Your Next Applicant Is Already Using AI to Apply, Research, & Negotiate
Sign Up For Daily Job Market News
A new survey from The Harris Poll, conducted with Ruth AI, found that 81% of Americans would let an AI agent handle at least one part of their job search. If you hire people, that number describes the candidates already landing in your inbox.
Look at what people are willing to delegate. Two-thirds would let AI search for jobs (67%) and run their pre-interview research (67%). Nearly two-thirds would let it update their resume (65%), and more than half (55%) would let it apply to jobs automatically. 45% say they have already used a tool like ChatGPT, Claude, or Gemini for career or work advice, and that jumps to 66% of Millennials and 63% of Gen Z.
It goes further into pay. One in three U.S. adults has used AI for advice on salary, raises, bonuses, or negotiation, and roughly half would hand the negotiation itself to AI: 47% for salary, 49% for benefits.
The awareness gap is the uncomfortable part. 76% of Americans have never heard that independent research has found AI can produce biased career and salary guidance, and 72% agree AI can sound confident even when its advice is wrong. People are delegating faster than they are checking.
Candidates are arriving more prepared, more polished, and more automated than they were even a year ago. None of it is a problem to solve. It is the new baseline to plan around. A resume that reads perfectly and a candidate who name-drops the right facts about your company tell you less than they used to, because AI can manufacture both. The screening burden shifts back to conversation: what can this person actually do, and how do they think when the script runs out. Sorting real capability from a good prompt is the part of hiring we spend most of our day on, and it is why the human read still decides the hire.
Google’s Data Says AI Is Assisting Workers, Not Taking Their Jobs
While workers are handing more of the job search to AI, Google put out research this week arguing that at work, AI is mostly a helper rather than a replacement.
The study, called ATLAS, is built from about 15 million de-identified interactions with Google’s AI tools, including Gemini. Instead of guessing which jobs could be automated someday, it looks at how people actually use AI right now. The finding: most use is what Google calls “shallow,” with the average worker using AI for about 21% of their tasks and leaning on it as a collaborator for research, drafting, and troubleshooting rather than turning over whole jobs.
The most interesting piece for me is the reach beyond desk work. AI use showed up across roughly 70% of occupations, including skilled trades. Electricians and auto mechanics are uploading photos and videos to diagnose problems, which makes AI a hands-on tool, not just a writing and coding assistant.
The practical read is to treat AI as a productivity multiplier rather than a headcount plan. The organizations pulling ahead are the ones helping their people fold AI into real workflows, not the ones counting on it to justify cutting the team. If you are weighing where AI changes your staffing math, that is a conversation worth having before you set next year’s plan, and it is the kind of thing we work through with clients through our staffing services.
Staffing your team doesn’t have to be hard.
Reach out and see how we can help.
Upwork’s Data Shows Companies Hiring to Build AI, Not Just Test It
Upwork’s June hiring report points the same direction from the demand side. The fastest-growing category was AI Apps & Integration, up 34% month over month and 33% year over year. Companies are past the tinkering stage and are hiring people to embed AI into products, workflows, and business systems.
The rest of the report fills in the picture. Hiring for ERP and CRM software projects grew 13% month over month, and Lead Generation & Telemarketing rose 9%, so the money is going into revenue and operations infrastructure alongside AI. Among small and midsize businesses, the growth was in execution: QA Testing up 20% and Data Entry & Transcription up 19%.
I keep coming back to those last two numbers. There is a loud narrative that AI erases QA and data-entry work, and Upwork’s own data cuts against it. As Jennifer Brett of the Upwork Research Institute put it, “continued demand for QA testing and data entry is a reminder that AI adoption still depends on human execution.” Somebody has to validate the outputs, keep the data clean, and turn a model’s answer into a reliable result.
Jobless Claims Fell to 187,000, and the Labor Market Held Steady
The hard data this week backs up the calmer story. The Labor Department reported 187,000 initial unemployment claims for the week ending July 18, down 22,000 from the prior week’s revised 209,000. The four-week moving average dropped to 207,500, so this was not a one-week blip.
Continuing claims held steady at 1,796,000, with the insured unemployment rate unchanged at 1.2%. A year ago that continuing-claims figure was 1,941,000, so more people are staying employed than at this point last summer.
The state detail is where the caution lives. For the week ending July 11, the largest increases in initial claims came from New York (+12,580), Michigan (+3,143), Florida (+2,799), Texas (+2,676), and South Carolina (+2,113). States tied the jumps to layoffs in transportation and warehousing, healthcare, education, manufacturing, construction, retail, and hospitality. New York cited transportation, healthcare, and education; Michigan pointed to manufacturing; Florida spread it across agriculture, construction, manufacturing, wholesale, and retail.
A low claims number does not mean recruiting gets easy. Employers are making targeted cuts, not broad ones, which keeps skilled candidates in demand and competition tight. If you wait too long to fill a critical role, you can still lose the candidate you wanted. For job seekers, this is a market that rewards preparation, especially keeping technology and AI skills current as companies keep changing how the work gets done. Pay is part of that preparation too, and our salary data is a good place to ground a number before you negotiate.
