More than 90% of organizations have deployed AI in talent acquisition, yet fewer than 5% report transformational outcomes. That is the headline finding of a study titled "The New Talent Equation: Building Better Talent Decisions," commissioned by ManpowerGroup Talent Solutions and produced with Everest Group, published on 23 June 2026. For US job seekers already frustrated by slow pipelines and vanishing follow-ups, the data puts a number on something that felt personal but is actually structural.

The study surveyed 80 C-suite, CHRO, and senior talent acquisition leaders across the United States and the United Kingdom. Its diagnosis is direct: AI is being bolted onto sourcing, resume screening, and candidate engagement, but fragmented systems, isolated tools, and siloed data are preventing most of the promised efficiency from reaching candidates at all.

What the research actually found

ManpowerGroup Talent Solutions and Everest Group frame a wide adoption-to-impact gap. Near-universal usage sits alongside modest, uneven gains. Most respondents report only moderate improvements in decision quality, workforce agility, and strategic capacity. The transformational tier, which is exactly what vendors advertise against, is a sliver of the sample.

Three blockers surface repeatedly:

• Fragmented workflows. An applicant tracking system, a sourcing tool, a scheduling bot, and a CRM that do not communicate with each other cleanly.
• Siloed data. Candidate history, assessment scores, and interview notes sitting in separate systems, so no single model has a complete picture of any one applicant.
• Governance gaps. Rules about bias review and human-in-the-loop oversight that are inconsistent across tools, which slows anything sensitive to a crawl.

The report also flags a second-order problem: candidates are now using AI too, drafting resumes and cover letters with generative tools. That changes the signal recruiters are trying to read, and the current hiring stack was not designed for that shift.

For employees and candidates: what this means for your applications

The frustrating experience many US job seekers describe, slow pipelines, silence after a submission, an interview that then goes cold, is not usually the story of one disorganized recruiter. It is a systems problem. When ManpowerGroup Talent Solutions and Everest Group describe isolated tools and siloed data, they are describing the reason an application status page still reads as pending five weeks after submission. A Society for Human Resource Management (SHRM) analysis of hiring practices has consistently found that process fragmentation, not recruiter intent, drives the majority of candidate experience failures in corporate hiring funnels.

A few practical shifts help you work around a broken pipeline:

• Treat the resume as the first filter, not the interview. AI screening is the most common use case in the ManpowerGroup and Everest Group research. Your resume needs to pass a machine reader before a human ever sees it. Keep the structure clean, mirror the language of the job posting, and cut decorative formatting that parsing tools cannot read.
• Do not rely on a single application channel. Because tools often do not communicate with each other, a referral, a direct message to a hiring manager, or a submission through a separate jobs platform can each be visible to different parts of the same organization. Using multiple channels is a realistic response to a fragmented system, not a sign of desperation.
• Ask about the process on your first call. A recruiter who can tell you exactly how many interview stages there are, who reviews each one, and when a decision is expected is operating inside a functioning pipeline. A vague or shifting answer is often a sign that the pipeline itself is fragmented.
• Follow up on your own timeline. If a role stays open past the date the recruiter cited, a polite check-in is reasonable. Silence frequently means a record is stuck between two systems, not that a rejection decision has been made.

One caveat worth stating plainly: using generative AI to draft an entire application is now visible to the other side, and the research is clear that recruiters have noticed. Use AI as an editor and a structure check, not as a ghostwriter. The parts that sound like you are the parts that survive review.

For HR and talent leaders: why the license has not moved time-to-hire

The uncomfortable read of the ManpowerGroup Talent Solutions and Everest Group findings is that AI in recruitment is now table stakes and, on its own, no longer a differentiator. Ninety percent adoption with fewer than five percent transformational outcomes is not a technology gap. It is an integration and governance gap.

Several observations are load-bearing for US hiring operations specifically. The US operates under a federal-plus-state compliance framework that makes governance gaps in AI recruiting tools a legal exposure, not just an operational inefficiency. Under Title VII of the Civil Rights Act of 1964 and the Americans with Disabilities Act (ADA), automated screening tools that produce disparate impact on protected classes can generate Equal Employment Opportunity Commission (EEOC) liability regardless of intent. The EEOC issued technical assistance in May 2023 specifically addressing employer responsibility when AI-driven hiring tools screen out protected-class candidates, putting the compliance obligation squarely on the employer, not the vendor. Separately, the Department of Labor (DOL) has signaled ongoing scrutiny of automated hiring practices under Executive Order 13985 frameworks. For HR leaders, that regulatory context makes the governance gaps described in the ManpowerGroup and Everest Group research an urgent operational matter, not a future-state concern.

Against that backdrop, four observations stand out for US talent leaders:

• The stack is the strategy. Adding a new AI point solution on top of an ATS that already does not integrate with your CRM tends to add latency, not remove it. The teams reporting real gains in the research are consolidating tools, not accumulating them.
• Governance decides speed. Bias review, human-in-the-loop rules, and audit trails are not the enemy of velocity. Inconsistent governance across tools is. Standardizing sign-off authority once, across the full stack, typically runs faster than reworking each hire under a different internal rulebook. EEOC audit readiness is far easier when that rulebook is unified.
• The candidate side is now an AI system too. If your screening logic assumes a human-written resume and a human-written cover letter, and your candidate pool is increasingly using generative tools, your signal-to-noise ratio is quietly degrading. That is a scoring model problem, not a sourcing problem.
• Measure the pipeline, not the tool. The right question is not whether the AI is working in the abstract. The operational question is: has time-to-hire moved, has quality-of-hire held or improved, has candidate drop-off between stages fallen. If the answer is no on all three, the license is not the problem worth solving next.

The 80-leader sample in the ManpowerGroup and Everest Group research is not representative of every US employer, and the report is commissioned rather than independent. But the direction it describes aligns with what SHRM member surveys and DOL workforce data have long flagged: adoption of new hiring technology is fast, integration into functioning processes is slow, and the gap between them is where recruiting budgets are spent without much to show for it.

The honest bottom line

AI has not fixed recruiting. It has amplified whatever process was already in place. A tidy pipeline gets a bit tidier. A fragmented one gets faster at being fragmented.

For candidates navigating the US job market, the practical implication is that your interview timeline is still, mostly, a human process problem sitting inside a technology wrapper. Tools like applicant tracking systems and AI screeners set the pace, but process design and governance decisions made by people determine whether any of it works. For HR and talent leaders, the implication from the ManpowerGroup Talent Solutions and Everest Group research is direct: the next unit of value is not another tool. It is the integration layer, the unified data model, and the governance framework that makes the tools already installed actually work together, and that keeps the organization defensible when the EEOC comes asking about automated screening outcomes.