Why review papers get rejected: common pitfalls and how to avoid them

Review articles play an important role in the production of scientific knowledge in social sciences and beyond (Chen and Hitt, 2021; Cronin et al., 2025; Kunisch et al., 2023a; McMahan and McFarland, 2021). As a stand-alone research endeavor, review research (articles) ^[1] can be understood as “secondary data research” that uses scientific methods to produce new knowledge about a specific topic (Cronin et al., 2025; Kunisch et al., 2023a). In correspondence with their important function in the production of scientific knowledge, many academic journals publish review articles ^[2]. Yet, the rejection rates at respected journals in the field of business and management, including in the logistics and supply chain management domain, are often high, exceeding 90% ^[3]. This shows that, just like other forms of research, producing and publishing review articles is challenging.

To date, a wealth of advice exists to help scholars meet these challenges in producing review articles as stand-alone research. Some scholars have provided valuable insights into the overall research process (Fisch and Block, 2018; Petticrew and Roberts, 2008; Rousseau, 2024; Sauer and Seuring, 2023; Tranfield et al., 2003), and others have offered advice for individual steps, such as conducting a literature search (e.g. Adams et al., 2017; Gusenbauer and Gauster, 2025; Hiebl, 2023). Researchers have offered advice on specific literature reviews types, such as integrative reviews (Cronin and George, 2023; Elsbach and van Knippenberg, 2020; Torraco, 2016), critical reviews (Wright and Michailova, 2023), problematizing reviews (Alvesson and Sandberg, 2020), bibliometric reviews (Donthu et al., 2021; Lim and Kumar, 2023; Marzi et al., 2025; Zupic and Čater, 2015) and methodological reviews (Aguinis et al., 2023). Others have focused on types of knowledge contributions (Breslin and Gatrell, 2023; Durand et al., 2017; Krlev et al., 2025; Post et al., 2020) and quality criteria such as systematicity, transparency and rigor (Paré et al., 2016; Simsek et al., 2023, 2025). Collectively, these works provide a wealth of valuable advice for how to produce publishable review papers.

Still, many submitted review papers are rejected, and most of them are rejected at an early stage. This may be a tell-tale sign that scholars are overwhelmed by the abundance of advice, which is often rather technical, with a primary focus on conducting the actual research. Despite their merits, these works say little about what matters in the review process. Against this backdrop, the purpose of this invited article is to take a different inroad and focus on why manuscripts get rejected, including during early stages by editors (desk-rejection) or later stages (rejection after reviews).

Based on our own experience as authors, reviewers and senior editors, combined with a screening of author guidelines and editorials, desk-rejection decisions and reviewer comments, as well as discussions with several editors-in-chief of journals that publish review articles, we have identified six recurring reasons (common pitfalls) for why papers are rejected. For each of these pitfalls, we offer literature-based guidance on how to identify and mitigate them, along with illustrative examples of literature reviews that have successfully navigated these challenges. Our overarching argument is that publishable review articles need to be both systematic (i.e. rigorous and transparent) and generative (i.e. advance knowledge). This paper seeks to guide researchers to key issues and methodological references to reach these twin objectives and thereby reduce the likelihood of being rejected.

It is important to note that we focus on review articles that are submitted to scientific journals for publication, which represent a “stand-alone research project” (Krnic Martinic et al., 2019; Kunisch et al., 2023a; Rousseau, 2024). This means that they “employ scientific methods to analyze and synthesize prior research to develop new knowledge for academia, practice and policy-making” (see also, Krnic Martinic et al., 2019; Kunisch et al., 2023a, p. 3). Using the scientific method with the aim to make an original knowledge contribution is what “distinguishes review research [articles] from literature reviews that are not stand-alone research projects (e.g. in introductions to Ph.D. theses, empirical papers, or research grant proposals)” (Kunisch et al., 2023a, p. 11). As Kunisch et al. (2023a) argued: “Notably, the role of prior research and knowledge as the data is the key distinguishing feature of review research, which makes this type of research distinct ….” (p. 11) from other genres. We thus consider review articles as one type of research, next to other forms of empirical research ^[4], conceptual work and other forms of knowledge contribution.

It is important to note that review research is a broad tent that spans a variety of research purposes, review types and research methods (both qualitative and quantitative), which can be rooted in different onto-epistemological traditions with different quality criteria (Kunisch et al., 2023a). For example, integrative reviews and problematizing reviews differ substantially with respect to theorizing. As Patriotta (2020) noted, “The two approaches express a fundamental tension between two sets of expectations about theorizing (see Locke and Golden-Biddle, 1997): one supposes that reviewing organizes a knowledge space with the purpose of making it more accessible (integrative reviews), and the other that reviewing subverts that space with the purpose of extending the field of possibilities (problematizing reviews)” (p. 1274). Similarly, reviews that rely on quantitative and qualitative data and analyses can differ substantially. For example, consider the differences between meta-synthesis of qualitative studies (Hoon, 2013; Rauch et al., 2014) versus reviews that aim to examine “true” effect sizes across multiple studies using quantitative meta-analytical methods (Combs et al., 2019; Geyskens et al., 2008; Glass et al., 1981; Stone and Rosopa, 2017).

Yet arguably, two generic quality dimensions set all forms of review research apart as scientific inquiry, which we refer to as systematicity and generativity. Systematicity means that the (review) research process is rigorous and transparent (Paré et al., 2016; Simsek et al., 2023). A systematic approach implies a transparent, consistent and coherent research method across all steps of the research process including defining the research objective, carrying out search and selection (including theory-driven keywords, well-considered inclusion and exclusion criteria) across sources, as well as analysis and synthesis and detailed documentation of each step in the process (Durach et al., 2017; Rousseau, 2024). This level of rigor minimizes bias and allows others to understand, trust and (if permitted by the review type/method) reproduce the review research.

Generativity focuses on the contribution the review makes to refine and revise knowledge and conversation in the field (Bartunek and Lei, 2023; Fan et al., 2022). As reflected in the mission of many business and management journals, review research articles must advance theory (Wong, 2021). Rather than simply summarizing existing studies, a review that is generative manages to synthesize insights that help refine or revise theory (Breslin and Gatrell, 2023; Durach et al., 2021; Post et al., 2020). It transforms the review into a tool that contributes meaningfully to scholarly progress. Our use of “generativity” aligns with recent discussions on how review research can shape and advance theoretical conversations. For instance, Breslin and Gatrell (2023) proposed a miner–prospector continuum to describe different modes of review work. While “mining” focuses on extracting and organizing established knowledge, it tends to remain descriptive. In contrast, “prospecting” involves reinterpreting, reconfiguring, or challenging the literature to uncover new theoretical insights—this is where we locate “generativity.” Similarly, Kunisch et al. (2023a, b) outlined a range of review purposes such as classifying, interpreting, explaining and problematizing. These frameworks provide useful orientation for thinking about how review articles can meaningfully contribute to theory development, hence, be “generative.”

These two dimensions are orthogonal—neither precludes the other and impactful review papers successfully combine systematicity and generativity, meeting the expectations of editors, reviewers and readers (Figure 1). However, many submissions fall short by focusing narrowly on one dimension while neglecting the other. The shortcomings lamented in most, if not all, rejection letters speak to one or both dimensions, which in turn, underlines that “submission-ready review research” needs to employ a transparent, consistent and coherent method AND advance theory and conversation in the field ^[5]. Thus, in line with discussing the common pitfalls in review papers, we offer recommendations for how each of these pitfalls can be addressed, which, in turn, will foster systematicity and generativity. This approach should help authors aim for the top-right quadrant of Figure 1, producing a submission-ready review paper that is rigorous, transparent and promises a meaningful knowledge contribution.

In this section, we discuss six common pitfalls of review research. Each subsection begins with a brief description of the pitfall and typical manifestations (i.e. observable indicators or features of this pitfall) found in the review papers that appear to suffer from these pitfalls. For each pitfall, we then offer mitigation strategies (i.e. what can be done) and guide readers toward relevant methodological literature and review papers with illustrative examples that (partly) address the pitfall under discussion. We consider these mitigation strategies neither as silver-bullet solutions nor as mandates that must be implemented in all review articles. Rather, we consider them as exemplary ways for circumventing the pitfall under discussion. Table 1 provides a summary of these insights.

While some of these pitfalls resonate with phases or themes addressed in existing guidelines for conducting review research (e.g. Durach et al., 2017; Sauer and Seuring, 2023), our intent here is different. Those frameworks focus on “how to” conduct strong review research. Though, in the present article, we flip the lens to look at “why” review papers fail. Our goal is to offer reflective, experience-based insights into common issues in the review process. Our perspective is inherently shaped by our editorial vantage point, and we share it with the aim of helping authors better align their work with the expectations of journals and reviewers.

Also, please note that while we present the pitfalls as distinct categories for clarity, we acknowledge that they are often interrelated. For example, a weak conceptual foundation can lead to poor search strategies, and insufficient research design may undermine both systematicity and generativity. Our intent is not to offer a rigid typology but to flag common, recurring issues that undermine the quality of review manuscripts.

Probably the most common reason why review papers get rejected is that they lack a compelling motivation. Just like any other research paper, a publishable review article needs to create a “research space” (Swales, 1990) and contain a compelling “hook” at the beginning of a manuscript to help “readers fully recognize and appreciate what [the] research has to offer and intrigues them enough to read further” (Grant and Pollock, 2011). This includes addressing questions such as: (1) “who cares” or “so what”; i.e. why is the research topic interesting and important to address for research and practice; (2) what do we know, and what do we not know about the topic; and (3) what will we learn from the study; i.e. how does the study challenge or refine our understanding of the topic. Yet, many reviews lack a compelling motivation (maybe because reviews are often still not approached as research papers or because authors simply shoehorn what they have learned in line with other forms of research).

The following manifestations are frequently observed. First, the chosen topic is not suitable for a review (e.g. too narrow or too broad or little relevance for theory and practice) or the review focusses on a rather narrow topic that is only of interest for a small, specialized audience. Parmigiani and King (2019) noted the following about review proposals submitted to the Journal of Management: “The novelty and importance of the review topic (i.e. that recent reviews have not covered the same ground and that the ground itself is important) are weighed heavily. An appropriate balance of breadth and depth, given the particular topic and research focus, is also important; AEs commonly remark that a topic is too narrow or that a review is trying to cover too much ground” (p. 3085). Indeed, two out of six criteria for assessing proposals for review articles refer to topic choice: “(1) relevance, […] (3) scope of interest, […]” (Parmigiani and King, 2019).

Second, there is no (or a weak) argument for what the review aims to achieve and how it aims to advance or change a scholarly conversation. Sometimes papers simply state: “no prior review exists on this topic.” While the same or a similar review should indeed not exist, the absence of a review is certainly not sufficient to motivate a need for it. A related issue is the lack of a review purpose or too many different purposes (e.g. often in the form of many research questions). In the case of the former, it is not clear which problem (in the literature) should be addressed. For the latter, the paper tries to achieve too much beyond what can be achieved in a single research paper.

Third, there is no justification for why a review was the appropriate form of research (e.g. rather than a primary data study or a theory-building paper). It may come as a surprise for an experienced producer of reviews, but statements such as “there is no or little research on the topic” are not unusual. While this statement may suggest the need for some sort of research, it is a discussion stopper in a review motivation. It literally suggests that there is no/not enough literature to warrant a review; but a review uses prior research and knowledge as the data (Kunisch et al., 2023a). This does not mean that a minimum number of studies must exist (e.g. for a forward-looking review on an emerging topic, only a few studies may be enough) or that a certain number of studies is a good enough argument to motivate a review; but a compelling argument for why a literature review is the right method choice is indispensable.

To mitigate the risk of being rejected for a lack of compelling motivation, authors could consider the following: First, choose a suitable topic for a review. A suitable topic for a review should be interesting and important (e.g. concerns an important scholarly debate and/or practice issue; see also Colquitt and George (2011), Kunisch et al. (2023b)) and have appeal to a broader audience. Topics typically focus on a certain “research stream [… i.e.] a body of work that either (1) is focused on a specific topic or (2) applies a particular theory across different topics” (Ketchen and Craighead, 2023, p. 164). That is, topics could be inter-disciplinary, phenomenon-based, theory-driven and/or practice/policy focused. As an example, Soundararajan et al. (2025) argued: “In global supply chains, subpar working conditions are a critical issue affecting organizations, workers, civil society, and policymakers alike. Our objective is to evaluate the approaches to improving working conditions within global supply chains and their implications. Through a comprehensive review that integrates insights from various social science disciplines, we offer a fresh perspective on this challenge” (p. 230). A list of example topics for reviews in management is offered by Parmigiani and King (2019; see Table 2, p. 3086). While topic choice can indeed set a project on the right track early on, more needs to be considered.

Second, establish the common ground and current knowledge claims. Establishing what we know and what we do not know about the topic is key for review research and requires substantial knowledge about the topic (ex-ante). This could be achieved in different ways. One practice is to conduct a scoping study of the literature before initiating the review (see Durach et al., 2017, Step 1). For example, Kembro et al. (2018) established current knowledge claims about warehouse operations and design and thereafter synthesized literature from various fields such as logistics and supply chain management, operations research, information technology, retailing and marketing to investigate how the transformation to omnichannel logistics impact and pose contingencies for warehousing. Another practice is to list prior reviews. For example, Bapuji et al. (2024) included a table with related reviews and detailed how they differ from the study at hand (see Appendix 2, pp. 84–85).

Third, identify a knowledge problem in the extant literature that needs to be addressed through a review research article. In other words, describe a complication in the existing literature that calls for action (Lange and Pfarrer, 2017) and highlight the justified need (Tingelhoff et al., 2025) for a research project that uses the existing literature as data to develop new knowledge. This includes a convincing argument for why the authors chose to approach the research in the form of a review (Fisch and Block, 2018). As an example, Hohn and Durach (2023) argued: “a substantial share of the [socially sustainable supply chain management] research entails valuable background and context information … Hence, substantial information should be available across existing yet disconnected literature. Building on prior work, it is our aim to combine, integrate and extend these peripheral insights to formalize current knowledge and to move this topic to the center of theoretical discussion” (p. 14).

Fourth, formulate a review purpose. This could be done in the form of an explicit research question or an explicit purpose statement. As an example, Kembro and Näslund (2014) stated the following purpose: to evaluate “empirical evidence for [and against] benefits of information sharing in supply chains” (p. 180) and concluded “a clear disconnect between what is being researched, the concluded benefits of information sharing in supply chains, and what is recommended to supply chain managers” (pp. 192–193). This also includes stating the anticipated contributions and implications of the review (e.g. how the review might change a scholarly conversation). We recommend formulating one research question (Barney, 2018). For in-depth discussions about developing compelling research questions, we refer to Simsek et al. (2022) and Tihanyi (2020).

A weak or lack of a sound conceptual foundation (which affects the scope and boundaries, the empirical parts and ultimately the potential contributions of the review) is another important reason for submission rejection. The lack of definitional clarity and boundaries may result in a fragmented body of work that cannot be meaningfully synthesized (“comparing apples with oranges”). For example, when scholars focus on topics that are either inconsistently defined, such as “complexity” or “innovation” or not clearly differentiated from related constructs, such as “transparency” versus “visibility,” to use examples from the operations and supply chain domain. In such cases, the reviewed literature might use common terminology while referring to a wide range of distinct ideas and phenomena. These issues significantly hinder the researcher’s ability to achieve meaningful generativity and advance the field’s understanding. Without a sound conceptual foundation, the review is figuratively built on sand and has low potential to offer a meaningful knowledge contribution.

We observe the following typical manifestations in rejected literature reviews. First, authors often jump at search terms without discussing and delineating the underlying constructs and concepts. The absence of clearly defined constructs at the outset of the review can derail the project from the start. Basing a review on a vague and unmotivated set of keywords causes many issues later in the process. In most cases, it leads to a fuzzy and incoherent set of studies that may use similar terminology but address fundamentally different phenomena.

Second, other rejected reviews are based on poorly defined or too broad constructs. Sometimes, the concepts are not delineated from related concepts and literature. For example, the same term can have different meanings in different communities (e.g. “coding” in qualitative research versus in information systems) or different terms can have similar meanings (e.g. “strategic leadership” versus “top management teams”). In these cases, the weak conceptual foundation leads to a range of issues, such as unclear scope of the review and unclear inclusion and exclusion criteria. As a result, the sample of primary studies is questionable and/or it is difficult to make meaningful comparisons between the papers.

To mitigate the risks of being rejected for a lack of conceptual foundation, authors could do the following. First, start by defining and discussing the key constructs (e.g. in a dedicated background or conceptual foundations section) and their potential relationships (see Durach et al. (2017; step one, p. 70)). An example is Simsek et al. (2018) who first developed an encompassing definition of strategic leader interfaces and an organizing framework that interrelated context, contact and consequences and thereafter reviewed 122 prior studies across 3 decades. Another example is Madhavaram et al. (2023), who began with a detailed overview of the key constructs in their research domain, as well as the theoretical background of the underlying sub-processes. Only then did they embark on identifying the primary literature. Here again, they first defined their field, which guided subsequent decisions about which studies to include. For most review purposes, clearly defined concepts should be the starting point rather than the outcome. An exception may be when the review purpose is to clarify concepts (see, e.g. Castañer and Oliveira, 2020).

Second, specify the scope and boundaries of the review, that is what is in scope and out of scope (see, e.g. Armstrong et al., 2011). Elsbach and van Knippenberg (2020) offered a selection of impactful integrative reviews and their respective justifying condition and the boundary conditions of the review (see Table I, pp. 4–5). Defining the conceptual scope and boundaries not only provides authors with clear guidance for selecting literature but also offers journal editors and reviewers the necessary clarity to assess the review’s contribution (see, e.g. Fisch and Block, 2018 and vom Brocke et al., 2015). As an example, Ateş and Luzzini (2024) explicitly explained in a separate section that their review included studies on supply network complexity, encompassing dimensions like upstream complexity and the actively managed supply base, while excluding studies focused solely on internal or downstream complexity. In another example, Hohn and Durach (2023) conducted a review on how firms develop toward an integrative view on socially sustainable supply chain management. They first provided clear definitions of key constructs early in the review process and explicitly stated the boundaries of the review’s scope. Thereafter, they identified (i.e. selected) a “relatively narrow set of coherent studies, which allows for the consideration of study artifacts when developing generalizations with respect to the identified drivers of social SSCM development” (p. 17).

Another important reason why review papers are rejected is that the research design and the “big” methodological choices are not explained and justified. These choices are similar to other forms of empirical research (Bono and McNamara, 2011). However, while most scholars have probably taken a research methodology course during their PhD education, such courses rarely cover review methodologies. Moreover, the research design aspect is often overlooked in the plethora of “how to” advise, which focuses on specific types of or steps in the review research process.

Let us make a brief comparison with case study research: Once the case study is motivated as a research strategy, the next step is to describe and justify the research design. This includes the choice between a multiple-case study or a single case study (potentially with embedded cases), which depends on the purpose of the study as well as the underlying onto-epistemological views of the researcher. The design also (1) clarifies the unit(s) of analysis, (2) elaborates on purposeful sampling of suitable cases and (3) details how theory versus empirical data are used (inductive, deductive versus abductive study) to generate, elaborate, or test theory. Review research requires coherence and justification of methodological decisions (e.g. Seuring et al., 2021; Durach et al., 2021), similar to case study research, where design choices must be aligned with research purpose, theoretical framing and underlying epistemological assumptions.

Unfortunately, as Sauer and Seuring (2023, p. 1899) concluded, review research is often “considerably less stringently presented than other pieces of research.” Many authors fail to clarify how the study is designed and how their methodological choices (e.g. chosen type of review research; Durach et al. (2021) are aligned with the purpose/objectives of the review, the maturity of the research topic and the onto-epistemological view (Durach et al., 2021). As discussed by Simsek et al. (2023), the research quality criteria must “fit with their review goals and paradigmatic and methodological choices” (Paré et al., 2016, p. 496). These are typically “make or break” issues because the research design choices are big decisions that have implications for the entire research process and the quality criteria.

We observe several manifestations in rejected reviews. One is simply the absence of an articulation and justification of a “review type” (Durach et al., 2021), also referred to as “approach” (Krlev et al., 2025) or “review method” (Kunisch et al., 2023a; Rousseau, 2024). That is, authors jump from stating the review purpose directly to searching and selecting the papers without justifying or explaining the key research design choices.

Another manifestation concerns the misfit between the purpose or topic and the review type (Durach et al., 2021). For example, the review objective does not match the identified complication. An illustrative example is the choice between conducting an integrative review (Cronin and George, 2023; Elsbach and van Knippenberg, 2020; Torraco, 2016) or a problematizing review (Alvesson and Sandberg, 2020). This is an important decision that will have an impact on many of the following choices (e.g. identification and selection of papers). Another example is when the prior state of theory [nascent, intermediate, or mature (Edmondson and Mcmanus, 2007)] and the selected review type do not match. For instance, when reviews on mature topics employ an inductive theory-building approach or reviews on new research topics with a nascent state of theory sometimes inappropriately employ deductive, theory-testing methods (see Durach et al., 2021).

A third manifestation is the mismatch between the onto-epistemological assumptions and the research design and methods. This is important because different review types (e.g. meta-analysis and critical/problematizing reviews) are based on different “worldviews,” which have implications for the research quality criteria and perspectives on generalization. Durach et al. (2021) contrasted theory-testing reviews (e.g. using meta-analytical methods) and interpretive sensemaking, where the latter “would need to acknowledge the subjective perspectives of the different actors instead of attempting to generalize” (p. 1100). As another example, the onto-epistemological assumptions are also linked to the data collection and data analysis, including coding (inductive, bottom-up approach or pre-defined scheme) approaches. Other examples include the analysis (coding scheme) not aligned with the type of review (Durach et al., 2021) and a theory testing meta-analysis relying on inductively generated codes rather than on a pre-defined scheme.

To avoid these pitfalls, authors may want to consider the following. First, choose an appropriate review method to ensure fit between the review purpose and method. That is, reflect on the topic and the stated purpose/research question of the planned review study to determine what would be the most appropriate research design. Motivating this choice is important because, as discussed by Grant and Booth (2009, see Table 1), each review method has its strengths and weaknesses as well as idiosyncratic steps for selecting, evaluating and synthesizing the reviewed literature. An illustrative example is the choice between conducting a scoping review (Arksey and O’Malley, 2005), an integrative review (Elsbach and van Knippenberg, 2020), or a problematizing review (Alvesson and Sandberg, 2020). For example, Watson et al. (2025) conducted a scoping review on structures within organizations and justified this choice as follows: “it is particularly suited for capturing the breadth of evidence for topics with a broad scope (Arksey and O’Malley, 2005) that do not have an existing encompassing framework (Simsek et al., 2022) and that have not been reviewed comprehensively before (Arksey and O’Malley, 2005; Bolino et al., 2024)” (p. 311).

Second, choose a review method that fits with the prior state of theory. This choice requires a good understanding of the extant literature to determine whether the topic has been nominally or extensively researched. Simply put, an inductive review is more appropriate for new research topics with a nascent state of theory, while a deductive, theory-testing method is a better fit for mature topics (such as supply chain integration and information sharing). For guidance, authors can refer to the consistency model described in Durach et al. (2021; see Figure 5). As additional guidance, Krlev et al. (2025) proposed “two dimensions [degree of reflexivity and substantiveness] to develop a directional space that will help authors fit their articles to the status of the field they are reviewing, instead of blindly following a single idealized model or procedure” (p. 1; see also Figure 1, p. 9, and Figure 2, p. 12).

Finally, authors could explicitly discuss the onto-epistemological assumptions (“worldviews”) for their review and research design. These assumptions also guide the quality criteria for the research (compare theory testing and interpretative sensemaking; Durach et al., 2021). For example, Denyer and Tranfield (2006) illustrated the onto-epistemological assumptions of meta-ethnography as follows: “Meta-ethnography is driven by interpretation, not analysis, is seen as an alternative to the positivist paradigm and assumes that the social and theoretical contexts in which substantive findings emerge should be preserved through the synthesis (Noblit and Hare, 1988, pp. 5–6)” (p. 221). These assumptions have implications for research design. As a rare example, Habersang et al. (2019) clearly described and justified the chosen qualitative meta-analysis research design: “[This research design] is particularly suitable to develop process models of organizational failure for the following reasons. First, qualitative case studies provide rich, contextualized empirical descriptions of the dynamics of a single setting across multiple levels of analysis. […] Second, […] drawing on a large number of case studies allows us to reconcile previously disparate and irreconcilable empirical evidence about processes of organizational failure and therefore provide more robust, generalizable and comprehensive findings. Third, […] With a qualitative meta-analysis researchers can identify new relationships between concepts which existing theory did not account for …” (p. 22).

Another reason for rejecting papers refers to empirical issues. Like other forms of empirical research, reviews may suffer from weaknesses along various steps of the research process. This involves shortcomings in the search and selection as well as in the analyses and synthesis of the literature (Breslin, 2024). For example, key reasons why papers were rejected at International Journal of Management Reviews included: “Poor coverage of the literature with major gaps in key areas (17%)” and “Omitted key journals from review (5%)” and “Weak analysis of the literature – too descriptive (21%)” (Jones and Gatrell, 2014, p. 255). While some of these issues are more substantial than others, many of them could potentially be addressed in revisions. However, it is often not just one issue that makes or breaks the paper (i.e. leads to a rejection) but the number and combination of issues. Some issues, such as unmotivated search and selection (e.g. keywords and journals), may require a total redo of the selection of papers, which in turn would also require a complete redo of the analysis and synthesis.

The manifestations in rejected reviews relate to intransparency in search, selection, data analysis and synthesis. In terms of search and selection, one manifestation concerns the choices and (lack of) justification for sources (Gusenbauer and Gauster, 2025; Hiebl, 2023)—such as why few top journals or dubious ones were chosen, why one database or many, why a specific timeline or none and the rationale for specific inclusion and exclusion criteria. It also concerns the choice to include or exclude grey literature (Adams et al., 2017). Another common manifestation concerns the search terms (keywords). Frequently, the search terms lack a conceptual foundation, which creates several issues (see section 3.2). These manifestations undermine the quality of the review research (i.e. transparency, consistency, reliability and validity), where biased choices may lead to biased findings/conclusions.

In terms of data analysis and synthesis, probably the most prevalent shortcoming is a lack of depth. One of the reviewers for one of our own papers once commented on the analysis, “Your review is a mile wide and an inch deep,” and they were correct. Superficial analysis often leads to merely listing existing studies as annotated summaries rather than creating novel insights which require in-depth analysis. This applies especially to reviews that use qualitative approaches that often lack profound analysis techniques focused on textual data such as thematic analysis (Braun and Clarke, 2006; Fereday and Muir-Cochrane, 2006; Nowell et al., 2017), categorizing (Pierce, 2025), topic modeling (Hannigan et al., 2019) and constructing interpretation (rather than analysis in meta-ethnography) (Hoon, 2013; Noblit and Hare, 1988). While many of these techniques stem from qualitative research, arguably, there is still a need for methodological advancements focused on specific techniques and templates for reviews. Of course, reviews that rely on quantitative approaches such as meta-analysis (Villiger et al., 2022) and bibliometric analysis (Donthu et al., 2021; Marzi et al., 2025; Zupic and Čater, 2015) also need to address various challenges in the implementation.

To avoid being rejected for these reasons, authors could consider the following. First, employ state-of-the-art review methods and maybe even go beyond and thereby make a methodological contribution. In terms of search and selection, exemplary good practices include, among others, fit-for-purpose search and selection by using pioneering works as an anchor for search (Cronin and George, 2023), using gray literature (Berrone et al., 2023) and combining academic and practice literature (e.g. Stouten et al., 2018). Other scholars have used a combination of different search strategies (Cronin and George, 2023; Schätzlein et al., 2023; Schlütter et al., 2024). Making considered and justified choices is key. In terms of analyses, good practices include, among others, using templates in qualitative approaches such as the Gioia method (Hurmelinna-Laukkanen and Yang, 2022; Jeon and Maula, 2022; Kerr and Coviello, 2019) and coding for outcomes and mechanisms (Bapuji et al., 2024, see Appendix 1, p. 83). It is important to note that the bar in terms of methods is constantly rising, and what was sufficient for publication in the past may not suffice in the future. This is true for all research, and it seems especially true for review articles, which are still rather new in management research (see, e.g. Rousseau, 2024). Thus, we encourage scholars to keep an eye on methodological advice (e.g. in Organizational Research Methods and International Journal of Management Reviews).

Second, emphasize reporting and transparency. This facilitates “an audit trail for the reader” (Breslin, 2024, p. 5), which clarifies the “what,” the “how” and the “why” in terms of stating, clarifying and motivating all steps and choices for searching, analyzing and synthesizing (e.g. keywords, inclusion criteria and coding structure). For this purpose, authors may consider Simsek et al.’s (2023) protocol with multiple considerations for envisioning, explicating, executing, evaluating and encoding reviewed papers (see Table 4, pp. 313–314) as well as other reporting guidelines specified by the Cochrane Foundation and the Campbell Foundation.

Reporting and transparency can be addressed on two levels: the overall process and specific steps. The former concerns reporting the entire process by including flow charts and decision trees (e.g. Simsek et al. (2018, see Figure 1, p. 287) and Marculetiu et al. (2023, see Figure 2, p. 261). The latter concerns providing further details about individual steps of the process, such as the analysis. This includes important characteristics (e.g. unit of analysis, unit of data collection and study context) that must be reported to enable the reader to understand if/how papers can be compared for the analysis and synthesis (Simsek et al., 2023, 2025). For example, Kembro and Näslund (2014) analyzed the reviewed papers based on unit of data collection and research method versus unit of analysis (see Table V, p. 185), while Salmon et al. (2023) “reviewed 90 years of research on management and organizations in indigenous communities … [and] coded studies in several iterations to reveal cross-cutting themes” (Rousseau, 2024, p. 400). The reviewed publications were displayed by theoretical frameworks and thematic areas (see Salmon et al., 2023, Table 3, p. 445).

Third, address potential biases that can occur in the steps of the review process, including retrieving, selecting, analyzing and synthesizing the literature. Rousseau (2024) highlighted risk of bias (see Step 4 on “Two facets of study quality are methodological appropriateness and risk of bias”) and Durach et al. (2017) proposed remedies for four key biases for review research, namely sampling bias, selection bias, within-study bias and expectancy bias (see Table 2, p. 77). Thus, like other forms of empirical research, addressing potential biases can help increase trustworthiness in the review research and thereby increase the odds of not being rejected.

To be publishable in leading academic journals, review articles must make a novel and significant knowledge contribution and offer value. This means that reviews should produce new scientific knowledge (i.e. be generative), which in business and management typically means refining and revising theories (methodological reviews are an exception). This requirement is explicitly stated in many editorial guidelines at leading journals such as the Academy of Management Annals, International Journal of Management Reviews, Journal of Management or Journal of Management Studies. The Journal of Management Studies mission statement says that to be publishable, review articles must make a substantive contribution to theory, that is, they must “advance conceptual and empirical knowledge, and address practice in the area of management and organization” (JMS website; see also Post et al., 2020). Yet, many papers lack novelty and significance in providing new insights, such as proposing original frameworks, rethinking current knowledge claims, challenging conventional wisdom, or contributing as conversation changers; that is, they offer little new or different from what we already know (Hoon and Baluch, 2020). Ketchen and Craighead (2023, p. 168) stated in the Journal of Business Logistics that “many good literature reviews appear in leading journals, but the volume of exemplars pales in comparison to the number of reviews that offer little value beyond amassing articles together” (p. 168).

Indeed, part of the problem is that while “making a contribution” seems intuitive, what it means is often less clear because the concept of “contribution” itself is rather ambiguous. For our purpose, we mean the “intellectual product” and thus the review’s “utility” (see also Craighead and Ketchen, 2024) for scholars, practice and policy-making. Most scientific journals in management prioritize theories as a major contribution (Chen and Hitt, 2021; Hambrick, 2007). Thus, although reviews can also make “methodological contributions” (e.g. methodological reviews) and “contributions to practice and policy,” our discussion here focusses on “theoretical contributions and implications.”

We observe several manifestations in rejected reviews. First, review studies are too descriptive, focusing on descriptive insights (e.g. statistics) of the studies’ meta information and basic characteristics, including frequencies of topics, methods, theories, journals, key authors and citations. While descriptive insights can be interesting, they mainly describe the sample and thus could be presented in the method section or appendix and commonly do so without engaging critically with contradictions, inconsistencies or silences in the literature (Hoon and Baluch, 2020). Other reviews do a bit more by identifying and categorizing the subjects of existing research, stating what has or has not been studied. Still, these reviews do not deepen theoretical insight. Tranfield et al. (2003, p. 207) said that such descriptive reviews foster “tolerance to loss of knowledge.”

Second, many reviews offer only annotated summaries, putting emphasis on analysis but too little on synthesis. Typically, these reviews include an impressive number of studies without effectively synthesizing or integrating them into a cohesive narrative. This results in an overwhelming but ineffective (i.e. non-generative) synthesis of the literature. Author- and article-centric summaries tend to describe individual studies (who did what, which theory or method they used, and which topics they investigated) without integrating insights to revise or refine our theoretical understanding. As another example, even when papers produce frameworks, such frameworks are often criticized for their tendency to be topic-based rather than genuinely conceptual or theoretical (Wong, 2021). Such descriptive approaches have drawn criticism for lacking critical assessment and for failing to meaningfully synthesize knowledge (Webster and Watson, 2002). Topic-based collations often tend to group articles by broad subjects or topics, presenting one topic after another without deeper synthesis. Topics like “sustainability,” “procurement,” “logistics,” or “innovation” often serve as basic categorical labels rather than as lenses for theoretical development. The emphasis of such reviews often remains on identifying gaps in topic coverage, rather than on advancing theoretical perspectives.

Third, many reviews lack substantial implications and tend to be backward- rather than forward-looking with little to no implications for future research, practice and policy making. Yet, a review that makes a difference should have substantial implications, with our view of the world being different after the review research than it was before. Indeed, some journals require explicit implications for future research (e.g. Journal of Business Logistics, Journal of Management Studies). Even if authors state such future research implications, one specific, albeit frequently observed issue is a disconnect between review insights (“past/present”) and future research (“future”). Here, the suggested future research directions do not emerge from the review and are not linked to the status quo but rather come out of the blue.

The following practices could help mitigate the risks of being rejected for insufficient knowledge contribution: First, explicitly state the review’s new knowledge contribution and utility and deliver on it. Many review studies claim to contribute to theory, but how they do so remain vague or underexplained (Hoon and Baluch, 2020). Being explicit helps shift focus on producing an “intellectual product” (rather than a summary and mere listing of what the existing studies have done. In order to theorize effectively, Hoon and Baluch (2020) recommended that review authors should engage in dialectical interrogation—a reflective and critical process that makes explicit how the review engages with existing literature, whether by consolidating established knowledge or disrupting it to generate novel theoretical insights. Just to provide one example here, Hohn and Durach (2023) set out to refine an initial theoretical framework, which was derived from prior literature in the field of interest, addressing the “what,” “how,” and “why” of the phenomenon in question. Based on this starting point, the authors set out to challenge or refine scientific knowledge by critically assessing and integrating the primary literature into a coherent framework. The review proposed a new theory for the development of firms toward an integrative view of socially sustainable supply chain management. They depicted the proposed theoretical linkages, along with moderators and mediators, in figures.

While the intellectual product can take many different forms and vary widely across review purposes/types/methods as indicated above, authors can consider both more generic avenues as well as specific ways for contributions. Kunisch et al. (2023b) proposed three generic ways for advancing theoretical knowledge about grand challenges, namely, clarifying key concepts, advancing programmatic theory and overarching frameworks and revealing onto-epistemological assumptions, ideologies and values (see Table 1, p. 242). Post et al. (2020) “propose[d] a non-exhaustive set of avenues for developing theory with a review article: exposing emerging perspectives, analyzing assumptions, clarifying constructs, establishing boundary conditions, testing new theory, theorizing with systems theory, and theorizing with mechanisms” (p. 351). With respect to “problematizing,” Alvesson and Sandberg (2020) described different ways to “problematize” the existing knowledge base. Durand et al. (2017) focused on “integration” (integrative reviews) and identified five broader opportunities including, for example, “[r]econciling different theories and results that address the same phenomena but with different, sometimes conflicting, predictions and outcomes … [or] [a]pplying a single theoretical approach to a range of separate but related phenomena” (Table 1, p. 12).

Second, discuss implications for future research. A review that contributes to theory should also have no difficulty identifying a wealth of future research opportunities. Arguably, the most prominent form to do so is by offering “suggestions for future research” including exemplary research questions. Salmon et al. (2023) summarized their key insights and future research directions generated from their analysis in an extensive table (see Table 6, pp. 460–463).

When it comes to knowledge contributions, bibliometric reviews have especially raised controversies. While bibliometric reviews using bibliometric analyses are known for their systematicity, often producing visually engaging and well-structured charts and graphs, they are often descriptive in nature, primarily focusing on mapping connections between articles and identifying themes, which can sometimes limit their generativity (Russo and Wong, 2024). This focus on field-mapping rather than advancing theoretical knowledge may at first glance constrain their ability to meet the expectations of respectable academic journals. However, bibliometric reviews—or maybe better, bibliometric analyses—are not inherently limited or devoid of a theoretical contribution. Just like any other review method, reviews using bibliometric analyses also need to produce an “intellectual product” (i.e. demonstrate generativity) by moving beyond description. Marzi et al. (2025) proposed several strategies through which bibliometric analyses can transcend their descriptive nature and make significant contributions.

Recent bibliometric studies published in leading journals have mostly tried to attract the reader’s interest not through a direct advancement of theory, but through asking other questions of interest to the community. Nerur et al. (2016) mapped the evolution of the intellectual structure of the strategic management field, uncovering input–output dependency relationships with other disciplines. Among their findings was a noted decline in practitioner orientation, an insight that should spark interest as it serves as a critical signal to the scholarly community. Rabetino et al. (2020) examined the philosophical assumptions and prevalent conventions underpinning various research communities and theoretical approaches, offering foundational insights for problematizing existing literature and fostering theory development. So rather than advancing specific theory, they explicated the often implicit onto-epistemological foundations of extant theories and current knowledge and thereby nurturing further knowledge production. Kohtamäki et al. (2022) took stock of the Strategy-as-Practice domain and proposed a new research agenda, blending descriptive insights with forward-looking recommendations. Notably, these studies also succeeded because they engaged not only with the linkages between studies or keywords but also, at least partly, with the substantive content of the reviewed literature.

While the previous pitfalls mainly focus on designing and conducting rigorous and generative review research, reasons for rejecting a review paper can be directly related to the art of “paper crafting.” Academic writing is in itself an important skill (Huff, 1999) as it is “an act of communication, based on an established set of conventions, involving a plurality of actors (authors, editors, reviewers, and other scholars), and aimed at conveying a core message (contribution) to an audience of (management) scholars and practitioners” (Patriotta (2017), p. 4). Editors and reviewers, “[…], and ultimately the readers, want […] a well-thought-through and well-articulated” review (Michailova, 2023, p. 363). Yet, while many scholars have participated in some sort of academic writing courses during their PhD education, courses on review research and the peculiarities related to crafting a review paper for publication are still rare.

A poorly crafted review paper is ill-structured, lacks articulation and fails to persuade the reader. Bhardwaj (n.d.) argued that the art of persuasion, which plays a key role in science, rests on both an appeal to the rational (e.g. evidence) and on an appeal to the non-rational (e.g. language and values). Michailova (2023) elaborated “The importance of articulation is not to be underestimated; according to Short (2009, p. 1314), ‘a review can also be viewed as a work of art’” (p. 363).

While paper crafting concerns many different aspects, there are a few prominent manifestations that can be observed frequently. One set of problems relates to the structure of the paper and logical flow of the arguments. This includes unbalanced or missing chapters and a lack of well-organized and clear arguments. Many papers lack a proper method section or implications for future research. In addition, the sections are often not well-connected. In many review papers, the past (insights from the review) and the future (suggestions for future research) are disconnected. In other words, although there are suggestions for future research, they do not emerge from the review.

A second manifestation concerns the quantity and quality of graphical elements (tables and figures) and appendices. In terms of quantity, while some papers have little to no graphical elements to aid communication, many papers have too many tables and figures. Sometimes papers contain up to a dozen tables and figures, which relates to the “Presenter’s Paradox” (Connelly et al., 2023). Review researchers should focus on documenting the research process and summaries of key insights, but not all the details need to be included in the paper; instead, it is recommended to use appendices and supplementary material. In terms of quality, Short (2009) noted the following problem in relation to communication: “The tables span many pages, but provide little in terms of themes concerning specific trends concerning the review topic” (p. 1316).

Paper crafting is an art, and review articles come in all shapes and colors. There is neither a silver-bullet solution nor one formula on how to craft a review article. Of course, one needs to master the fundamentals of scientific writing and paper crafting as in all other forms of research (for advice, see Huff, 1999; Pollock, 2025). Yet, we see a few things that are specific to crafting a review article.

First, pay attention to telling a coherent story for a review article as a prospective “conversation changer” (Healey et al., 2023). This means “mov[ing] beyond reporting new phenomena to play[ing] an active role … [and] process of shaping other academics’ and practitioners’ understanding of the phenomena” (Fawcett et al., 2014). Indeed, good reviews have a common core in what they seek to accomplish and what various audiences expect (George and Cronin, 2024). In this process, researchers need to shift from sensemaking to sensegiving. Authors should “distinguish between doing a [literature review] (sensemaking) and writing up a [literature review] (sensegiving). Do both well, but submit the latter, not the former. And take care of the artistic elements of the review, demonstrating mastery” (Michailova, 2023, p. 363). Writing a review for publication is a story-telling process, with characters, a plotline and a clear beginning, middle and end. The written review simplifies a complex and lengthy process and presents a unique snapshot of the literature. This story needs to be convincing to editors, reviewers and readers with a good punch line or takeaway that rewards the reader after numerous pages of text, tables and figures!

Second, write clearly. As Parmigiani and King (2019) stated, “Clarity of communication is key” (p. 3085). This includes developing a suitable structure for the review. To this end, reviews are different from other forms of research, so the structure is also different. As a rule of thumb, a well-crafted review can follow this formula: introduction with research motivation and purpose (5%), background and conceptual foundation (10%), review method (15%), review insights (40%), discussion and implications (20%) and conclusion and future research (10%). Be transparent by making the data available as supplementary information.

Third, produce a limited number of high-quality tables and figures to facilitate clarity of communication. As Parmigiani and King (2019) noted, “tables and figures can be very helpful in conveying the central ideas and concepts” (p. 3085). Producing high-quality tables and figures, which complement the text but also work standalone, is often underestimated and requires a lot of effort. They may appear simple, but, as with models, beauty often lies in simplicity. Of course, the numbers can vary, but typically, it is a small number of tables and figures (plus appendices) that are needed to convey what the author intends. As a recommendation, one could limit tables and figures as one for the background, one for the method section, one to two for the review insight section and one for the future research section.

Finally, revise, revise and revise! Many submitted papers do read like a first draft that has never been read by peers or been subjected to language editing. Revising helps improve coherence and logical flow. This includes mastering the fundamentals of paper crafting (e.g. a well-thought-through and well-articulated review). To this end, Campbell and Aguilera (2022) suggested the following: “Each section needs to logically follow from the preceding section, each paragraph needs to perfectly follow the preceding paragraph, and each sentence needs to naturally follow the previous sentence. And anything that is not absolutely essential can—and should—go. [One technique] that might help you revise your work and test whether your main ideas clearly come through, stand out, and are supported by theory are, first, distilling the entire paper to bullet points in a PowerPoint presentation or “reverse outlining”—which entails identifying the topic sentences in each paragraph.” (p. 525).

We now turn to generative artificial intelligence (GenAI). With the rapid advancements in digital technologies and general-purpose technologies such as GenAI, there is a rapidly growing interest in the role of technology in various types of research (van Dis et al., 2023). Several writings have discussed the prospects and limitations of GenAI in research in the field of business and management (see, e.g. Bechky and Davis, 2025; Cornelissen et al., 2024; Dwivedi et al., 2023; Gatrell et al., 2024; Lindebaum and Fleming, 2024; Lorenz et al., n.d.). Against this backdrop, it comes as little surprise that scholars have also started to discuss the novel opportunities and distinct challenges for the production of review articles (see e.g. Block and Kuckertz, 2024; Chen and Hitt, 2021; Glickman and Zhang, 2024; Pan et al., 2023; Tingelhoff et al., 2025; Verma and Yuvaraj, 2023).

Advocates have emphasized the potential of AI and related technologies to enhance the review process. Krlev et al. (2025) argued that as AI advances, it can render certain review purposes obsolete and scholars must develop new, technologically enabled reviewing practices. Krlev et al. (2025) suggested that “technologically infused and collective ways of reviewing the literature could help make researchers’ engagement more reflexive, dynamic, and impactful” (p. 1). Similarly, Antons et al. (2023) introduced the concept of “computational literature reviews”—a method that augments rather than replaces human researchers by employing computational tools such as text mining, machine learning, and AI to automate key review tasks. Simonetti et al. (2025) offer a concrete and advanced toolkit to automate systematic literature reviews using NLP and Network Analysis, which complements and extends Krlev’s call for new “technologically infused” review practices and Antons’ computational review framework. These approaches make large-scale and real-time literature synthesis more feasible.

Sceptics have raised concerns about the limitations of these technologies and pitfalls in their application. Bhardwaj (n.d.) cautions that “in a marketplace of ideas where AI can generate theories [and reviews] cheaply, understanding how to do so is even more critical” (p. 1). One journal editor reflected similarly: “AI offers lots of possibilities, but it is also the reviews we often reject because they don’t go beyond what the AI tells them. This also raises the issue of transparency in the method.” Such critiques highlight the ongoing need for human judgment, critical reflection and methodological clarity when integrating AI into the review process.

Next, we anticipate manifestations related to reasons for rejection related to the use of AI and offer insights on mitigation strategies (summarized in Table 2).

While the use of technology may influence various aspects of generativity and systematicity—and consequently affect the reasons for rejection in diverse ways—we believe that these two quality dimensions will stand the test of time, even in the age of AI and algorithms. Drawing on and complementing recent contributions, we identified two generic pitfalls that warrant particular attention when using AI to produce publishable review articles. We refer to them as (1) misuse of technology (unacceptable use) and (2) overreliance on technology (acceptable but unproductive use).

The first pitfall, misuse of technology, concerns the inappropriate or unethical use of AI, which raises serious ethical concerns and may undermine core scholarly values such as reflexivity, agency and responsibility. Lindebaum and Fleming (2024) discussed the knowledge produced by ChatGPT and “what this means for our reflexivity as responsible management educators/researchers, and how an absence of reflexivity disqualifies us from shaping management knowledge in responsible ways” (p. 567). They “unpack its intrinsic epistemological limitations. Using high-probability choices that are derivative, ChatGPT has no stake in the knowledge it produces and is thus likely prone to offering irresponsible outputs. By contrast, genuine human thinking—embodied in a contingent socio-cultural setting—uses low-probability choices both “inside” and “outside” the box of training data, making it creative, contextual and committed. We conclude that the use of ChatGPT is wholly incompatible with scientific responsibility and responsible management” (p. 566). Similarly, Felin and Holweg (2024) “argue[d] that AI’s data-based prediction is different from human theory-based causal logic and reasoning” (p. 346) while Bechky and Davis (2025) “propose[d] a set of reforms to preserve the sacredness of craft and community at the core of our scholarly work” (p. 1). In addition, Tingelhoff et al. (2025) listed several areas of inappropriate or wrong use of technology in reviews.

The second pitfall, overreliance on technology, refers to the permissible yet problematic dependence on AI tools that may hinder rather than enhance scholarly contribution. Although the term “generative” might suggest parallels to recent advances in generative AI tools, such technologies offer only limited support (see, e.g. Block and Kuckertz, 2024; Tingelhoff et al., 2025). While they can streamline the summarization of existing research and highlight overarching topics, they cannot replicate the human capacity for deep interpretive reasoning, theoretical rigor and critical inquiry (Wong, 2021). Consequently, GenAI alone cannot illuminate what we truly do and do not understand about a given phenomenon, nor can it meaningfully shape the trajectory of our knowledge creation. Felin and Holweg (2024) argued that AI is more likely to produce backward-looking knowledge rather than forward-looking knowledge. Antons et al. (2023) also pointed toward the risks of researchers over-relying on algorithms, which can lead to misinterpretation of results and biased data selection. Additionally, poor scoping, lack of transparency and technical barriers can reduce the validity, replicability and ethical integrity of the review process – garbage in, garbage out. In other words, scholars must have a profound grasp of the reviewed literature and the skills to synthesize diverse insights, identify gaps and forge meaningful connections that advance the field. Otherwise, they might end up in the lower-right quadrant of Figure 1, where technology has helped them achieve systematicity, but the results contribute little to advancing knowledge.

GenAI is certainly a “moving target” and we need to stay on our toes. However, to ensure that AI tools enhance rather than undermine the integrity and value of review articles, we propose three guiding principles: (1) responsible use, (2) effective and efficient use and (3) transparent use and communication.

The responsible use of AI necessitates adherence to high ethical and scientific standards. AI should support rather than replace human scholarship, ensuring that integrity, reflexivity and accountability remain central in literature reviews. Tingelhoff et al. (2025) discussed the responsible use of GenAI in literature reviews by emphasizing that while GenAI can enhance efficiency, its integration must adhere to scientific quality standards, transparency and ethical principles. Their approach shifts the focus from what AI can do to what AI should be allowed to do in literature reviews. They outlined eight principles that should guide responsible AI use in literature reviews: beneficence, respect for persons, integrity, responsibility, rigor, impact, reproducibility and transparency.

Moreover, ethical considerations surrounding AI-generated content have been widely discussed. Dasborough (2023, 2024) and Tingelhoff et al. (2025) emphasized the need for explicit acknowledgment of AI’s role in academic work, cautioning against implicit reliance on AI-generated content. Scholars should ensure that intellectual ownership remains with human researchers. Moreover, some journals have begun to articulate editorial policies regarding the appropriate use of AI in research for publication (e.g. Cornelissen et al., 2024; Gatrell et al., 2024).

Effective and efficient use of AI can help to automate routine tasks and augment analytical capabilities, much like its application in empirical research. However, researchers must critically assess when AI tools add value and when human oversight is indispensable. Gatrell et al. (2024), in their editorial for the Journal of Management Studies, discussed how AI can assist in various phases of literature reviews, but they also caution against overreliance. Similarly, Tingelhoff et al. (2025) proposed a structured process for AI integration, outlining which steps in a review may benefit from AI automation (e.g. literature retrieval, summarization) and which require human judgment (e.g. theoretical synthesis and critical analysis)—particularly the latter part ensures the achievement of generativity, that is, contribution to theory in a literature review. In line with our thoughts, they highlighted that AI’s role should remain complementary, enhancing efficiency without compromising the intellectual depth of a review.

Transparency is fundamental in ensuring that AI-generated research remains trustworthy and reproducible. As AI becomes more prevalent ^[6], clear disclosure of AI involvement is crucial for maintaining scientific integrity. Some leading journals have already introduced formal AI disclosure policies. Gatrell et al. (2024) stated: “Our new guidelines require that authors make a clear and transparent statement about any use of AI in their manuscript, ideally in their methods section where it is clearly visible to reviewers and ultimately readers” (p. 746). This aligns with broader discussions on responsible AI use. Daigh et al. (2024) stressed the need for guidelines that ensure AI contributions are openly acknowledged, helping readers and reviewers assess the rigor and originality of a study. Transparent reporting should specify where and how AI was used, whether for language refinement or data extraction.

Review articles are crucial in the production of scientific knowledge. Despite a wealth of advice on how to produce rigorous and impactful review articles and what makes great review articles, many review papers submitted to journals are rejected. Because reasons for failure are not merely mirror-inverted success factors, our and the editors’ aim with this article was to lift the veil behind which rejections are often hidden and thereby help authors better navigate existing guidelines. Thus, in this article, we took a different approach and focused on why manuscripts get rejected.

We started by emphasizing that unsuccessful review papers fail to meet either or both generic quality criteria, systematicity and generativity. On this basis, we outlined common pitfalls and how these pitfalls typically manifest in manuscripts. Thereafter, by pointing to method papers and illustrative examples, we provided targeted mitigation strategies for each of the pitfalls. The suggested remedies for the pitfalls offer a pathway to achieve both systematicity and generativity. To summarize our arguments, in Table 3, we provide a practical checklist of questions that may help authors critically evaluate their review articles before journal submission. This checklist is intended to help authors identify potential vulnerabilities or areas for further improvement by reflecting on whether their reviews might be susceptible to one or more of the pitfalls.

We hope this work lifted the veil behind the review process and thereby offers useful guidance that will help researchers avoid the most common pitfalls and strengthen the quality of their review papers in terms of systematicity and generativity.