From Matching Models to Recruiting Agents: A Systematized Narrative Review of AI Recruitment Systems, Evaluation, and Governance

Computer Science > Artificial Intelligence
[Submitted on 3 Sep 2026]
Title:From Matching Models to Recruiting Agents: A Systematized Narrative Review of AI Recruitment Systems, Evaluation, and Governance
View PDF HTML (experimental)Abstract:Artificial intelligence in recruitment has shifted the object being automated from profile pairs and ranked lists to multi-stage workflows that retrieve evidence, compare candidates, and support or execute actions. This systematized narrative review traces that development from bilateral retrieval and behavioral ranking through neural person--job matching, large language model (LLM) components, and tool-using recruiting agents. Using a purposive search and coding protocol updated through 23 July 2026, plus targeted updates through 2 September 2026, we organize 40 representative works with supporting industrial and legal sources. This synthesis is not a prevalence estimate. We analyze three coupled transitions: from similarity to reciprocal suitability, from a model to a compound workflow, and from offline prediction to evidence- and productivity-aligned evaluation. Across document understanding, retrieval, ranking, assessment, interviewing, sourcing, and human handoff, we distinguish field-, pair-, list-, case-, trajectory-, and outcome-level evidence. Persistent gaps arise because behavioral labels confound exposure, preference, and qualification; private and synthetic data limit external validity; final-output scores conceal pipeline failures; and, within the coded set, privacy is not directly evaluated and no row jointly evaluates utility, fairness, privacy, and security. These observations describe the coded set rather than the field as a whole. We therefore introduce a staged mapping from evaluation evidence to the strongest defensible claim, together with an agenda for reciprocal, evidence-grounded, temporally controlled, selective, and auditable systems. Progress should be judged by whether workflows retrieve the right evidence, preserve uncertainty, support contestable decisions, and improve outcomes under explicit cost and risk constraints.
References & Citations
Loading...
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

* The focus of AI in recruitment has shifted from profile pairs and ranked lists to multi-stage workflows involving evidence retrieval, candidate comparison, and action execution.
* Development traces a path from bilateral retrieval and behavioral ranking to neural person-job matching, LLM components, and recruiting agents using tools.
* A systematic review organized 40 representative works with industrial and legal sources.
* Three transitions are analyzed: similarity to reciprocal suitability, model to compound workflow, and offline prediction to evidence- and productivity-aligned evaluation.
* Evidence levels distinguished include field-, pair-, list-, case-, trajectory-, and outcome-level evidence.
* Gaps exist because behavioral labels confound exposure/qualification; private data limits external validity; final scores conceal pipeline failures.
* The coding set does not jointly evaluate utility, fairness, privacy, and security.

Artificial intelligence in recruitment has evolved from automating the matching of profile pairs and ranked lists to managing multi-stage workflows that retrieve evidence, compare candidates, and execute actions. The development traces a progression from initial bilateral retrieval and behavioral ranking through neural person-job matching, integration of large language model components, and the deployment of tool-using recruiting agents. A systematic review organized 40 representative works with supporting sources, analyzing three coupled transitions: moving from similarity to reciprocal suitability, from a singular model to a compound workflow, and from offline prediction to evaluation aligned with evidence and productivity. The review distinguishes various levels of evidence—field, pair, list, case, trajectory, and outcome level—but notes persistent gaps due to confounding factors like behavioral labels, data privacy constraints, and the lack of joint evaluations for utility, fairness, privacy, and security in coded sets.

The narrative outlines a critical tension between the sophisticated capabilities of AI in recruitment and the need for defensible, auditable systems. The transition from simple similarity matching to complex workflow management reflects an attempt to capture the nuanced reality of candidate suitability through evidence-based assessment. However, the systematic limitations identified—where behavioral labels obscure true qualification, data privacy limits external validation, and output scores mask pipeline failures—suggest that current evaluation methods are inherently incomplete when attempting to assert system efficacy. The call for a staged mapping from evaluation evidence to defensible claims suggests a necessary pivot toward systems where progress is judged by their ability to retrieve correct evidence, preserve uncertainty, support contestable decisions, and operate within explicit cost and risk constraints. This implies the core challenge shifts from merely achieving better predictions to establishing epistemological accountability within automated hiring processes.

From Matching Models to Recruiting Agents: A Systematized Narrative Review of AI Recruitment Systems, Evaluation, and Governance

Facts Only

Executive Summary

Full Take

Sentinel — Human

From Matching Models to Recruiting Agents: A Systematized Narrative Review of AI Recruitment Systems, Evaluation, and Governance

Facts Only

Executive Summary

Full Take

Sentinel — Human