Praktika's talent head returned from leave to Claude Cowork and applicant pools over 1,000

Two decisions sit inside Ashley Wallace's account, and they run on different clocks. The returner's gap is a question about one person's first weeks back. The applicant pool is a standing condition of every open role. A manager can settle the first with a plan this quarter, and the second will still be there next quarter.

"I came back thinking, oh, I'll ease myself in," Wallace said. "It definitely isn't the case in the landscape that I came back to." ^[4] An easing-in plan written for the person who left assumes the tools stayed put. At Praktika they did not, and six months was long enough for the system everyone was talking about to arrive without her ^[1][3].

Only Wallace returned to a job at an employer she had left on leave. Max Cherry took a new public relations job after three years away ^[8], so his case is ordinary onboarding after a gap. Jess Clark pivoted her own business and reacted to posts on her phone ^[10]. A skeptic would say a week of worry over social media is not a skills gap, and for Clark that is largely right: "I was in a full spiral for about a week," she said ^[11]. What she did next was individual effort. She taught herself to code with Claude and reoriented her courses toward it ^[12]. Cherry described the same pattern: "I have to train myself like, 'hey, it's OK to put this into the AI agent,'" he said ^[9].

The surveys show how fast adoption is moving. Gallup's July poll found nearly half of workers say their companies have adopted AI, up from 38% last fall ^[5]. McKinsey's spring survey found 40% of workers at large companies say their firms are scaling agents, up from 27% last year, a rise of 13 percentage points ^[6][18]. PwC found 64% of workers have used AI this year ^[7]. These are adoption measures. They do not show how fast a single role changes or what employers do for returning staff.

Elapsed time does not separate the cases: Clark's absence of about three weeks produced a week of worry ^[17], and Wallace's six months produced a changed job ^[1]. We'd tie a plan to events instead: any absence during which a tool was adopted. Business Insider describes new enterprise systems, workflows and AI sprint experiments that rise and fall ^[15], so a plan that lists today's products goes stale for the next returner. A plan that says where the tools are and who answers questions lasts longer, at the price of manager time now.

The evidence on screening is thinner. Praktika's pool, a few hundred before and more than 1,000 for some jobs now, is Wallace's own count ^[19], and the cause is her suspicion that applicants used AI ^[20]. Whether an earlier filter drops good candidates will show only if someone samples the rejected applications next quarter.