ATSRadar Blog — #job search (2 posts)

Most job seekers know that "apply early" is good advice. The harder question is: how early actually matters?

Some openings stay active for weeks. Others look fresh one day and are gone the next time you check.

This report uses ATSRadar's own historical jobs and polling data to answer four practical questions:

1. How long do openings usually remain active after ATSRadar first sees them?
2. How much of the active market is truly fresh right now?
3. Do some ATS ecosystems or role families move faster than others?
4. What should job seekers change if they want to apply earlier?

Key takeaways

• Observed median time-to-inactive: 8.95 days for jobs that have already closed after ATSRadar first saw them.
• Active jobs are not mostly brand new: the median currently open role has already been in ATSRadar for 24.54 days.
• Early closure is real: 14.02% of mature cohorts go inactive within 7 days, and 31.53% are inactive within 14 days, rising to 49.03% within 30 days.
• The truly fresh slice is small: only 1.59% of currently active jobs were first seen in the last 24 hours, and 21.30% were first seen in the last 7 days.
• What this means for job seekers: the best openings do not vanish instantly across the board, but the early-application edge is strongest in the first 7-14 days.

What "fast-moving" means in this report

This article measures observed opening speed, not the exact amount of time a job has existed on the employer side.

That distinction matters.

• First seen means when ATSRadar first ingested the opening.
• Inactive means when ATSRadar detected that the opening had left the active set.
• postedAt is shown for coverage context only because many ATS records do not provide it consistently enough to anchor lifecycle math.

So when this article says a role "closed within 14 days," it means ATSRadar first saw it, then later detected it inactive within 14 days.

That is the right frame for job seekers deciding how often to check alerts, how quickly to apply, and how much freshness should influence prioritization.

Data breakdown

What the data is measuring

This article measures observed opening speed, not employer-side posted duration. In practice that means we anchor lifecycle math to the moment ATSRadar first sees a job (created_at) and the moment ATSRadar detects it has left the active pool (became_inactive_at).

Current analysis window: 90 days ending 2026-03-31 UTC. Jobs analyzed in-window: 441,014.

Chart 1: Observed time-to-inactive buckets

This chart uses only jobs that have already gone inactive, so it describes observed closed-role lifetimes after ATSRadar first saw the posting.

Static chart rendered server-side for reliable loading.

Show full table Hide table

Bucket	Inactive jobs	Share of inactive jobs
<3 days	29,189	16.83%
3-7 days	26,554	15.31%
7-14 days	64,411	37.13%
14-30 days	46,162	26.61%
30+ days	7,156	4.13%

Bucket	Inactive jobs	Share of inactive jobs
<3 days	29,189	16.83%
3-7 days	26,554	15.31%
7-14 days	64,411	37.13%
14-30 days	46,162	26.61%
30+ days	7,156	4.13%

Chart 2: How fresh the current active pool actually is

Static chart rendered server-side for reliable loading.

Show full table Hide table

First seen bucket	Active jobs	Share of current active jobs
24h	4,263	1.59%
3d	15,411	5.76%
7d	56,978	21.30%
14d	88,103	32.93%

First seen bucket	Active jobs	Share of current active jobs
24h	4,263	1.59%
3d	15,411	5.76%
7d	56,978	21.30%
14d	88,103	32.93%

Chart 3: Daily flow of openings into and out of the active pool

Static chart rendered server-side for reliable loading.

Show full table Hide table

Day	New jobs	Jobs became inactive	Net active pool change
2025-12-31	0	0	0
2026-01-01	0	0	0
2026-01-02	0	0	0
2026-01-03	0	0	0
2026-01-04	0	0	0
2026-01-05	0	0	0
2026-01-06	0	0	0
2026-01-07	0	0	0
2026-01-08	0	0	0
2026-01-09	0	0	0
2026-01-10	0	0	0
2026-01-11	0	0	0
2026-01-12	0	0	0
2026-01-13	0	0	0
2026-01-14	0	0	0
2026-01-15	0	0	0
2026-01-16	0	0	0
2026-01-17	0	0	0
2026-01-18	0	0	0
2026-01-19	0	0	0
2026-01-20	0	0	0
2026-01-21	0	0	0
2026-01-22	0	0	0
2026-01-23	0	0	0
2026-01-24	0	0	0
2026-01-25	0	0	0
2026-01-26	0	0	0
2026-01-27	0	0	0
2026-01-28	0	0	0
2026-01-29	0	0	0
2026-01-30	0	0	0
2026-01-31	0	0	0
2026-02-01	0	0	0
2026-02-02	0	0	0
2026-02-03	0	0	0
2026-02-04	0	0	0
2026-02-05	0	0	0
2026-02-06	0	0	0
2026-02-07	0	0	0
2026-02-08	0	0	0
2026-02-09	0	0	0
2026-02-10	0	0	0
2026-02-11	0	0	0
2026-02-12	0	0	0
2026-02-13	0	0	0
2026-02-14	0	0	0
2026-02-15	0	0	0
2026-02-16	0	0	0
2026-02-17	0	0	0
2026-02-18	56,308	366	+55,942
2026-02-19	64,168	777	+63,391
2026-02-20	5,815	3,895	+1,920
2026-02-21	1,091	969	+122
2026-02-22	77	153	-76
2026-02-23	6,522	1,810	+4,712
2026-02-24	5,063	3,737	+1,326
2026-02-25	2,454	2,760	-306
2026-02-26	3,317	3,140	+177
2026-02-27	33,356	29,172	+4,184
2026-02-28	6,167	5,359	+808
2026-03-01	17,612	237	+17,375
2026-03-02	3,487	1,723	+1,764
2026-03-03	3,093	4,449	-1,356
2026-03-04	5,065	4,842	+223
2026-03-05	3,700	3,874	-174
2026-03-06	89,154	3,404	+85,750
2026-03-07	2,880	2,721	+159
2026-03-08	4,891	444	+4,447
2026-03-09	3,525	5,154	-1,629
2026-03-10	3,418	7,794	-4,376
2026-03-11	1,047	973	+74
2026-03-12	2,429	2,150	+279
2026-03-13	3,203	3,081	+122
2026-03-14	703	701	+2
2026-03-15	10,730	10,262	+468
2026-03-16	670	850	-180
2026-03-17	322	265	+57
2026-03-18	3,386	3,394	-8
2026-03-19	17,565	18,878	-1,313
2026-03-20	8,400	6,166	+2,234
2026-03-21	2,890	3,695	-805
2026-03-22	0	2	-2
2026-03-23	1,252	6,952	-5,700
2026-03-24	8,879	5,577	+3,302
2026-03-25	4,126	7	+4,119
2026-03-26	34,904	10,188	+24,716
2026-03-27	748	1,414	-666
2026-03-28	14,126	6,584	+7,542
2026-03-29	210	280	-70
2026-03-30	2,844	5,275	-2,431
2026-03-31	1,420	0	+1,420

Day	New jobs	Jobs became inactive	Net active pool change
2025-12-31	0	0	0
2026-01-01	0	0	0
2026-01-02	0	0	0
2026-01-03	0	0	0
2026-01-04	0	0	0
2026-01-05	0	0	0
2026-01-06	0	0	0

Showing first 7 of 91 rows.

Chart 4: Which ATS ecosystems move faster

Only providers above the minimum volume threshold (5,000 jobs and mature 14-day cohorts) are included.

Static chart rendered server-side for reliable loading.

Show full table Hide table

ATS	Jobs in window	Median observed time-to-inactive	Inactive within 7d	Inactive within 14d	Inactive within 30d
Greenhouse	214,222	13.15 days	6.52%	20.41%	41.18%
Ashby	88,512	8.94 days	15.51%	52.74%	64.10%
Lever	77,072	6.51 days	25.87%	33.33%	41.76%
SmartRecruiters	35,125	5.39 days	27.32%	33.29%	37.50%

ATS	Jobs in window	Median observed time-to-inactive	Inactive within 7d	Inactive within 14d	Inactive within 30d
Greenhouse	214,222	13.15 days	6.52%	20.41%	41.18%
Ashby	88,512	8.94 days	15.51%	52.74%	64.10%
Lever	77,072	6.51 days	25.87%	33.33%	41.76%
SmartRecruiters	35,125	5.39 days	27.32%	33.29%	37.50%

Chart 5: Which role families close faster

Family inference coverage: 66.54% of jobs landed in a non-Other family.

Static chart rendered server-side for reliable loading.

Show full table Hide table

Role family	Jobs in window	Median observed time-to-inactive	Inactive within 7d	Inactive within 14d	Inactive within 30d
Engineering	85,864	8.94 days	13.61%	33.60%	49.22%
Design	71,990	8.98 days	20.51%	35.43%	48.03%
Security	33,783	12.50 days	10.38%	29.08%	47.91%
Sales	29,214	8.95 days	7.90%	30.72%	50.59%
Operations/RevOps	16,438	8.94 days	11.55%	37.05%	55.75%
Marketing	12,565	8.95 days	11.29%	37.14%	55.78%
Data	9,757	8.95 days	14.53%	33.35%	52.28%
HR/People	9,293	13.10 days	12.61%	40.90%	62.62%
Finance	8,571	9.38 days	9.68%	32.38%	56.34%
Product	6,561	8.98 days	8.92%	31.38%	50.20%

Role family	Jobs in window	Median observed time-to-inactive	Inactive within 7d	Inactive within 14d	Inactive within 30d
Engineering	85,864	8.94 days	13.61%	33.60%	49.22%
Design	71,990	8.98 days	20.51%	35.43%	48.03%
Security	33,783	12.50 days	10.38%	29.08%	47.91%
Sales	29,214	8.95 days	7.90%	30.72%	50.59%
Operations/RevOps	16,438	8.94 days	11.55%	37.05%	55.75%
Marketing	12,565	8.95 days	11.29%	37.14%	55.78%

Showing first 6 of 10 rows.

Geography note

ATSRadar stores a raw country value for only 36.16% of jobs in this historical first-seen window, so this article does not publish country speed rankings. That is deliberate: weak normalization would create false precision.

What job seekers should do with this

• If a role matters, treat the first 7-14 days as the highest-value application window instead of assuming it will sit open for a month.
• Keep one broad alert for volume and one tighter alert for the exact titles you want, so you can react quickly without drowning in noise.
• When a provider or role family in this report moves faster, check those openings daily rather than weekly.
• Use freshness as a filter, not a guarantee: a job first seen yesterday is not automatically better, but it is less likely to be late-stage in the hiring process.

Why applying early still matters

The data does not say that every job disappears instantly.

It does say that the first week and the first two weeks matter more than many job seekers assume.

If you only check once a week, you will still catch part of the market. But you will be consistently late on the slice that moves fastest, and that is often the exact slice with the least competition slack.

In practical terms:

1. Use alerts or searches that surface newly seen roles daily.
2. Prioritize the first 7-14 days, not just the first 24 hours.
3. Use freshness to rank your queue, especially in faster ATS ecosystems and faster-moving role families.
4. Do not rely on postedAt alone when it is missing or stale. First-seen timing is often the more reliable operational signal.

Methodology

Analysis window: 90 days ending 2026-03-31T04:08:06.489Z (UTC).

Dataset coverage: first observed job 2026-02-18T06:53:08.863Z, latest observed lifecycle timestamp 2026-03-31T04:08:06.489Z.

Field choices: first seen = jobs.created_at is used as first seen by ATSRadar because it marks first ingestion into the platform. last seen = jobs.last_seen_at is used as the most recent confirmed observation of an opening. became inactive = jobs.became_inactive_at is used as the detected closure timestamp when a previously active opening drops out of the active set.

postedAt coverage: 219,224 jobs in-window (49.71%).

Role-family coverage: 66.54% classified outside Other.

Exact vs Approximate

• Exact: first seen = jobs.created_at (job first ingested by ATSRadar)
• Exact: last seen = jobs.last_seen_at (most recent successful observation)
• Exact: became inactive = jobs.became_inactive_at (when ATSRadar detected the job left the active set)
• Exact: freshness buckets use current active jobs only and first-seen timestamps
• Exact: daily flow uses UTC day buckets from created_at and became_inactive_at
• Approximate: time-to-inactive is exact to ATSRadar detection time, not the exact moment an employer removed the posting
• Approximate: median observed lifetime only covers jobs that have already become inactive inside the analysis window
• Approximate: postedAt reflects source-provided employer timestamps and is used for coverage notes only because completeness is mixed
• Approximate: older inactive rows that predate became_inactive_at were backfilled from updated_at / last_seen_at / scanned_at fallbacks during migration 0033
• Approximate: geography is not published in this article because stored country coverage is not strong enough for a reliable all-cohort speed comparison

Try ATSRadar for free

Want to spot openings earlier instead of refreshing company pages manually?

Register Now, it's free and track fresh openings by title, company, and source.

AI proficiency is no longer a niche signal in hiring, especially for knowledge workers.

Using ATS Radar job data (all geographies), we analyzed how often postings mention AI-tool proficiency and whether employers frame it as required, preferred, or general context.

This report focuses on knowledge-worker roles first, then compares that with the broader workforce.

Charts and Data Breakdown

Snapshot refreshed: Mar 11, 2026 02:54 UTC.

Chart 1: Share of job postings mentioning AI tools (30/60/90 days)

Static chart rendered server-side for reliable loading.

Show full table Hide table

Window	Total jobs	AI mention jobs	AI mention %	Required+Preferred %
30d	110705	27963	25.26%	5.95%
60d	136590	32906	24.09%	5.81%
90d	144462	34169	23.65%	5.71%

Window	Total jobs	AI mention jobs	AI mention %	Required+Preferred %
30d	110705	27963	25.26%	5.95%
60d	136590	32906	24.09%	5.81%
90d	144462	34169	23.65%	5.71%

Chart 2: Top role families by AI-mention rate

Static chart rendered server-side for reliable loading.

Show full table Hide table

Role family	Total jobs	AI mention %	Required+Preferred %
Data	4902	45.90%	18.32%
Product	2417	43.03%	15.64%
Engineering	27478	39.29%	9.56%
Customer Success	1324	38.75%	10.73%
Security	6280	36.27%	6.02%
Design	4595	31.19%	10.82%
Finance	5259	22.78%	7.00%
Marketing	8453	21.21%	4.02%
Operations/RevOps	11753	19.65%	3.03%
Sales	15363	17.26%	2.43%

Role family	Total jobs	AI mention %	Required+Preferred %
Data	4902	45.90%	18.32%
Product	2417	43.03%	15.64%
Engineering	27478	39.29%	9.56%
Customer Success	1324	38.75%	10.73%
Security	6280	36.27%	6.02%

Showing first 5 of 10 rows.

Chart 3: Top industries by AI-mention rate

Static chart rendered server-side for reliable loading.

Show full table Hide table

Industry	Total jobs	AI mention %	Required+Preferred %
Software/SaaS	44821	34.21%	8.04%
AI/ML	21947	29.58%	10.45%
Marketplace/Logistics	17686	24.25%	3.56%
Fintech	8817	21.15%	4.58%
Cybersecurity	8107	20.11%	4.05%
Education	3165	16.05%	0.95%
Energy/Climate	1711	15.55%	2.28%
Other/Unknown	20051	11.64%	2.34%
Healthcare	12169	9.59%	3.20%
Government/Nonprofit	1891	5.55%	1.22%

Industry	Total jobs	AI mention %	Required+Preferred %
Software/SaaS	44821	34.21%	8.04%
AI/ML	21947	29.58%	10.45%
Marketplace/Logistics	17686	24.25%	3.56%
Fintech	8817	21.15%	4.58%
Cybersecurity	8107	20.11%	4.05%

Showing first 5 of 10 rows.

Chart 4: Required vs preferred vs general AI mention split

Static chart rendered server-side for reliable loading.

Show full table Hide table

Class	Jobs	% of AI mentions	% of all jobs
Required / expected	7160	20.95%	4.96%
Preferred / bonus	1096	3.21%	0.76%
General mention	24320	71.18%	16.83%
Company / product context	1593	4.66%	1.10%

Class	Jobs	% of AI mentions	% of all jobs
Required / expected	7160	20.95%	4.96%
Preferred / bonus	1096	3.21%	0.76%
General mention	24320	71.18%	16.83%
Company / product context	1593	4.66%	1.10%

Chart 5: Entry-level vs mid-level vs leadership AI mention rates

Static chart rendered server-side for reliable loading.

Show full table Hide table

Seniority	Total jobs	AI mention %	Required+Preferred %
Mid-level	91985	23.47%	5.85%
Leadership	44105	24.82%	5.24%
Entry-level	8372	19.58%	6.75%

Seniority	Total jobs	AI mention %	Required+Preferred %
Mid-level	91985	23.47%	5.85%
Leadership	44105	24.82%	5.24%
Entry-level	8372	19.58%	6.75%

Chart 6: Most commonly named AI tools

Static chart rendered server-side for reliable loading.

Show full table Hide table

Tool	Mentioning jobs	% of AI-mention jobs
Claude	1286	3.76%
ChatGPT	956	2.80%
GitHub Copilot	798	2.34%
OpenAI	782	2.29%
Cursor	712	2.08%
Anthropic	708	2.07%
Gemini	707	2.07%
Midjourney	102	0.30%
Perplexity	57	0.17%
Llama	54	0.16%

Tool	Mentioning jobs	% of AI-mention jobs
Claude	1286	3.76%
ChatGPT	956	2.80%
GitHub Copilot	798	2.34%
OpenAI	782	2.29%
Cursor	712	2.08%

Showing first 5 of 10 rows.

Chart 7: Knowledge workers vs broader workforce

Static chart rendered server-side for reliable loading.

Show full table Hide table

Segment	Total jobs	AI mention %	Required+Preferred %
All jobs	144462	23.65%	5.71%
Knowledge workers	92154	29.25%	7.06%
Non-knowledge workforce	52308	13.80%	3.34%

Segment	Total jobs	AI mention %	Required+Preferred %
All jobs	144462	23.65%	5.71%
Knowledge workers	92154	29.25%	7.06%
Non-knowledge workforce	52308	13.80%	3.34%

Key findings (last 90 days)

• 23.65% of analyzable postings mention AI tools or AI proficiency in some form.
• In knowledge-worker roles, that rises to 29.25%.
• Explicitly required or preferred AI proficiency appears in 5.71% of all postings and 7.06% of knowledge-worker postings.
• AI mention rates have increased from 23.65% (90d) to 25.26% (30d).
• Remote roles show higher AI-mention rates (35.94%) than hybrid (26.94%) or onsite (18.02%).

30 / 60 / 90-day trend

Window	Total jobs	Jobs mentioning AI	AI mention rate	Required rate	Preferred rate
90 days	144,462	34,169	23.65%	4.96%	0.76%
60 days	136,590	32,906	24.09%	5.07%	0.74%
30 days	110,705	27,963	25.26%	5.26%	0.69%

Takeaway: AI mentions are rising in recent postings, but the biggest growth is still in general expectation language, not explicit preferred/bonus language.

Knowledge workers vs broader workforce

Segment	Total jobs	AI mention rate	Required + preferred
Knowledge-worker roles	92,154	29.25%	7.06%
Non-knowledge roles	52,308	13.80%	3.34%
All analyzable roles	144,462	23.65%	5.71%

This gap is material: knowledge-worker postings are about 2.1x as likely to mention AI.

Which knowledge-work role families mention AI the most?

Role family	Total jobs	AI mention rate	Required + preferred
Data	4,902	45.90%	18.32%
Product	2,417	43.03%	15.64%
Engineering	27,478	39.29%	9.56%
Customer Success	1,324	38.75%	10.73%
Security	6,280	36.27%	6.02%
Design	4,595	31.19%	10.82%
Finance	5,259	22.78%	7.00%
Marketing	8,453	21.21%	4.02%
Operations / RevOps	11,753	19.65%	3.03%
Sales	15,363	17.26%	2.43%

AI expectations are strongest in technical and analytical families, but non-technical corporate functions are clearly involved.

Which industries mention AI proficiency most?

Industry	Total jobs	AI mention rate	Required + preferred
Software/SaaS	44,821	34.21%	8.04%
AI/ML	21,947	29.58%	10.45%
Marketplace/Logistics	17,686	24.25%	3.56%
Fintech	8,817	21.15%	4.58%
Cybersecurity	8,107	20.11%	4.05%
Education	3,165	16.05%	0.95%
Energy/Climate	1,711	15.55%	2.28%
Healthcare	12,169	9.59%	3.20%

Interpretation: AI language is broadest in software-heavy sectors, but it is not limited to AI-native companies.

Required vs preferred vs general mention

Across all AI-mention postings (90d):

Class	Jobs	Share of AI mentions	Share of all jobs
Required / expected	7,160	20.95%	4.96%
Preferred / bonus	1,096	3.21%	0.76%
General mention	24,320	71.18%	16.83%
Company/product context only	1,593	4.66%	1.10%

The big signal right now: AI is often treated as baseline context in job language, while explicit “must-have” phrasing is growing but not yet dominant.

Entry-level vs leadership: who gets AI expectations?

Seniority group	Total jobs	AI mention rate	Required + preferred
Entry-level	8,372	19.58%	6.75%
Mid-level IC	91,985	23.47%	5.85%
Leadership (Manager+)	44,105	24.82%	5.24%

Leadership postings mention AI more often overall, but explicit required/preferred phrasing is not dramatically higher than IC roles.

Which AI tools are named directly?

Named-tool mentions are still a minority vs generic AI language.

Tool	Mentioning jobs	% of AI-mention jobs
Claude	1,286	3.76%
ChatGPT	956	2.80%
GitHub Copilot	798	2.34%
OpenAI	782	2.29%
Cursor	712	2.08%
Anthropic	708	2.07%
Gemini	707	2.07%

For job seekers: optimize for AI workflows and outcomes, not just one branded tool.

Practical read for job seekers

If you are applying into knowledge-work roles in 2026, assume AI literacy is now an expected part of your toolkit, even when postings are not explicit.

A practical approach:

1. Add concrete AI workflow examples to resume bullets (analysis, drafting, automation, QA).
2. Show business outcomes, not just tool names.
3. Prepare interview stories for responsible AI use and quality control.
4. For data/product/engineering roles, expect materially higher AI expectations.
5. For non-technical roles, focus on productivity, documentation, and process automation use cases.

Methodology (short)

• Analysis run date: March 10, 2026 (America/Los_Angeles).
• Windows: rolling last 30 / 60 / 90 days.
• Geography: all available geographies in ATS Radar source data.
• Population: active jobs with coalesce(posted_at, scanned_at) in window.
• Analyzable text filter: description_text length >= 150 and non-harvester parser rows.
• Analyzed 90-day base: 144,462 postings.
• AI detection: rule-based mention detection across AI terms and named tools, then classification into:
  • required/expected,
  • preferred/bonus,
  • general mention,
  • company/product context.
• Job family and seniority: ATS Radar’s existing rule-based role classification heuristics.
• Knowledge-worker definition: all ATS Radar job families except Other.

Limitations and confidence

• This is a text-signal analysis, not a verification of how strictly hiring teams enforce each requirement.
• Some ATS sources provide shallow descriptions; those were excluded from proficiency classification.
• Multi-language posting text and non-standard formatting can cause under-detection or misclassification.
• Classification is designed to be conservative for explicit required/preferred labels.

Confidence level: moderate-high for directional trend and segment comparisons, moderate for exact required/preferred split.

Chart-ready data files are available in content/blog/data/ai-tool-proficiency.

Track your own AI-skill market shifts

Use ATS Radar to watch how AI expectations move in your target role family, location, and seniority band.

Register Now, it's free

Related ATS Radar reads:

• Remote Jobs Reality Check (Last 30 Days)
• Entry-Level vs Senior: Where the Market Actually Is
• Daily Hiring Trends by Geography & Industry