The future of knowledge management: an agenda for research and practice

ABSTRACT

This paper proposes a research agenda for knowledge management research and practice. To do this, it first reviews selected past knowledge management activity, especially that published in KMRP, including descriptions, predictions, initiatives and other research agendas. This merges into a consideration of the current states of knowledge management literature and knowledge management practice, and some issues that need to be taken into account for the future. These are used to synthesise an agenda whose key feature is research on practice, not just research and practice.

KEYWORDS:

• Artificial intelligence
• knowledge management journals
• knowledge management practice
• knowledge management research
• research agenda

1. Introduction

Knowledge Management Research & Practice (KMRP) completed its twentieth Volume at the end of 2022. This is an appropriate time for reflection, but the purpose of this paper is mainly to look at the future. We look backwards only in order to look forward: following the advice of many sages from Confucius to George Santayana, the past is reviewed to provide context, and especially to consider how accurate past predictions and forecasts about knowledge management (KM) turned out to be. However, Huber (2004) has pointed out how predictions and forecasts tend to be inappropriately anchored to the past. By reflecting, we hope to learn from what we and others have predicted, proposed, or tried to initiate, whether it proved to be successful or unsuccessful.

KM as an identified field of activity began in the mid-1980s, and was in full flow by the mid-1990s. The first issue of KMRP was published in May 2003, in print-only format. However, producing a new scholarly journal like KMRP did not happen overnight. The idea was first proposed in a conversation between one of the authors (JSE) and Meliha Handzic at the 2nd European Conference on Knowledge Management in Bled, Slovenia in 2001. Turning that into a physical journal, with the collaboration of the Operational Research Society, took 18 months. As the journal’s name implies, the focus of the journal has always been on the interplay between research and practice, as recently restated by the current editors (Lönnqvist et al., 2022).

KM is a broad area. This has been, and remains, both a blessing and a curse. There is scope for contributions from any area of expertise, and the potential for cross-fertilisation between the understandings of different disciplines. The literature reflects this, with inputs from sociology, economics, finance and accounting, computer science, library and information science, human resource management and operational research/management science, to name just a few. However, there is still no generally agreed definition of KM, nor any sign of one being achieved in the foreseeable future. This is not necessarily a problem: artificial intelligence faces the same issue, and does not seem to be being held back by it. On the other hand, it does make it harder to explain KM, its benefits and its challenges to others, whether they are people outside the field or those who might possibly be encouraged to join it.

Indeed, one point that is hard for those in the field to judge is what non-specialists actually know about KM. Far more people than 20 years ago will probably be aware of KM as a label, but what detail, if any, do they know about it? Lönnqvist (2017) articulated the resultant concern that “many practitioners seem to find it difficult to appreciate its added value for their managerial work” (p.184). He went on to describe the KM literature as dealing with “somewhat technical tasks related to knowledge” whereas managers, even in knowledge-intensive organisations, dealt more with “general management type challenges”. Bailey and Clarke (2001) had identified the “added value” issue, and felt that what they called the “ultimate breakthrough for KM” (p.67) would come when there was “a realisation that managing knowledge in the twenty-first century is what managing organisations and their change actually is”.

Even within the field, authors often show a curious reluctance to be precise about what they mean by KM. Müller-Merbach (2006) argued forcefully that KM authors should follow the advice of Eysenck (1979), who suggested that “Everybody who talks or writes in some depth about any concept ought to communicate [their] personal understanding of the concept, that is present an ‘operational definition’”. Müller-Merbach also warned that such an operational definition of knowledge or KM should not be regarded as apodeictic (certain beyond dispute), but rather was essential in order to make their “territory” and assumptions clear. It is difficult to disagree with this viewpoint. Nevertheless, some 15 years later, Jevnaker and Olaisen (2022), analysing papers from three years of the IFKAD (International Forum on Knowledge Asset Dynamics) series of conferences, still found that two-thirds of the papers did not define their terms fully at all. So when we discuss KM, are we even sure that we are talking about the same thing(s)?

For our purposes in this paper, we can only adopt the pragmatic position that knowledge management covers anything that the people doing it or writing about it say is KM. A key consequence of this is that there are also many different definitions and types of knowledge, as Müller-Merbach explained in several short articles in KMRP between 2006 and 2008.

The remainder of the paper is structured as follows. We look briefly at what past authors have said, under the headings of descriptions, predictions, initiatives and research agendas. We then mention some specific aspects or issues that seem to be important to KM going forward, and finally offer our own agenda for the future of KM.

2. Looking backwards to look forward

Contributions inevitably overlap between our four headings of descriptions, predictions, initiatives and research agendas, especially the first category and one of the others. We consider each paper under the heading(s) where it is most relevant to our discussion.

2.1. Past descriptions of KM

There have been numerous descriptions of the state of KM, and/or how KM got into that state. Initially, these tended to consist of “what I think about KM” or “what I and others have experienced of/as KM”. As the number of publications about KM increased, literature review papers followed, and then even papers that reviewed other literature review papers about KM. We select just two for specific mention here.

Baskerville and Dulipovici (2006) contributed what we regard as still the most useful account of the theoretical background to KM: as its strength is in its detail, we do not attempt to summarise it here. Rather, we recommend those interested to read it in full.

Schiuma et al. (2023) offer a comprehensive summary of the authorship, provenance, content and literature foundations of KMRP papers published from 2006 to 2019. Their cluster analysis of the content of the papers yielded four main groups. The first and largest cluster concerned knowledge sharing, including knowledge transfer and knowledge exchange, and also covered trust and organisational support. Cluster 2 papers centred on knowledge management itself, including tacit knowledge, codification and innovation. The third cluster focused on the various aspects of intellectual capital, including social capital, relational capital and knowledge assets. The smallest of the four clusters concentrated on social networking, absorptive capacity, networked innovation, supply chain and collaboration. One might think of this as multi-organisational KM. Schiuma et al. pointed out that the majority of the papers in this cluster were published after 2015, making this work a more recent development than the other clusters.

2.2. “Pre-KMRP” predictions about the future of KM

The person generally credited with popularising the term KM, Wiig (1997), writing about the past and future of KM, thought that over a timescale of a decade or two “we can expect to find that what we today think of as explicit KM practices and activities, will have been assimilated into the daily mainstream work – they will become automatic and ‘second nature’”. This was based on what he had observed in organisations that were early adopters of KM. Wiig specifically mentioned Dow Chemical as an example. Another of the pioneers of KM, Larry Prusak, writing in 2001, saw Wiig’s view as one of two possible outcomes for KM over the next five years. One possibility was that “Knowledge management may become such a natural part of how people organise work that it becomes invisible” (Prusak, 2001, p. 1006), following along the lines of the quality movement, which he thought had achieved that status. However, Prusak also identified another possibility, that knowledge management would go the way of re-engineering, whose practical legacy, he felt was “almost nil”, because it would be hijacked by “sales representatives and sloganeers”. Liebowitz (2001) made a similar point: “if the knowledge management marketplace creates hype, overexpectations, and vaporware, then this could ultimately kill the ‘goodness’ of a knowledge management strategy” (p.6) and suggested widely agreed standards were a way to avoid or prevent it.

This does not seem to have happened to KM, but nor, in most organisations, has it become a natural part of how people organise work. There has been much work on standards: for example, in 2018 the British Standards Institution published an ISO Standard implementation document dedicated to KM (ISO 30401:2018). However, KM does not seem to be as widely known or used as might be hoped.

Smith and Farquhar (2000) presented, from a perspective based on their experience working for Schlumberger, a road map of technology and practice for KM for the next 10 years. This was strongly based on the possibilities they thought artificial intelligence (AI) could offer, and very much stressed the positive side: Liebowitz (2001) offered rather more measured reasoning about then-future interactions between AI and KM.

Akeroyd (2001) made several reasonably accurate predictions about the future of academic libraries. For example, that “physical collections will mutate into multi-functional spaces”. He also predicted that “Knowledge management will become increasingly important” to libraries. While this has happened to some extent, it has not had a major impact.

2.3. Past research agendas

Grover and Davenport (2001) offered what they labelled “a pragmatic framework for KM research”. This consisted of very broad dimensions (e.g., individuals/groups/organisations) that are arguably more useful for categorising research than directing it.

Shin et al. (2001) identified three more specific pillars of a research agenda for KM: extracting context independent knowledge; knowledge location (i.e., finding knowledge); and rewarding knowledge sharing.

Edwards et al. (2003), setting the agenda for KMRP itself, believed there should be “a constructive tension in place, between difference and integration. KMRP, and indeed KM in general should embrace difference – differences in perspective, differences between disciplines, differences in levels of analysis, differences between the concerns of research and those of practice. These differences are to be embraced rather than eliminated; they serve to make the field more fruitful and relevant. However, they need to be embraced in a spirit of integration, of debate, of complementarity, of building bridges; and the most important of these bridges is the one between research and practice, whether research leads practice or follows it. All of this needs to be done with the necessary rigour to convince people, whether academic or practitioner, of the usefulness of KM and its study”.

Argote et al. (2003) took a very different approach, identifying what they called six emerging themes in KM research. These were: significance of social relations, “fit” of properties of contexts, organisational boundaries, nature of experience, environmental factors and embedding knowledge.

Many other authors have offered their view of a KM research agenda since this selection from either side of the year 2000. For example, Gunasekaran and Ngai (2007) proposed several research topics for KM in manufacturing; Venkitachalam and Busch (2012) identified no fewer than 15 research directions in tacit knowledge; and Levallet and Chan (2019) specifically addressed knowledge retention and knowledge loss.

Perhaps the most comprehensive attempt at a research agenda has been by Heisig et al. (2016), who surveyed 222 KM experts from around the world on future research needs for KM and business, and came up with eight very broad themes. These were business strategy, intellectual capital, decision-making, knowledge sharing, organisational learning, innovation performance, productivity and competitive advantage. Heisig et al. made the point that some of these themes had already been extensively researched, and that this shows KM is such a broad and complex field that their experts were not aware of all of the research that had been done. This may well be related to the issue of KM’s multi-disciplinarity. It could be argued that KM scholars themselves need better knowledge management!

2.4. Past initiatives

Many authors have suggested frameworks for KM practice, such as those by Gunasekaran and Ngai (2007); Shin et al. (2001); and Smith and Farquhar (2000) already mentioned. It is after all one of the things that academic authors do in the hope of making a name for themselves. Arguably more usefully, Liebowitz (2001) offered five specific practical suggestions for the early days of KM in an organisation: run a series of KM forums; conduct a knowledge audit of a targeted area; attend KM seminars and conferences; bring in KM advisors to shape a KM strategy; develop a repository for best practices/lessons learned/“yellow pages”. These would still be well worth following for any organisation that has never engaged with KM.

Others have proposed more specific directions in which KM should assist and/or develop. A significant one of these is Hasan (2008), who discussed how KM might contribute to what she called “sensible organisation”. She described this as “the art of making common sense decisions and judgements”. Her observation (p.27) that “Sensible enterprises will become agile, flexible and adaptable by incorporating more creativity and diversity into their structures, processes and human resources” has surely become even more relevant over the past 15 years.

Massaro et al. (2016) suggest that “editors and business schools should provide room for discussions of research findings between scholars and stakeholders”. They observe that conferences are good for this, but how many stakeholders actually attend KM conferences? In our experience, practitioner attendees at KM conferences are more usually people whose responsibilities are specifically for KM.

Dumay (2022) and Jevnaker and Olaisen (2022) both call for more problem-driven KM research. Dumay comes from what might best be termed a critical paradigm, and also compares KM unfavourably with intellectual capital (IC) in respect of problem-driven research. The content of the TAKE (Theory and Applications in the Knowledge Economy) conferences supports that this is a reasonable claim – there is a strongly problem-driven element there, and the conference covers IC more than KM. These arguments suggest that much of the research work currently reported in KMRP may be quite far (too theoretical, too generic) from the actual work of KM practitioners.

3. Aspects and issues

3.1. Is KM dying?

Authors remained obsessed for a remarkably long time with whether or not KM was a fad that would soon disappear, until analyses such as those of Wallace et al. (2011) and Serenko and Bontis (2013) laid that to rest. However, Serenko and Bontis, analysing journals in KM and IC, did find (p.321) that “One of [many KM/IC researchers’] key concerns is the lack of clear identity and external recognition of KM/IC as a distinct scientific field”. This concern continues today, and has led to some conference discussions of whether or not KM is dying. There seem to be two possible causes for this. First, there have not been many breakthrough developments in KM research in recent years. Much of the research is about making minor additions to what is already known (e.g., testing connections between some knowledge-related concepts using statistical data). Many of the articles are still relying on well-known models such as SECI (Nonaka & Takeuchi, 1995) or the Knowledge-Based View (Grant, 1996) as their starting point. At the same time, some related disciplines such as analytics, big data, and artificial intelligence are developing very quickly. Second, the extent to which KM is actually practiced in organisations seems unclear; is KM a real organisational activity or mainly an academic exercise? For example, in Finland it is hard to find discussions about KM in practice. There is a lot of discussion about analytics and how to effectively use the information available to manage operations, but not about how to manage knowledge.

A specific issue relevant to this heading, as well as more broadly to our overall theme, is that there are very few published studies looking at the long-term effectiveness and/or impact of KM. Jevnaker and Olaisen (2022) found no longitudinal studies at all in 3 years of IFKAD conference papers.

Another issue to consider is that knowledge-related research themes have become popular within many established research fields. General management and business scholars as well as administrative and political scientists, to mention just a few examples, are exploring various research topics related to the potential of data, information and knowledge in their respective domains. For example, Saarijärvi et al. (2014) have explored the potential of customer data in developing new service-based business models. The KM research field originally started to evolve as a result of the growing significance of knowledge, the related management challenges and the lack of dedicated research focus on the topic. However, now it seems that KM research as a theme or perspective has become popular within many disciplines. As this is the case, we may question whether a specialised research field of KM is still needed. Time will tell, but it is our current belief that there definitely still is a need for a dedicated KM research field. KM research continues to evolve; as mentioned earlier, it is a broad and complex topic, and thus it is vital to have a dedicated research field pursuing and collecting the latest research findings for all those interested.

3.2. Repeating the same mistakes

Fahey and Prusak (1998) outlined the eleven deadliest sins of KM:

1. Not developing a working definition of knowledge

2. Emphasising knowledge stock to the detriment of knowledge flow

3. Viewing knowledge as existing predominantly outside the heads of individuals

4. Not understanding that a fundamental intermediate purpose of managing knowledge is to create shared context

5. Paying little heed to the role and importance of tacit knowledge

6. Disentangling knowledge from its uses

7. Downplaying thinking and reasoning

8. Focussing on the past and the present and not the future

9. Failing to recognise the importance of experimentation

10. Substituting technical contact for human interface

11. Seeking to develop direct measures of knowledge

Dervin (1998) observed that these formed a list she had “heard frequently over the past 25 years in other fields” (p.38). Experimentation (#9), for example, is something still frowned upon by many organisations in all areas, not just KM. #8 is a general human failing, as pointed out in our opening paragraph. Progress has been made on some of the eleven issues in KM over the intervening 25 years, especially knowledge flow, as witness the amount of research on knowledge sharing (#2), tacit knowledge (#3/#5/#10) and knowledge application (#6/#7). We have already mentioned that lack of common definitions remains an issue (#1). A fair summing-up of #4 might be that its importance is understood by those working in KM, but actually achieving and maintaining the shared context in an organisation often turns out to be difficult. #11 has seen many developments in the IC field, but arguably the significance of knowledge is not what it is, but what you do with it.

The eleven sins can still serve as a useful “memory jogger”.

3.3. KM and other new(ish) technologies

A natural direction for research is to study the relationship between KM and any new technological developments.

The technology capturing the public’s imagination most at present is AI, much as this “new” technology is considerably older than the term KM. There has already been significant research on AI and KM: recently, Leoni et al. (2022) found that the effect of AI on manufacturing firm performance is fully mediated by knowledge management processes. In other words, AI without effective KM is a waste of effort and money. This is a really important result that deserves to be more widely known, and replicated in other contexts. Interestingly, this was not published in a KM journal. There remains scope for much more work on KM and AI, for example the uses of generative AI systems based on large language models (Dwivedi et al., 2023).

Analytics and big data is another area where lessons already learned from KM should be highly relevant. There have been a few examples already (e.g., Chinnaswamy et al., 2019; Edwards & Rodriguez, 2016; He et al., 2019; O’Connor & Kelly, 2017). It is noteworthy that these have all appeared in either the KM or general management literatures. Most existing research in the big data literature does not seem to have been informed by KM concepts at all.

Other technologies where research into the interaction with KM might be fruitful include cloud computing, virtual/augmented reality and the metaverse.

3.4. Problem-driven research

There is an issue about publishing problem-driven research in journals. A common form of problem-driven research is action research. One author (JSE) has been involved in supervising two successful doctoral theses based on action research in the past decade, one of them a truly excellent piece of work. However, writing them up comprehensibly within the length typically allowed for a journal paper proved to be impossible, even with some journals allowing around 50% more words than KMRP’s recently increased limit of 8000. There are good KM book series available. These include the Springer series on Knowledge Management and Organizational Learning, edited by Ettore Bolisani and Meliha Handzic, and the Emerald series on Working Methods for Knowledge Management, edited by Denise Bedford and Alexeis Garcia-Perez. However, these do not lend themselves well to reporting single-organisation case studies either.

Despite the challenges in making managerially relevant, problem-based research publishable, we would like to see more of that in the future. On the other hand, the majority of submissions to KMRP nowadays are questionnaire-based statistical studies. While many of these are of high quality and provide novel findings, some may be considered mechanistic and repetitive with only marginal additions to existing knowledge. For example, we already know quite a lot about various organisational conditions and factors supporting or hindering knowledge sharing. We believe that doing more problem-driven research might result in novel findings that would also be relevant and interesting for practitioners and perhaps even the general public.

3.5. Scope

KM can operate on several levels, including at least those shown in Figure 1: personal, group, profession, organisation, network of organisations (a supply chain or other form of collaboration), industry, nation, society. The placing of the levels is indicative, and is not clear-cut. For example, some professions sit clearly within a single industry, while others do not. Linkages have been omitted deliberately, to avoid blinkered thinking. Most published KM research focuses on the organisation: it could be argued that this focus has been taken too far. In particular, the levels are not separate from each other, as the spiral part of the SECI model emphasised many years ago (Nonaka & Takeuchi, 1995). One of the main benefits of the KM movement has been the realisation that within organisations, trying to do KM at the group level can be more of a hindrance than a help – the classic criticism relating to organisational silos. Dumay (2022) draws attention to the wider societal implications of what an organisation does: Bailey and Clarke (2001) assert that all KM, including organisational KM, stems from the personal relevance of KM, and thus needs to be grounded in personal KM.

Figure 1. Levels of knowledge management.

Serenko (2021) identified (p.1889) that “There is a need for knowledge brokers that may deliver the KM academic body of knowledge to practitioners”. He thought that the gap between academics and practitioners had widened since a similar publication in 2013. This chimes with the work of Collins (2004) on what he termed interactional expertise, in the context of the public understanding of science.

3.6. Fake knowledge

There are two related issues here, stemming from misinformation and disinformation, the latter being deliberate. Knowledge issues arising from misinformation cover three overlapping aspects:

1. Out-of-date knowledge. Knowledge that has not been questioned recently and as a result is ineffective, because something crucial in the environment in which the knowledge is used has changed. This could be happening at any level from the individual to the whole organisation. It could be in regular use (“this is the way I/we do it”) and so not questioned, or it could only be used occasionally, and so rarely scrutinised.

2. Knowledge drawn from misleading data. This is a particular problem with the growth of machine learning, but can apply in any circumstances, as the old IT adage “garbage in, garbage out” testifies, as does the more recent inclusion of veracity as one of the “5 Vs” of big data. Indeed, it can occur at the opposite end of the scale – small data – as well, in the form of inappropriate reasoning from a sample of one or two.

3. Weak knowledge. Knowledge acquired from or used by somebody who is not really an expert in the relevant domain, but more like a journeyman/woman. The silo effect may come in here, or perhaps as a separate point that would be headed partial knowledge.

Turning to disinformation, and the “Trumpism” kind of fake knowledge, this lifts the discussion about KM to a political and societal level. Nowadays, many people seem to reject institutional knowledge (e.g., that provided by healthcare professionals related to vaccines) and become “Facebook or YouTube experts” by educating themselves with the message that seems most appealing to them. This comes back to the old “knowledge is power” slogan, and to the use of propaganda (fake stories) to change how people think and to advance one’s political goals. These issues are most apparent at the personal, national and societal levels of Figure 1, though they do also appear at the industry level, as with the past actions of the oil companies with respect to climate change.

There is also a small and very specific problem of deliberately fake knowledge in organisations – but that is one more within the orbit of ethics or criminal behaviour than KM.

4. Agenda

We synthesise the preceding sections to come up with what we must stress are our personal suggestions for a research agenda for KM.

4.1. For those doing research

More research on what KM achieves – its impact on practice (UK academics and some others will already be feeling this impulse).

More research that is not biased towards “the organisation”.

Progress towards a set of common processes, even if there will be no agreed definition.

Don’t let the AI opportunity slip by: AI desperately needs KM, even if its practitioners do not realise it themselves.

More research on what KM actually is in organisations (or elsewhere in society). Who is doing KM in practice and what is it that they are doing?

More papers linking KM to some new phenomena. Nowadays, there are more papers about KM and sustainability. Perhaps there could be papers connecting KM and societal security or resilience.

Our aim of proposing an agenda for research and practice should perhaps better read an agenda for research on practice.

4.2. For journal editors and reviewers

Insist that authors make it clear what they mean by KM, and perhaps by knowledge as well.

The “so what” test (particularly for KMRP) – what will we do differently in future now that we have the result of this research?

Be open-minded for new ideas and approaches even if the technical or methodological characteristics of the studies are different from what we are used to.

4.3. For those in and around KM

Talk it up, don’t talk it down. Diversity of perspectives is a strength, not a weakness. The wider literature shows the importance of KM, but (except during the 1990s) the key work has not always received the attention it deserves.

One of the points made by Massaro et al. (2016) was that “more interesting research that questions established conclusions is required”. Is what we have said interesting enough, we wonder?

Disclosure statement

No potential conflict of interest was reported by the authors.