INTEGRATION OF AI AGENTS IN RECRUITMENT:

OPPORTUNITIES AND CHALLENGES

Labiba Fairuz Hassan

, Md. Aminul Islam

, Abdullah Al Mamun

Department of Management Information Systems

Independent University, Bangladesh

Abstract

This research illustrates the potential implications of AI agent deployment in the

recruitment ecosystem, primarily through a Bangladeshi multinational cultural lens. The

goal is to streamline procedures, accelerate recruitment, introduce a paperless workow,

and facilitate more rigorous, robust evaluations of prospective candidates, reducing human

error and bias to ensure optimal job alignment. However, a few international studies have

highlighted AI’s utility in the hiring process, but its actual use and impact in Bangladesh

remain largely unstudied. Hence, the results of this study are anticipated to have a broader

impact on Bangladesh’s digitalization of the traditional hiring process, reducing time and

human intervention while addressing algorithmic and behavioral biases. More importantly,

the research also emphasizes the need to educate applicants and recruiters about technical

errors in AI systems, including the trial period of operation and the duration of data feeding.

Ultimately, this study supports not only AI implementation but also responsible adoption

of AI, backed by a regulation framework, stakeholder understanding, and capacity-

building initiatives, which can help digitally transform the hiring landscape. These gaps

were addressed through a mixed-method approach combining quantitative surveys with

qualitative interviews of job seekers and human resource professionals. Data has been

collected from multiple organizations through surveys at different levels, secondary sources,

interviews, and relevant case studies. Here, Gibson’s Theory of Affordance serves as the

analytical structure supporting the study by evaluating user acceptance and engagement

with the AI-agent-based system. Future studies and cases should investigate the long-term

feasibility of integrating AI agents into HR operations, especially the recruitment process,

and evaluate the new challenges arising in other Bangladeshi companies.

Keywords: Articial Intelligence (AI), AI-Agent, Hiring, Opportunities, Challenges

Introduction

Recruitment or hiring is an essential part of any organization, enabling it to operate

and survive in a rapidly evolving, competitive environment based on the merit of its

people. Hence, recruitment is considered one of the most resource-intensive functions of a

* Corresponding Author: Labiba Fairuz Hassan, E-mail: 2130002@iub.edu.bd

5757

modern organization, involving multiple stakeholders at various levels, such as the Human

Resource team, Functional Personnel, Managers, and sometimes external units and third-

party agencies, culminating in a process that is simultaneously time-consuming, nancially

burdensome, and inconsistent in terms of evaluation. This complex process often leads

to delays, inefciency, and resource waste, while incurring signicant nancial and non-

nancial costs. In recent times, the rigorous implementation of the recruitment process has

come with lengthy procedures, signicant time and resource investments, human error,

and, sometimes, bias, especially in multinational companies that often work with large

pools of data.

On the other hand, AI is a disruptive technology in the modern era, making work

easier and more time-efcient. AI is said to help reduce time-consuming activities through

automation and the streamlining of redundancy, match job requirements with candidates’

existing skills more efciently, and enable faster, more effective decision-making

(Horodyski, 2023; Gusain et al., 2023). Additionally, such changes require considerable

effort from recruiters. To cope with these challenges, technological innovations such as

articial intelligence (AI) have been used in the recruitment process (Fritts and Cabrera

2021; Vedapradha et al. 2019; Geetha and Reddy 2018). Backed by data, this technology

has been used in the recruitment industry. According to a Statista survey (2017) of 8815

respondents, including talent acquisition professionals and hiring managers, 67% of the

respondents said that articial intelligence helps reduce the time-consuming nature of

the hiring process (Statista 2022). As a result, global corporations have made signicant

progress in using AI for Talent Acquisition, as reected in a report by Intelion Systems

(2023), which states that AI use in recruitment is growing, with 35% to 45% of companies

already using tools for this purpose. They also pointed out that as many as 99% of Fortune

500 companies were already using AI in some way in recruitment. However, Bangladesh

is far behind in practically adopting these new norms, particularly in large corporations.

This research explores the underexplored area of AI agent use in Bangladeshi Recruitment

systems and considers a signicant development in resource management and waste

reduction.

Global data underscore that AI tools improve organizational efciency and accuracy

and reduce human subjectivity (Mujtaba & Mahapatra, 2024). Moreover, AI integrates

the traditional people-oriented approach with a greater emphasis on data analytics

(Gelinas et al., 2022). This also evaluates recruiters’ awareness and willingness to adopt

this technology, taking into account the risks, training period, AI-borne inaccuracies, and

ethical and responsible implementation. It is evident that 88% of organizations worldwide

have experimented with AI in recruitment activities; among these, 41% employed AI-

based chatbots for candidate engagement, 44% used AI to identify candidates via social

media and public data, and 43% leveraged AI for training recommendations (Mujtaba &

Mahapatra, 2024). Therefore, the primary focus is to evaluate the efciency improvement

of AI agent recruitment in terms of time, cost, human effort, and errors in Bangladesh.

In general, AI agents are designed to reduce costs, human intervention, and time while

expediting the selection process, minimizing errors, coordinating with the operations

Integration of AI agents in recruitment: opportunities and challenges

Independent Business Review, Vol 15, No. 1 | June 2026 5858

team and other relevant stakeholders, and more. Therefore, this study positions itself as a

connecting point between two realities: global advancement and local barriers, examining

how AI agents can be meaningfully installed in the traditional recruitment system. It is

said that AI agents are becoming very popular globally, but empirical evidence remains

scattered and undocumented, especially in developing countries, where the struggle with

proper infrastructure and technical frameworks is very likely, and a limited understanding

of cost-benet implications is prevalent. Studies on similar developing countries remain

unexplored and outdated, narrowing the scope for further advancements and insights.

Hence, guided by Gibson’s Theory of Affordance as the theoretical framework, this study

focuses not only on stakeholders’ mindset towards adopting AI tools but also on how they

perceive these tools’ capabilities, based on organizational context, digital literacy, and

cultural disposition.

Research Objective

The study focused on two primary research questions, each having a distinct dimension

of the AI-agent-enabled hiring system in Bangladesh:

RQ 1: What will be the impact of AI agents on the conventional recruitment ecosystem

in Bangladesh with respect to cost analysis, feasibility, accessibility, technological

infrastructure, efciency, error reduction, and expectancies?

RQ 2: Will the implementation of AI agents create a notable change in organizational

and natural resource optimization, enhancing waste management and eco-friendly methods?

These questions were developed to capture the overall operational dynamics of

recruitment with AI-agent adoption. It also aims to assess broader socio-environmental

perspectives. These orientations present a blended theoretical grounding, methodological

structure, and interpretive design framework for the study.

Literature Review

Interest in deploying articial intelligence applications in human resource functions

has surprisingly grown among scholarly activities and larger companies, with AI adoption

expanding organizational footprints. Tambe, Cappelli, and Yakubovich (n.d.) noted that AI

is not a linear enhancer of productivity, but rather a disruptive technology that recongures

the entire decision-making process, accountability, and risk-management for work, which

is ultimately the most appropriate theoretical explanation over time. They highlighted that

AI-based decisions pose challenges that are often difcult to comprehend for both regulatory

bodies and candidates, further complicating the process and its legal compliance. In the

Bangladeshi context, this has been a particular observation: employment law has yet to

properly adapt across industries to the digitalized workplace realities; however, candidates

may have limited resources to address algorithmic decisions.

Mujtaba and Mahapatra (2024) provided a comprehensive, empirical assessment of

AI’s growing role in recruitment systems and selection processes globally, documenting

widespread organizational experimentation alongside persistent concerns about fairness,

5959

mitigation of bias, and the erosion of human judgment in hiring personnel. Their

observation is that AI-backed recruitment is not a technology-based solution but an

evolving sociotechnical pattern that requires consistent governance. Gelinas et al. (2022)

have complemented this analysis by tracing the momentum shift from people-centric

to data-centric HR decision-making, arguing that AI’s predictive analytics capabilities

are a representation of a genuinely new epistemology for talent assessment; that one

premised on pattern recognition across large datasets, different than that of the holistic

human evaluation approach. This shift has been cautioned, stating that it must be carefully

managed to ensure that the historical biases embedded in training data do not become an

algorithmic perpetuation of structural inequality.

However, different perceptions have emerged in studies of AI and its implications.

Akhter, Bhattacharjee, and Hasan (2024) provided indispensable grounding in the

Bangladeshi context, exposing the current status of AI adoption in HR management. They

also pointed out the structural and conceptual challenges that distinguish local implication

trajectories from global practices. The ndings showed a landscape in which large corporate

entities and their subsidiaries began precautionary engagement with AI-based tools, while

many domestic organizations, including medium- to large-sized enterprises, remained at

the “pre-adoption” (awareness) stage. Primarily, the documented structural constraints

concern infrastructure limitations, the lack of trained human resources in AI, and cost

analysis. Rahman et al. (2025) extended the discussion by proposing a structural framework

for AI-driven recruitment that integrates the objectives of inclusivity, sustainability, and

efciency, demonstrates a contextually grounded integration, and reconciles competitive

imperatives.

In contrast, AI agents are a comparatively new concept in the recruitment domain, which

are distinct from earlier AI-based tools because of their architectural sophistication and

operational scope. Earlier AI applications simply depicted keyword-matching algorithms

and structured questionnaire platforms, where they were used to perform discrete, rule-

bound tasks; whereas AI agents are meant for the automation of the process, with multi-

step action, dening the environment, forming and pursuing goals while making decisions

under an uncertain structure, and adapting the behavior as the response to feedback (Colbak,

2025). This quality represents a qualitative expansion of AI capability, with remarkable

implications for restructuring recruitment workows.

Koman, Kubina, and Borsos (2024) systematically document the current and emerging

possible applications of AI agents across the recruitment lifecycle. At the primary stage,

AI agents can autonomously scan professional networks, job boards, and social media

platforms to identify candidates whose proles match the role specications for a

specic position, often surfacing passive candidates who would not have applied through

conventional channels. At the screening stage, NLP (natural language processing) models

resume content, assess writing quality and coherence, and score candidate proles against

designated, weighted criteria, reducing manual review time by up to 75% in documented

case studies (Horodyski, 2023). During the assessment stage, AI-powered video interview

platforms analyze verbal content, speech patterns, and, more controversial, facial

Integration of AI agents in recruitment: opportunities and challenges

Independent Business Review, Vol 15, No. 1 | June 2026 6060

expressions, to generate candidate evaluations. Furthermore, scheduling, communicating,

and offering management can be handled by conversational AI agents (conversational)

capable of maintaining contextually appropriate dialogue across multiple candidates

simultaneously.

In another view, Colbak (2025) described the developmental trajectory, referring to AI

as a ‘co-pilot’ that augments human decision-makers with data and recommendations. AI

essentially acts as an ‘autopilot’, executing end-to-end workows with minimal to no human

intervention. These changes raise signicant governance questions about the tolerable extent

of AI autonomy in substantial employment decisions, as well as about the accountability

and regulatory frameworks needed when AI agents make errors and produce discriminatory

outcomes. However, Fritts and Cabrera (2021) and Vedapradha et al. (2019) examined the

organizational and ethical implications of this AI-powered transition, arguing that the shift

to autonomous AI agents in hiring requires the development of oversight frameworks, audit

mechanisms, and candidate selection processes.

The ethical reasoning of AI agents in recruitment is crucial and multidimensional.

Mori, Sassetti, Cavaliere, and Bonti (2025) conducted a literature review on the use of AI

in recruitment through the lens of ethical norms, which further identied three primary

concerns: rst, the replication of historical biases in training data; second, the transparency

of algorithm-based decision-making systems, delivering a meaningful candidate resource;

and lastly, the threat of potential replacement of human intervention and judgement in the

nal output that would have profound consequences for individual livelihoods and the

organizational culture of hierarchy and positions. Their analysis proposes an ethics-by-design

framework that integrates fairness, transparency, and accountability into AI ecosystems from

the outset, rather than treating ethical compliance as a post hoc corrective.

However, Lashkari and Cheng (2023) examined the transparency scale in AI-driven

hiring from a technical perspective, segmenting the metrics used in algorithmic operations

and reviewing bias-cancellation methods. One key nding is that multiple fairness criteria,

such as demographic parity, individual fairness, and equalized odds, are not mathematically

compatible, suggesting that satisfying one criterion may often require sacricing another.

This technical complexity underscores the limitations of algorithmic fairness resolution

through xes alone; with organizational commitments, the fairness concept takes priority

in this context. In the Bangladeshi landscape, where religious, gender, ethnic, and most

importantly, socioeconomic dimensions of inequality intersect in historically evident ways,

these decisions have a profound impact on errors.

Marabelli, Newell, and Handunge (2021) contributed to research with a wider perspective

on the ethical dimension related to the algorithmic decision-supporting system, enforcing

the theory that the challenges would not normally appear during the deployment, rather

would arise gradually across the system’s cycle, as the technology interacts with the work

demographics. This note is important for organizations in Bangladesh in the context of rst-

generation AI-agent adoption: the ethical risks of this AI-agent-based recruitment are not

fully acknowledged. Therefore, a frequent monitoring framework, periodic evaluation, and

6161

adaptive governance are required beyond a single compliance assessment.

The theoretical structure of this study was developed based on Gibson’s (1979) “Theory

of Affordance,” which originated in ecological psychology and often describes the probable

actions evident in the environment relative to their capacities. Translating to the domain

of information systems and technology adoption, affordance theory states that “The ‘use

value’ of a technology is not inherent in the technology itself but emerges relationally

from the interaction between technological features and user attributes, which includes

capabilities, motivations, expectations, and organizational context” (Faraj & Azad, 2012).

An AI-agent-based hiring chatbot, for instance, affords rapid, low-friction, and less redundant

communication for a digitally equipped and literate HR manager but may cause confusion,

exhaustion, and distrust for a candidate with limited smartphone access.

This relational theory species that affordance theory is particularly well-suited to

analyze AI-agent adoption as heterogeneous in Bangladesh’s corporate sector, where digital

literacy, infrastructure access, and organizational culture vary signicantly across companies

and stakeholder groups. The theory empowers the study to move beyond binary adoption and

non-adoption queries, asking about the different approaches stakeholders take to engage with

AI-powered recruitment tools and how organizational and policy interventions can realign

perceived affordances with desired outputs.

The concept of advancing AI- or AI-based technology in recruitment originates in

economies with higher GDP, higher income, and technological maturity, where organizational

functions are digitally mature, regulatory sophistication is advanced, and the labor market

structure differs signicantly from that in Bangladesh and similar South and Southeast Asian

economies. This geographic and institutional imbalance in the study, with no proper further

investigation specically on AI-agent usage, is the research gap addressed in this study.

Rahman et al. (2025), however, represented a valuable exception, providing a study with

a proposed framework suitable for the South Asian context, which foresaw the AI-driven

sustainable recruitment foregrounding local conditions such as the heterogeneity of digital

infrastructure, dynamics of the skills market, regulatory, and an adoption trajectory with a

so-far promising outcome.

Geetha and Reddy (2018), in a pre-foundational analysis of AI-based applications in

the recruitment process, developed a prototype of AI-powered hiring tools, which remains

analytically useful for emerging markets, although it depicts the current state of the AI

agent paradigm. It has been emphasized that organizational readiness denes the alignment

of leadership commitment, tech infrastructure, and workforce capability, all of which are

prerequisites for an effective AI-enabled environment. AI adoption has yet to stabilize in

Bangladesh’s work culture, where a few big corporations are integrating AI tools into daily

operations, while many are still on their way to building the foundational tech competencies

needed to leverage advanced AI systems. This readiness gap in the pre-stage is rather

institutional, encompassing data governance practices, ethical considerations, readiness to

adopt the digital shift, and legal and compliance policies for a more responsible AI-backed

ecosystem.

Integration of AI agents in recruitment: opportunities and challenges

6262

Research Methodology

Project Overview and Approach

The team carried out a project during the internship, focusing on the implementation

and analysis of AI agents in the recruitment processes of participating organizations

in Bangladesh. The goals of the project were to design, model, and test a prototype

recruitment system using AI agents and evaluate the practical feasibility, efciency

gains, and organizational impact of the system in the context of the Bangladeshi business

environment.

Project Approach

The internship project was a mixed-methods study that incorporated system design,

stakeholder engagement, and empirical observation.

Recruitment System Design

The research team created and tested a planned and efcient recruitment system that

included the following ve AI agents:

• Agent 1: (Campaign Start for Recruitment): This agent streamlines the

commencement of recruitment campaigns to ensure that they are regular and on

schedule.

• Agent 2: (CV Collection & Pre-screening): The use of this agent will save CV

screening over 70% of the manual time, allowing the ability to quickly identify

qualied candidates.

• Agent 3: (Appointment Data Preparation): This agent helps enhance logistics,

reduces the administrative burden, and tests a standardized procedure for

interviews.

• Agent 4: (Transmittal of the Final Document): This agent simplies the paperless

recruitment process by allowing appropriate digital ling and providing an audit

trail for compliance.

• Agent 5: (Analysis). This agent can be regarded as a powerful tool for generating

data-driven insights, benchmarking performance, and providing process

optimization suggestions.

Data Collection Methods

System Design Documentation

Process mapping was used to systematically and comprehensively document the

current recruitment processes in partner organizations. This involved planning and dening

integration points for each AI agent, specifying the technology infrastructure and tools,

creating user interface mockups, and outlining system ow diagrams. In addition, the

integration specications and API requirements are outlined.

Independent Business Review, Vol 15, No. 1 | June 2026

6363

Stakeholder Interviews & Feedback Sessions

Informal interviews and discussions with HR managers and recruitment professionals

from the pilot organizations (company representatives) (n = 15-20), as well as with hiring

managers, team leaders, and job seekers who had contact with the AI agent system were

conducted. Face-to-face discussions were conducted during these sessions.

Practical Testing and Observation

Direct observation and testing of the prototype system were conducted, followed by

trials with real recruitment data (with appropriate permissions). This included taking time

to process each percentage point in the automation process, as well as recording manual

interventions, technical problems, system constraints, output quality metrics, and total

resources used.

Document Review

Existing documentation from the organization was analyzed, including current

recruitment workow documentation, CV Evaluation Criteria and scoring sheets, Hiring

Statistics, recruitment timelines, technology infrastructure assessment, and cost of current

recruitment processes.

Quantitative Analysis

Several objective measures, including time reduction, were used to evaluate the output.

This included the time to screen the CV, the time to schedule an interview, and the overall

time to reduce the recruitment cycle time. The manual workload saved was estimated to

be approximately 70% of the previous workload. A cost analysis was conducted to assess

the monetization of tool costs and reduction in staff time costs. The total cost of ownership

(TCO) was also analyzed, and the cost-per-hire reduction was calculated and compared

with the current setup. For example, a reduction in paper consumption was estimated by

analyzing current paper consumption, which was estimated to be no more than 96.13% of

the paper when digitized. The environmental impact was also quantied.

Qualitative Analysis:

We collected and analyzed stakeholder feedback by coding these responses into

opportunities, concerns, and barriers. The comments were grouped into stakeholder groups

(candidates and HR managers). We also gathered common pain points, pain-point solutions,

and possible recommendations for system improvements, and attempted to document the

implementation challenges.

Technical Performance Assessment:

The evaluation included several areas of system performance, such as the accuracy

of AI candidate matching and screening results, reliability and uptime of the integration,

effectiveness of integration across platforms, ease of use of the UI/UX, errors and system

failures, and scalability assessment.

Integration of AI agents in recruitment: opportunities and challenges

6464

Validity and Reliability:

During system testing, each agent was independently evaluated using sample data.

Full end-to-end testing was conducted using real recruitment scenarios to evaluate the

validity of the AI’s suggestions. Edge cases and system limitations were uncovered, and

any bugs and workarounds were documented. Moreover, user feedback was veried by HR

professionals through system designs and requests for their suggestions for improvement.

Testing was conducted with real users, and design improvements were made iteratively.

Practical Constraints and Limitations:

Various problems in the eld were identied during the design and testing of this system.

These encompassed data quality issues, compatibility constraints with legacy systems,

differences in digital literacy among users, opposition to change within organizations, and

privacy and data security concerns unique to Bangladesh. Some limitations of the project

that affected the results were limited integration with legacy HR systems, API capacity

limits, the cost of AI tools, inadequate data quality in some enterprises, and limited time

for thorough system tuning. Due to time constraints and the system’s status as a prototype,

testing was conducted with a small sample size. It is worth noting that the situation in the

Bangladeshi business sector may differ from other markets, and AI technology is constantly

evolving, which may make the ndings of this study obsolete in the future. Moreover, there

is limited experience with AI agents in the local job market, and the regulatory landscape

for AI in Bangladesh is still evolving.

Key Findings

Despite the contemporary study and data collection currently underway, the anticipated

results suggest that AI agent installation in corporate hiring will signicantly enhance

recruitment efciency and has the potential to optimize resources, indicating a positive

change in terms of the sustainability mandate.

Additionally, this study will help assess the reduction of human bias and error, provided

that Bangladesh develops a relatable technological infrastructure, an accessible interface,

increasing awareness, and a regulatory framework. The results are expected to contribute to

modernizing the current hiring system, accelerate automated, data-driven decision-making,

improve internal coordination, and promote stakeholder education on the use of AI. However,

this prospect may also face barriers and backtracks, including recruiters’ concerns about

job security, transparency, cost analysis, lack of readiness for transformation, adoption

of new technologies, reliability, and discrepancies in AI-generated decisions. Notable

variations may be visible between global practices and local implementations, given the

varying conditions of technological advancement, infrastructure, and the economy.

Independent Business Review, Vol 15, No. 1 | June 2026

6565

Figure 1. AI-Agent System Modelling (Authors’ Developed)

A distinctive contribution of this study is its focus on AI recruitment through a

sustainability lens. Traditional recruitment in Bangladesh results in substantial paper use

throughout the process. AI agent installation is expected to reduce paper consumption,

signifying the organizational environmental, social, and governance (ESG) mandates

that are gradually becoming important within multinational corporations operating in

Bangladesh. Apart from paper reduction, AI-backed recruitment is said to optimize

energy and human resource expenditures associated with in-person early-stage interviews,

resulting in lower carbon footprints.

However, several barriers hinder AI adoption. One of the most signicant is algorithmic

bias, because AI systems are often trained on historical data on the hiring process,

which may reinforce the prevailing gender, socioeconomic, and educational biases. For

Bangladesh, this scenario could perpetuate the pre-dened preferences for candidates

from elite institutions and urban backgrounds. Moreover, data privacy and security are

Integration of AI agents in recruitment: opportunities and challenges

6666

major concerns, as AI agents will rely on enormous amounts of sensitive and personal

information, while Bangladesh currently lacks comprehensive, structured data protection

regulations. Moreover, resistance from hiring personnel and team leaders may seem evident

due to concerns about job security, reliability, transparency, and motivation to change the

system. Another important challenge is AI’s limitations, which mean it cannot fully capture

the contextual judgment, cultural understanding, and interpersonal insights that humans

possess.

The study targets to show both theoretical and practical views on AI agent

implementation: Theoretically, it uses the structure of Gibson’s Affordance theory on

the recruitment domain and related stakeholders’ perceived idea; on the other hand, the

practical part proposes a recruitment model that being AI Agent-enabled considering the

insights, ethical safeguards, and respective concerns ensuring fairness and adaptability.

Discussion And Implications

The expected contribution of this study goes beyond Bangladesh’s recruitment

context, offering insights into how developed economies can navigate optimal use,

ethical considerations, organizational challenges, and compliance with AI-agent adoption

in the hiring lifecycle. The study is based on a theoretical approach, mostly by applying

Gibson’s Affordance Theory, explaining perceptions of AI-agent-powered human resource

management, and shifting the prime focus from individual technological acceptance to

the broader intersections among stakeholders and AI tools, resulting in organizational

efciency and sustainable practice. This perspective focuses on the advantages of AI agents,

the creation of automated functions across various scopes, and the challenges at different

layers of operations. The study also proposes a prototype experimental model of an AI

agent-enabled recruitment system tailored to the multinational corporation environment of

Bangladesh, following a widely used recruitment process. This design can also be modied

to meet future implementation requirements in other emerging economies in the region. The

important views of this study include Bangladesh’s regulatory readiness, policy framework,

data protection laws, algorithmic accountability standards, and certication frameworks

for a just and transparent recruitment process, which are outlined as gaps in this research.

This requires coordinated and collaborative efforts from industry experts, policymakers,

and regulatory agencies. Deploying AI agents for functional work will not be enough for

any entity; this study also highlights the signicance of AI literacy among employees,

AI readiness, and an adoption mindset, which are crucial for full-edged operation. In

the long run, AI-agent-based recruitment will enable organizations to work towards their

sustainable goals through a paperless work system and ESG-aligned practices, which are

also applicable to the apparel, banking, RMG, and other sectors.

Independent Business Review, Vol 15, No. 1 | June 2026

6767

Table 1. Summery of Opportunities Challenges

OPPORTUNITIES CHALLENGES

Substantial Time Savings

- CV screening

- Interview scheduling

- Overall cycle time reduction

Algorithmic Bias Risk

- Systems trained on historical data may

perpetuate biases

- May disadvantage candidates from non-elite

institutions

- Requires ongoing bias auditing and mitigation

Cost Reduction

- Estimated 20-30% reduction in cost per

hire

- Reduced administrative staff time

Data Privacy and Security Concerns

- Bangladesh lacks comprehensive data

protection laws

- Sensitive personal information at risk

- Requires robust data security infrastructure

Paper Elimination

- 96.13% reduction in paper consumption

- Environmental sustainability benets

- ESG compliance and corporate

responsibility

Change Management Resistance

- Staff concerns about job security

- Preference for traditional hiring methods

- Requires signicant training and organizational

change

Improved Candidate Matching

- Objective evaluation standards

- Consistency across all candidates

- Better matching of skills to requirements

System Limitations

- Cannot fully capture contextual judgment

- Misses cultural t and interpersonal insights

- Accuracy issues with Bangladeshi CVs

Enhanced Efciency

- HR team can process 2.5-3x more

applications

- More thorough evaluation of qualied

candidates

- Reduced hiring backlog and delays

Legacy System Integration Challenges

- Many organizations have outdated HR systems

- API compatibility issues

- Requires workarounds and manual handoffs

Better Resource Optimization

- 70% reduction in manual work

- Frees HR professionals for strategic tasks

- Enables focus on employer branding and

retention

Infrastructure Limitations

- Digital infrastructure varies across Bangladesh

- Limited AI expertise and technical capacity

- Cost-benet analysis challenges for SMEs

Data-Driven Decision Making

- Real-time analytics dashboards

- Source of hire analysis and metrics

- Evidence-based recruitment strategy

Regulatory Uncertainty

- Bangladesh AI governance framework

incomplete

- Compliance requirements unclear

- Legal liability for algorithmic decisions

Scalability and Growth Support

- System handles high-volume recruitment

- Maintains quality as company grows

- Enables expansion without proportional

staff increase

Stakeholder Skepticism

- HR professionals worry about accuracy

- Candidates concerned about fairness

- Requires transparent communication and trust-

building

Integration of AI agents in recruitment: opportunities and challenges

6868

Transparency and Appeal Mechanisms

- Clear audit trails of decisions

• Can document decision criteria

• Enables appeal and review processes

System Errors and Failures

- AI recommendations sometimes inaccurate

(False positive rate)

- Technical bugs and integration failures

- Requires human oversight and verication

Organizational Learning and Innovation

- Positions organization as forward-

thinking

- Attracts tech-savvy talent

- Competitive advantage in talent

acquisition

Training and Skill Gaps

- Staff need training on new systems

- Limited AI literacy in Bangladesh

- Ongoing support and troubleshooting required

Fairness and Bias Monitoring

- Can implement fairness constraints

- Enables regular bias audits

- Maintains diversity hiring goals if

designed properly

Cost Barriers

- Implementation costs high for small

organizations

- Premium AI tools expensive

- ROI difcult to justify for lower-volume hiring

Process Standardization

- Consistent hiring procedures across

organization

- Reduces subjective decision-making

- Better compliance with hiring standards

Incomplete Solution

- AI agents cannot make autonomous nal

decisions

- Still requires human review and approval

- Efciency gains limited without human

oversight

Limitations And Future Directions

This study acknowledges the limitations of the scope of future research. The mixed-

methods design is theoretically robust; however, the actual implications may vary during

practical implementation, with concerns such as human adoption behaviors, unforeseen

organizational bottlenecks, and counterintuitive stakeholder perceptions. This may create a

divergence from the anticipations noted earlier. Since the AI-Agent has very little historical

evidence in the South Asian region, there are very few empirical studies. This led to a

completely new theory of emerging AI-agent-based perceptions proposed in this study. On

the other hand, this study has focused on Bangladeshi multinational corporations and their

recruitment processes, thereby imposing limitations on other sectors within and beyond

the global context. Recruitment stages vary across companies; however, the proposed

model and ndings can apply to most entities. Additionally, Bangladesh provides a specic

reference to technological infrastructure, feasibility, and digital conditions, which may

differ from other orientations.

Conclusion

This study examines a theoretically proven experiment of recruitment processes with

appropriate AI agent integration, with an analytical view of Gibson’s Theory of Affordance.

It emphasizes Bangladesh’s emerging economy, which has remained underrepresented in

existing studies to date. This research aims to provide concrete insights, especially those

that are locally relevant, and provides an equation for broader implications, navigating

Independent Business Review, Vol 15, No. 1 | June 2026

6969

the opportunities and challenges in the landscape. The study also highlights the AI-driven

workforce transition, showing improvements in operational time, human intervention, and

administrative workow, while promoting sustainable practices. Key constraints, such as

algorithmic bias, data mismanagement, AI errors, data protection policies, infrastructure

status and limitations, digital literacy, and cultural resistance, must be addressed through

responsible, appropriate implementation strategies, including periodic audits, transparency

practices, and necessary learning sessions. Despite the challenges, this study does not

discourage adoption; rather, it offers an optimistic path to integrating the AI agent into the

recruitment process, identifying the long-term benets, which can bring a paradigm shift

to industry operations.

References

Akhter, F., Bhattacharjee, A., & Hasan, A. (2024, January). Application of articial intelligence in

human resource management: A Bangladeshi perspective. Munich Personal RePEc Archive. URL:

https://mpra.ub.uni-muenchen.de/122222/

Bonti , M. (2025, April 7). A systematic literature review on articial intelligence in recruiting and

selection: A matter of ethics. In M. Mori, S. Sassetti, & V. Cavaliere (Eds.), Personnel Review,

54(3). DOI: https://doi.org/10.1108/PR-03-2023-0257

Borsos, P. (2024, July 19). The possibilities of using articial intelligence as a key technology

in the current employee recruitment process are discussed. In G. Koman & M. Kubina (Eds.),

Administrative Sciences, 14(7), 157. DOI: https://doi.org/10.3390/admsci14070157

Colbak, L. (2025, May). AI agents: From co-pilots to autopilots. Financial Times. URL: https://

www.ft.com/content/3e862e23-6e2c-4670-a68c-e204379fe01f

Faraj, S., & Azad, B. (2012). The materiality of technology: An affordance perspective. In P. M.

Leonardi, B. A. Nardi, & J. Kallinikos (Eds.), Materiality and organizing: Social interaction in

a technological world (pp. 237–258). Oxford University Press. DOI: https://doi.org/10.1093/

acprof:oso/9780199664184.003.0009

Fritts M., Cabrera F. (2021). AI recruitment algorithms and the dehumanization problem. Ethics and

Information Technology, 23(4), 791–801. DOI: https://doi.org/10.1007/s10676-021-09589-z

Geetha, R., & Reddy, D. M. (2018). Recruitment through articial intelligence: A conceptual study.

International Journal of Mechanical Engineering and Technology, 9(7), 63–70.

Gelinas, J., Vézina, M., & Dumont, B. (2022). Data-driven human resource management:

Prospects and challenges. Human Resource Management Review, 32(2), 100832. DOI: https://doi.

org/10.1016/j.hrmr.2022.100832

Gibson, J. J. (1979). An ecological approach to visual perception. Houghton Mifin.

Handunge , V. (2021, September). The lifecycle of algorithmic decision-making systems:

Organizational choices and ethical challenges. In M. Marabelli & S. Newell (Eds.), The Journal

of Strategic Information Systems, 30(4), 101695. DOI: https://doi.org/10.1016/j.jsis.2021.101695

Horodyski, P. (2023). Applicants’ perspectives on AI-assisted recruitment: Opportunities and threats.

Frontiers in Psychology, 14, Article 1020270. DOI: https://doi.org/10.3389/fpsyg.2023.1020270

Lashkari, M., & Cheng, J. (2023, January 27). Fairness in AI-driven recruitment: Challenges, metrics,

methods, and future directions. arXiv Preprint arXiv:2405.19699. DOI: https://doi.org/10.48550/

Integration of AI agents in recruitment: opportunities and challenges

Independent Business Review, Vol 15, No. 1 | June 2026 7070

arXiv.2405.19699

Marabelli, M., Newell, S., & Handunge , V. (2021, September). The lifecycle of algorithmic

decision-making systems: Organizational choices and ethical challenges. The Journal of Strategic

Information Systems, 30(4), 101695. DOI: https://doi.org/10.1016/j.jsis.2021.101695

Mori, M., Sassetti, S., Cavaliere, V., & Bonti , M. (2025, April 7). A systematic literature review on

articial intelligence in recruiting and selection: A matter of ethics. Personnel Review, 54(3). DOI:

https://doi.org/10.1108/PR-03-2023-0257

Mujtaba, B. G., & Mahapatra, S. (2024). Articial intelligence in recruitment and selection:

Organizational practices and ethical implications. International Journal of Human Resource

Management, 35(4), 1289–1315. DOI: https://doi.org/10.1080/09585192.2023.2267857

Rahman, M., Hossain, A., Miah, S., Alom, M., & Islam, M. (2025, January). Articial intelligence

(AI) in revolutionizing sustainable recruitment: A framework for inclusivity and efciency.

International Research Journal of Multidisciplinary Scope. URL: https://www.researchgate.net/

publication/388936696

Statista. (2022). According to talent acquisition professionals worldwide, articial intelligence

benets hiring. Statista Research Department.

Tambe, P., Cappelli, P., & Yakubovich, V. (n.d.). Articial intelligence in human resource

management: Challenges and a path forward. California Management Review. DOI: https://doi.

org/10.1177/0008125619867910

Vedapradha, K., Anuradha, J., & Hema, J. (2019). Machine learning techniques for recognizing the

emotional states of autistic children. Microprocessors and Microsystems 74, 103–111. DOI: https://