Hiring used to begin with a handshake, or at least a phone call. Today, many job seekers apply and get rejected without ever speaking to anyone at the company, and they may not even get a response.
Both sides of the process now lean on artificial intelligence. Candidates use it to turn out resumes and cover letters in bulk, while employers use it to sort the pile that results. Hiring managers end up with a stack of applications that read alike, and qualified people slip through without a real look. For cleaning and restoration companies that depend on finding dependable technicians and frontline workers, that gap gets expensive fast.
Matt Rosen sees the problem from inside the technology world. He is the founder and CEO of Allata, a Dallas-based consulting and technology services firm he started in 2016. The company has offices in the U.S., India and Argentina and helps clients with custom software, data work and AI adoption. Rosen said questions about AI never let up.
“Every day, that is at top of our clients’ minds,” Rosen said. “How is AI changing? When should they jump in? What should they not use it for? How do they make sure it doesn’t expose things it shouldn’t in their environment? And then how do they work it into their processes?”
“Everyone’s using it to some degree or another, but varying levels of maturity for sure,” he added.
A flood of applications that sound alike
Allata is always hiring, and much of that hiring is for engineers, analysts and other highly technical roles. Those candidates use AI tools constantly, and Rosen said they are very good at using them to build strong resumes and keep in touch with the company. The volume has created its own headache.
“We use AI as a firm to help do initial screens of resumes because we get thousands, and a lot of them are AI-generated resumes or it’s a lot of noise,” Rosen explained. “It’s like, is this real? Is this not real? How do I know there’s a person? How do I know they’re not working for two other companies at the same time?”
That last worry carries more weight when employees work remotely, as many at Allata do. The firm’s answer has been to keep hiring people-intensive.
“While we may use AI to help us screen resumes and create lists of questions for a unique role, our interviewing process is very much driven by people,” Rosen noted. “We would never have an AI conduct an interview with a person.”
Where the robot interview goes wrong
AI-led interviews are spreading anyway. Recordings of applicants struggling through sessions with software that mishears or misreads their answers have circulated online, and Rosen has little patience for the practice.
“My understanding is this is happening across the board, and I think it’s a pretty bad practice, especially if the first interaction that a human has with your company is that they’re literally talking to an AI, whether it’s a chatbot, whether it’s an avatar, whatever it is,” he said.
He does see limited room for automation at the front door. “I could see for an initial screen, some basic HR questions and telling people up front, ‘Hey, you’re going to do this because it’s going to help us save time later,’” Rosen acknowledged. In-depth conversations meant to get to know someone, or quizzes on technical skills, are another matter.
“I believe that if you’re going to hire good people, you’ve got to have good people hiring them and using AI to assist them,” he stated.
Rosen is comfortable letting AI handle the routine work around recruiting, such as screening resumes, drafting thank-you emails, preparing offer letters and sending reminders that move candidates through each step. The judgment call is where he draws the line.
“I’m totally okay AI enabling the recruiting and HR function in an organization,” Rosen said. “What I’m not okay with is AI taking the place of a human being, getting to know another human being and determining their fit within their organization.”
For a contractor, fit often comes down to things no screening tool can measure well, like how a technician talks to a homeowner standing in a flooded basement.
Look for numbers, not adjectives
Rosen doesn’t throw out a resume simply because AI helped write it. “I want them to make it their own,” he said.
After reading enough machine-written material, he noted, the tells become easy to spot. “AI is super verbose,” Rosen explained. “If I see this really flowery language and a lot of stuff, it’s not compelling.”
What he wants is specifics. “At the end of the day, a lot of times I’m looking for numbers,” he said. “If it’s a sales role, I want to know exact numbers. What did you sell? What was your quota? How did you perform?”
For engineers, Rosen looks for the type of projects a candidate worked on and the benefit those projects brought to the client. He admitted some roles can’t always be measured that way, though he still wants applicants to show the problems they have mastered and focus on accomplishments rather than filler.
Formatting gives people away too. “Claude has some very interesting tells in how it shades and color codes things,” Rosen noted. “If people are too lazy to take it and put it into their own professional format, that’s a pretty dead giveaway.”
Online job boards make the call harder. On sites like LinkedIn, where applicants mostly fill in blanks, Rosen said it can be tough to tell whether someone made up their profile with AI and whether they can actually do the job.
The same test works for a cleaning or restoration owner sorting through applicants. A candidate who can tell you how many jobs they ran in a week, or how they handled a customer complaint, has told you more than a page of polished phrases.
Getting past the gatekeeper
Qualified candidates who get filtered out before a person ever sees their application have an option, according to Rosen. Go around the system.
“Don’t be scared to pick up the phone or find the CEO or an executive’s email,” he said. “And if that annoys the recruiting team, then so be it.”
Rosen reads every LinkedIn message he receives, and when a candidate asks him for a look, he passes the name along to his recruiting team. “I don’t respond to all of them, but I do take the time to read them,” he shared.
He suspects many leaders would do the same, particularly those who are short on staff and don’t realize good people are being blocked by a process “put in a few levels below them.”
“If the company really needs folks and you feel like you’re really qualified, they shouldn’t be offended if you make a few extra calls or send a few extra emails,” Rosen said.
Plain persistence has worked for him personally. “That’s how I got many of my jobs, is just reaching out,” he added. “Or it’s through relationship or referral.”
Referred candidates at Allata still go through the full HR process, which provides checks and balances a business its size needs. “But it doesn’t mean that people don’t get a step ahead because they’ve reached out or have a relationship,” Rosen said.
That advice has a flip side for owners. If a sharp technician calls the shop asking for a shot, somebody should answer the phone.
Watch the complete interview with Matt Rosen or listen to the podcast below:
