Why AI has not made hiring faster

In early September ManpowerGroup published its quarterly employment outlook survey, and HR Dive reviewed the findings on 15 September. The survey draws on responses from nearly 40,000 public and private employers across 42 countries, including more than 6,000 in the United States. One of the more striking results is not about how many employers intend to hire, but about how fast they manage to do it.

Only a third of employers report that their time-to-hire has improved compared with 2025. Another 42 percent see no change, and 25 percent say the process has actually become slower. This is happening while AI tools have become close to universal in recruitment. HR Dive points to a Paylocity report in which 91 percent of HR leaders say they actively use AI in their recruitment process, with the remainder planning to do so within the next 18 months.

Among the obstacles employers cite are a shortage of candidates with the necessary skills, too few qualified candidates in local markets, mismatches between what candidates expect and what the job requires, and fewer applications coming through referral networks. Lengthy recruitment processes are also named as an obstacle in their own right.

It is worth being precise about what the survey shows and what it does not. It shows that widespread use of AI and improved time-to-hire do not go hand in hand across a large sample of employers. It does not show why. The rest of this article is the editorial team's reading of what typically happens in practice, and it should be read as interpretation rather than documented findings.

A recruitment process is a chain of steps. An AI tool that screens CVs in seconds saves time in one link of that chain. If candidates then wait a week for a test invitation because invitations are sent by hand, and several more days for someone to open the report and compare it with the CV, the overall gain has been consumed. Time-to-hire is set by the slowest link, not the fastest.

This is where the test workflow and its integration with the applicant tracking system become decisive. Not because integration is a mark of quality in itself, but because most delays in a testing stage are coordination delays. Who sends the invitation and when, who is notified when the result is in, and where the result sits when the decision has to be made.

When a testing stage is mapped step by step, the same bottlenecks tend to surface.

The handover from screening to testing. If a recruiter has to move candidates into a test stage manually and press send, the invitation depends on that person's calendar. A rule-based transition in the ATS removes the waiting time, but it requires that the criteria for who gets invited have been decided in advance.

Reminders and expiry. Candidates who never complete the test keep the pipeline open. A fixed deadline and automatic reminders are simple to configure, yet they are often overlooked when the integration is set up.

The result back in the ATS. Whether the system returns a single overall score, an overall score with sub-scores, or only a link to a full report affects how quickly a hiring manager can act. A link that requires a separate login is frequently never opened.

The decision rule. The slowest link is often not technical but human. There is no agreed rule for how the test result weighs against the CV and the interview, so the result becomes one more data point to be debated in a meeting rather than an input that moves the candidate forward or out.

Measure time per step, not only from application to signed contract. Without step-level measurements it is impossible to tell whether a new tool has shortened anything or has merely shifted the waiting time somewhere else.

Decide test criteria and decision rules before the integration is built. An integration can only automate what has already been decided.

And be sceptical of speed claims made for any single tool. The two surveys HR Dive cites have different populations and cannot simply be added together. Placed side by side, though, they paint a picture in which nine out of ten HR leaders use AI in recruitment while only a third of employers experience faster hiring. Whatever the cause, that is a signal that speed comes from the process as a whole, not from one link in it.

Source: HR Dive, "AI hasn't significantly improved the speed of hiring, report finds", 15 September 2026, https://www.hrdive.com/news/ai-hasnt-significantly-improved-the-speed-of-hiring/830316/