What AI recommends for applicant tracking systems
Across 35 live answers recorded on August 20, 2026, the applicant tracking systems most often recommended by ChatGPT, Claude, Gemini, Perplexity and Grok were Greenhouse, Lever, Workable. The five engines named 49 distinct brands between them, and all five agreed on 9 of the top 20. The rest of the field depends on which engine a buyer happens to ask.
The leaderboard
Ranked by how many of the 35 answers named each brand. Average position is where the brand appeared in the list the engine gave — 1.0 means it was named first every time it came up.
| # | Brand | Answers naming it | Engines | Avg position | Sentiment |
|---|---|---|---|---|---|
| 1 | Greenhouse | 21 of 35 | 5 of 5 | 1.9 | positive |
| 2 | Lever | 20 of 35 | 5 of 5 | 3.3 | positive |
| 3 | Workable | 16 of 35 | 5 of 5 | 3.6 | positive |
| 4 | Ashby | 15 of 35 | 5 of 5 | 5.0 | positive |
| 5 | iCIMS | 13 of 35 | 5 of 5 | 4.9 | positive |
| 6 | SmartRecruiters | 12 of 35 | 5 of 5 | 7.6 | positive |
| 7 | JazzHR | 10 of 35 | 5 of 5 | 6.5 | positive |
| 8 | Workday Recruiting | 9 of 35 | 5 of 5 | 5.9 | positive |
| 9 | BambooHR | 9 of 35 | 4 of 5 | 7.4 | positive |
| 10 | Breezy HR | 8 of 35 | 5 of 5 | 3.8 | positive |
| 11 | Zoho Recruit | 7 of 35 | 4 of 5 | 4.4 | positive |
| 12 | Workday | 7 of 35 | 4 of 5 | 5.4 | positive |
| 13 | SAP SuccessFactors Recruiting | 5 of 35 | 4 of 5 | 7.2 | positive |
| 14 | SuccessFactors | 5 of 35 | 4 of 5 | 9.0 | positive |
| 15 | Oracle Recruiting Cloud | 4 of 35 | 3 of 5 | 5.3 | positive |
| 16 | Workday HCM | 4 of 35 | 4 of 5 | 6.5 | positive |
| 17 | Taleo | 4 of 35 | 4 of 5 | 7.0 | mixed |
| 18 | Rippling | 4 of 35 | 3 of 5 | 9.5 | positive |
| 19 | SAP SuccessFactors | 3 of 35 | 2 of 5 | 5.7 | positive |
| 20 | Jobvite | 3 of 35 | 2 of 5 | 9.0 | mixed |
Where the engines agree, and where they do not
All five engines named Greenhouse, Lever, Workable, Ashby, iCIMS, SmartRecruiters, JazzHR, Workday Recruiting and 1 more. Those are the defaults in this category — the brands an AI answer reaches for regardless of which model a buyer happens to be using.
Across all measured categories the pattern is consistent: the leader is agreed on, the rest is not. The pooled study puts numbers on it.
Every brand on the leaderboard was named by more than one engine.
| Engine | Named most often |
|---|---|
| ChatGPT | Greenhouse, Lever, Workable, Ashby, iCIMS |
| Claude | Greenhouse, Lever, Workable, Ashby, iCIMS |
| Gemini | Greenhouse, Lever, Workable, Ashby, iCIMS |
| Perplexity | Greenhouse, Lever, Workable, Ashby, iCIMS |
| Grok | Greenhouse, Lever, Workable, Ashby, iCIMS |
What this means if you sell applicant tracking systems
Being absent from this table does not mean the engines dislike you. It usually means they have nothing to reach for: no comparison pages, no presence on the review and listicle sites they cite, no clear statement of who the product is for. The generative engine optimization guide covers what actually shifts this, and the audit process covers how to run the measurement yourself.
Note also what this page cannot tell you. It measures the category, not you. It does not include the brand-level prompts that reveal what an engine says when someone asks about your product by name, it carries no share-of-voice figure for your brand, and it does not list the domains AI cites when answering questions in this category. A $19 snapshot adds all three for one brand.
The prompts behind this page
Every category page uses the same seven prompts, each sent to all five engines:
What are the best applicant tracking systems right now? List the top options.Which applicant tracking systems would you recommend for a small business, and why?Compare the leading applicant tracking systems options — who stands out?What is the best free or affordable applicant tracking systems?What applicant tracking systems do you recommend for a larger company / enterprise?What are the most popular applicant tracking systems people are switching to?I'm choosing a applicant tracking systems. Give me a shortlist of 5 with one-line pros and cons.
Language models are stochastic, so a rerun on another day will not reproduce this table exactly. That is a property of the medium rather than a flaw in the measurement, and it is one reason to distrust any AI visibility score quoted to one decimal place. The methodology page sets out the limitations in full.