philosophy that shapes how each tool is evaluated and where it’s deployed.
Putting AI in a tie
Without clear boundaries, Eitner has watched some workers default to what he calls a “cyborg mentality,” placing blind trust in machine output. He confronted that vulnerability recently when an HR employee generated 16 reports in a single day without proofreading any of them, with the unverified output including hallucinated case law that could have moved into real workflows unchecked. The episode prompted him to reinforce the team’s review protocols and emphasize the importance of verifying every AI-assisted output before it reaches a decision-maker. “You need to go back and proof every single one of these to make sure that what you’re saying is legit,” Eitner says. “That’s the worst thing with AI right now, it will just make up cases. They’ll just come down and start making up stuff, and that’s not good at all.”
Eitner’s response to that risk didn’t come in the form of a software patch. At Eaton Capital Management, he leased a $200,000 physical humanoid robot named Bob to make AI’s presence visible inside daily HR operations. Bob now supports calendars, daily tasks, document review, copy preparation, resume review, and office workflows, with employee directories and team responsibilities uploaded directly into him. Putting AI in a physical form gave the team something concrete to interact with, learn from, and correct when needed, which Eitner has found more effective than abstract policy at calibrating how people trust the technology.
The adoption curve was steep at first, then leveled out as employees adjusted to Bob’s daily presence. He now circulates the office in a jacket and tie, checks in with team members on a regular cadence, and handles the kind of administrative work that previously slowed senior staff down. Mistakes still happen, with Bob occasionally needing direction on tasks he misreads, but Eitner sees the corrections as part of the learning process rather than evidence the model is broken. “People were absolutely scared out of their minds when we started but now, people know Bob,” he notes. “What we did is we uploaded everyone’s calendar into Bob. We upload daily tasks, we upload different areas of human resources, and Bob can function now. He’s not perfect, but he’s definitely a time saver and I think he’s a game changer for the better.”
A new set of eyes
Bob’s role inside Eaton Capital Management has moved past administrative work, with Eitner deploying the robot to address a problem that often goes unexamined. While much of the public debate centers on AI bias in hiring software, Eitner’s concern was human bias surfacing through the ATS itself. He noticed that traditional ATS parameters were nudging human screeners to filter candidates by irrelevant demographic data, including names and zip codes, so he blinded those fields to enforce a skills-first review and brought Bob into the resume process to read applications against a defined set of parameters. The combination treats the technology as a corrective layer for a hiring process that needed repair.
“I had to reteach everybody basically how to do it, and then I went into the ATS and blocked out names and zip codes, so the only thing they could read is the actual resume,” explains Eitner. “Bob is actually reviewing resumes, but he has a certain set of eyes. We could give him the parameters, like an ATS, only he’s going to actually read the resume and recognize that a candidate might not hit every point, but they are still worth a look and very well could do the job.” The broader point is that Bob is recalibrating human judgment rather than replacing it, with Eitner running Bob’s evaluations against his recruiting team’s to surface where unconscious patterns may be filtering qualified candidates out.
The exercise also addresses a real workforce tension surfaced in BambooHR’s State of the Workforce 2026 report, The Rising Cost of Dignity Debt, which found that 54% of workers report AI is disruptive to their daily work and 57% of leaders say they would fire employees who refuse to adopt AI. The combination signals an adoption environment where pressure is high and trust is uneven, which is exactly the dynamic Eitner’s calibration approach is designed to address. “We’re trying to calibrate my talent acquisition people to what Bob’s doing,” he says. “We’re not going to do it by speed because his speed is very different. But can we all look at a resume without looking at an ATS and independently decide that we want a specific candidate to come in for an interview.”
Reframing AI through tech history
Dropping a robot into a corporate office tends to trigger immediate workplace anxiety, and Eitner’s firm was no exception. He watched the finance team react with visible unease to a machine that could execute advanced math in seconds. His response was to reframe the technology as a standard workflow tool that will reshape more jobs than it replaces, grounding the conversation in historical context his team could recognize. “This is just another tool,” says Eitner. “Everyone thought Wikipedia was the end-all, be-all, but then they realized they could change the answers in Wikipedia. And before that, it was email. And before that, it was smartphones. Any type of new technology that comes out gets the same criticism.”
The pattern Eitner points to is the workforce arc that follows nearly every major technology, with each new tool triggering anxiety before settling into ordinary office use. The transparency he builds around that pattern runs both ways, with leaders acknowledging that mistakes happen, employees understanding how AI affects work and evaluation, and the team learning together where AI’s authority should end and human judgment should begin. “Even when the internet came out, we all were really hesitant and wondering what was going on,” Eitner concludes. “I remember I started on Prodigy, then I went to AOL. I’ll never forget my one friend told me three days before graduating, ‘Yeah, you should really try this website called Google.’ And where are we now? It’s crazy.”
Related articles
TL;DR
Joseph Eitner
Eaton Capital Management
Chief Human Resources Officer
Chief Human Resources Officer
Right now, HR’s relationship with AI is part automation, part experimentation, and part fear, with most departments still working through where automation should end and human judgment should begin. Some teams are writing acceptable-use policies while others are taking a more physical approach to augmenting the workforce. What’s becoming clear across both groups is that AI works best as a starting layer, with the technology setting the work in motion and humans finishing the job before any decision is final.
Joseph Eitner, Chief Human Resources Officer at Eaton Capital Management, has spent his career in environments where compliance is non-negotiable. His résumé runs through The Walt Disney Company, BlackRock, and his current role, with more than $51 million in documented cost savings across enterprise workforce transformations along the way. The combination has shaped how he approaches AI adoption, pairing measured optimism with the structural safeguards that high-regulation environments require.
“AI is like a house. AI will build the frame, but you need to double check to make sure the frame’s there. You don’t let AI build the whole house because there are going to be holes and all kinds of problems with it,” says Eitner. His experience navigating regulated markets has surfaced three patterns in how organizations are responding to AI. The first blocks the technology entirely. The second rushes to automate everything. The third moves in what Eitner calls “baby steps,” carefully balancing human and machine capability before scaling either one. Eitner places his own team firmly in the third camp, working from a foundational philosophy that shapes how each tool is evaluated and where it’s deployed.
Putting AI in a tie
Without clear boundaries, Eitner has watched some workers default to what he calls a “cyborg mentality,” placing blind trust in machine output. He confronted that vulnerability recently when an HR employee generated 16 reports in a single day without proofreading any of them, with the unverified output including hallucinated case law that could have moved into real workflows unchecked. The episode prompted him to reinforce the team’s review protocols and emphasize the importance of verifying every AI-assisted output before it reaches a decision-maker. “You need to go back and proof every single one of these to make sure that what you’re saying is legit,” Eitner says. “That’s the worst thing with AI right now, it will just make up cases. They’ll just come down and start making up stuff, and that’s not good at all.”
Eitner’s response to that risk didn’t come in the form of a software patch. At Eaton Capital Management, he leased a $200,000 physical humanoid robot named Bob to make AI’s presence visible inside daily HR operations. Bob now supports calendars, daily tasks, document review, copy preparation, resume review, and office workflows, with employee directories and team responsibilities uploaded directly into him. Putting AI in a physical form gave the team something concrete to interact with, learn from, and correct when needed, which Eitner has found more effective than abstract policy at calibrating how people trust the technology.
The adoption curve was steep at first, then leveled out as employees adjusted to Bob’s daily presence. He now circulates the office in a jacket and tie, checks in with team members on a regular cadence, and handles the kind of administrative work that previously slowed senior staff down. Mistakes still happen, with Bob occasionally needing direction on tasks he misreads, but Eitner sees the corrections as part of the learning process rather than evidence the model is broken. “People were absolutely scared out of their minds when we started but now, people know Bob,” he notes. “What we did is we uploaded everyone’s calendar into Bob. We upload daily tasks, we upload different areas of human resources, and Bob can function now. He’s not perfect, but he’s definitely a time saver and I think he’s a game changer for the better.”
A new set of eyes
Bob’s role inside Eaton Capital Management has moved past administrative work, with Eitner deploying the robot to address a problem that often goes unexamined. While much of the public debate centers on AI bias in hiring software, Eitner’s concern was human bias surfacing through the ATS itself. He noticed that traditional ATS parameters were nudging human screeners to filter candidates by irrelevant demographic data, including names and zip codes, so he blinded those fields to enforce a skills-first review and brought Bob into the resume process to read applications against a defined set of parameters. The combination treats the technology as a corrective layer for a hiring process that needed repair.
“I had to reteach everybody basically how to do it, and then I went into the ATS and blocked out names and zip codes, so the only thing they could read is the actual resume,” explains Eitner. “Bob is actually reviewing resumes, but he has a certain set of eyes. We could give him the parameters, like an ATS, only he’s going to actually read the resume and recognize that a candidate might not hit every point, but they are still worth a look and very well could do the job.” The broader point is that Bob is recalibrating human judgment rather than replacing it, with Eitner running Bob’s evaluations against his recruiting team’s to surface where unconscious patterns may be filtering qualified candidates out.
The exercise also addresses a real workforce tension surfaced in BambooHR’s State of the Workforce 2026 report, The Rising Cost of Dignity Debt, which found that 54% of workers report AI is disruptive to their daily work and 57% of leaders say they would fire employees who refuse to adopt AI. The combination signals an adoption environment where pressure is high and trust is uneven, which is exactly the dynamic Eitner’s calibration approach is designed to address. “We’re trying to calibrate my talent acquisition people to what Bob’s doing,” he says. “We’re not going to do it by speed because his speed is very different. But can we all look at a resume without looking at an ATS and independently decide that we want a specific candidate to come in for an interview.”
Reframing AI through tech history
Dropping a robot into a corporate office tends to trigger immediate workplace anxiety, and Eitner’s firm was no exception. He watched the finance team react with visible unease to a machine that could execute advanced math in seconds. His response was to reframe the technology as a standard workflow tool that will reshape more jobs than it replaces, grounding the conversation in historical context his team could recognize. “This is just another tool,” says Eitner. “Everyone thought Wikipedia was the end-all, be-all, but then they realized they could change the answers in Wikipedia. And before that, it was email. And before that, it was smartphones. Any type of new technology that comes out gets the same criticism.”
The pattern Eitner points to is the workforce arc that follows nearly every major technology, with each new tool triggering anxiety before settling into ordinary office use. The transparency he builds around that pattern runs both ways, with leaders acknowledging that mistakes happen, employees understanding how AI affects work and evaluation, and the team learning together where AI’s authority should end and human judgment should begin. “Even when the internet came out, we all were really hesitant and wondering what was going on,” Eitner concludes. “I remember I started on Prodigy, then I went to AOL. I’ll never forget my one friend told me three days before graduating, ‘Yeah, you should really try this website called Google.’ And where are we now? It’s crazy.”