AI Saves One HR Leader 20 Hours a Week Because She Knows Exactly What to Keep for Herself
AI is not going to handle my judgment calls. It’s not going to take my job for me. I have been an HR person for 25 years, and it is not going to handle my judgment.
Diana Alvis
VP of Human Resources
Crown Capital Management
The productivity gains from AI in HR are real, but they only show up when the person using the tools knows what to automate and what to protect. An HR leader who feeds everything into an AI system without boundaries ends up fact-checking every output and creating more work instead of less. One who draws a clear line between the tasks AI handles well and the judgment calls it cannot touch gets back 15 to 20 hours a week and puts that time toward the work that actually requires a human.
Diana Alvis is Vice President of Human Resources at Crown Capital Management, where she built the full HR infrastructure for a 600-plus employee, multi-entity construction organization during rapid statewide growth. She holds an SPHR certification, has over 20 years of experience across construction, healthcare, legal, retail, and public sector HR, and handles everything from recruiting and compensation design to compliance, benefits, and bilingual field communications. She uses AI extensively in her daily workflow but is deliberate about where it stops.
“AI is not going to handle my judgment calls,” Alvis says. “It’s not going to take my job for me. I have been an HR person for 25 years, and it is not going to handle my judgment.”
Where AI earns its time
Alvis estimates AI saves her 15 to 20 hours per week across a set of bounded, specific use cases. The gains are practical rather than transformational. She is currently migrating between payroll systems and uses AI to identify data gaps, mismatches, and missing fields across spreadsheets. “It’s able to analyze data at light speed,” Alvis says. “It would have taken me so much longer on my own.” She limits what goes in, using employee ID numbers rather than full employee records to keep sensitive data out of the system.
The most surprising win was deduction code cleanup. AI found the most commonly used codes across her system, standardized them, and resolved inconsistencies she had not been able to untangle manually. “I was amazed how quickly it was just cleaning up deduction codes in the background,” she says. “I said, oh, wow. That was fast.”
Crown Capital employs a large Spanish-speaking field workforce, and Alvis uses AI to translate and condense messages for MMS delivery simultaneously. “I can ask it to minimize something for an MMS message and translate it at the same time. I was never able to do that before.” She has also built a custom AI agent trained on her communication style to help polish day-to-day email responses. “It’s almost like I have an assistant helping me respond to things,” Alvis says.
Where she draws the line
Alvis is equally specific about what stays with her. AI does not make judgment calls. It does not evaluate situations with the contextual understanding that managers expect from HR. And it does not get access to data it does not need.
She configures her settings to require source citations so she can verify where information comes from. “We all know not everything on Wikipedia is factual,” she says. “You have to be able to fact-check your data.” Certain compliance-sensitive information never enters the system at all. Even on a private AI instance, Alvis limits inputs to single data points rather than complete employee records. “There are certain things compliance-wise that just don’t go in,” she says.
She also warns that newer HR professionals who lean on AI for judgment calls risk losing the skills that make them effective. “You can become too reliant on anything,” Alvis says. “You have to remember to exercise your brain. You have to remember to think for yourself.” BambooHR’s State of the Workforce 2026 report, The Rising Cost of Dignity Debt, found a similar gap: leaders report productivity gains from AI while employees report rising stress, a divide the report ties to unclear boundaries around how AI gets used.
The result is that Alvis gets the productivity of what would otherwise require additional headcount while keeping the human decision-making that HR exists to provide. AI polishes the communication, audits the data, and translates the message. The judgment about what to do with that information, how to advise the manager, and how to protect both the employee and the organization stays with the person who has 25 years of context that no model can replicate.
Related articles
TL;DR
Diana Alvis
Crown Capital Management
VP of Human Resources
VP of Human Resources
The productivity gains from AI in HR are real, but they only show up when the person using the tools knows what to automate and what to protect. An HR leader who feeds everything into an AI system without boundaries ends up fact-checking every output and creating more work instead of less. One who draws a clear line between the tasks AI handles well and the judgment calls it cannot touch gets back 15 to 20 hours a week and puts that time toward the work that actually requires a human.
Diana Alvis is Vice President of Human Resources at Crown Capital Management, where she built the full HR infrastructure for a 600-plus employee, multi-entity construction organization during rapid statewide growth. She holds an SPHR certification, has over 20 years of experience across construction, healthcare, legal, retail, and public sector HR, and handles everything from recruiting and compensation design to compliance, benefits, and bilingual field communications. She uses AI extensively in her daily workflow but is deliberate about where it stops.
“AI is not going to handle my judgment calls,” Alvis says. “It’s not going to take my job for me. I have been an HR person for 25 years, and it is not going to handle my judgment.”
Where AI earns its time
Alvis estimates AI saves her 15 to 20 hours per week across a set of bounded, specific use cases. The gains are practical rather than transformational. She is currently migrating between payroll systems and uses AI to identify data gaps, mismatches, and missing fields across spreadsheets. “It’s able to analyze data at light speed,” Alvis says. “It would have taken me so much longer on my own.” She limits what goes in, using employee ID numbers rather than full employee records to keep sensitive data out of the system.
The most surprising win was deduction code cleanup. AI found the most commonly used codes across her system, standardized them, and resolved inconsistencies she had not been able to untangle manually. “I was amazed how quickly it was just cleaning up deduction codes in the background,” she says. “I said, oh, wow. That was fast.”
Crown Capital employs a large Spanish-speaking field workforce, and Alvis uses AI to translate and condense messages for MMS delivery simultaneously. “I can ask it to minimize something for an MMS message and translate it at the same time. I was never able to do that before.” She has also built a custom AI agent trained on her communication style to help polish day-to-day email responses. “It’s almost like I have an assistant helping me respond to things,” Alvis says.
Where she draws the line
Alvis is equally specific about what stays with her. AI does not make judgment calls. It does not evaluate situations with the contextual understanding that managers expect from HR. And it does not get access to data it does not need.
She configures her settings to require source citations so she can verify where information comes from. “We all know not everything on Wikipedia is factual,” she says. “You have to be able to fact-check your data.” Certain compliance-sensitive information never enters the system at all. Even on a private AI instance, Alvis limits inputs to single data points rather than complete employee records. “There are certain things compliance-wise that just don’t go in,” she says.
She also warns that newer HR professionals who lean on AI for judgment calls risk losing the skills that make them effective. “You can become too reliant on anything,” Alvis says. “You have to remember to exercise your brain. You have to remember to think for yourself.” BambooHR’s State of the Workforce 2026 report, The Rising Cost of Dignity Debt, found a similar gap: leaders report productivity gains from AI while employees report rising stress, a divide the report ties to unclear boundaries around how AI gets used.
The result is that Alvis gets the productivity of what would otherwise require additional headcount while keeping the human decision-making that HR exists to provide. AI polishes the communication, audits the data, and translates the message. The judgment about what to do with that information, how to advise the manager, and how to protect both the employee and the organization stays with the person who has 25 years of context that no model can replicate.