Knowledge Management and Recordkeeping
Barbara Reed

Paper to AIC Document Management Conference,
Sydney, May 1998

Beware - there are a lot of snake oil salesmen out there. And perhaps I am one.

What I mean by that somewhat bizarre opening is that knowledge management is a concept that many many players in the information game can lay valid claims to be part of. However, the key word there is part of. Knowledge management is a large and complex conceptual framework which seeks to encompass previously existing technologies and insights into human behaviour to provide organisations with a way of leveraging and valuing information resources which exist within the systems and people. Knowledge management is not just information management repackaged. It has an evolutionary relationship to information management and because it assumes the use of technology, it can easily be packaged and presented with both the information management and the technology spins on it to the fore.

One of the key concepts in knowledge management is the distinction between the types of information used in organisations. This distinction between explicit and implicit information was outlined by Nonaka and Takeuchi in their book Knowledge-Creating Company (1995). Explicit information is that which is the domain of information management - information which is available for dissemination and sharing between individuals in a relatively static form - captured into systems in various formats. Implicit knowledge is that information which is 'highly personal and hard to formalise. Subjective insights, intuitions and hunches fall into this category of knowledge.' This is echoed in discussions of knowledge management, most approachably by statements such as Tom Davenport's: The best information environments will take advantage of the ability of IT to overcome geography but will also acknowledge that the highest bandwidth network of all is found between the water fountain and the coffee machine.'

Seeking to leverage this combination of information - explicit and implicit - to turn information resources into information assets and to value it as intellectual capital in organisations is the underlying goal of knowledge management. It links to a plethora of extremely active research and applied research communities:

• the CSCW (computer supported collaborative work) information systems community,
• information retrieval specialisits both technology and strategy oriented in the librarians and the computer scientists
• the management strategists theorising about the learning organisation and organisational learning
• business process reengineers, and
• the more focussed information technology communities working on specific applications such as intranets, web-casting, video-conferencing and the like.

To disentangle one stream and treat it independently of the whole is to distort the notion of knowledge management. And that is exactly what is happening. So, be wary of the claims of individual pieces of the information technology puzzle to be selling a knowledge management solution. The solutions are not purely technological, they will not come out of a box, but are highly contextualised to individual organisations seeking to exploit technologies and the capacities of people to share knowledge. This in turn links to the growing perception that the future viability of organisations is bound up with their ability to adapt, quickly and flexibly, to rapidly changing and increasibly unknoweable business environments.

Given that I prefaced my remarks by the warning that I may be one of the snake oil salespeople, what am I going to offer you? My aim is to offer some comments on the reconceptualisation of old forms which is taking place within the discipline which I represent - that of the recordkeeper. This reconceptualisation was perhaps well overdue, with many falling into the common trap of elevating particular practices into the only ways of seeing the world and the way we do things. The nature of the resulting products, still very much in formation, are increasingly being aligned to and absorbed into some of the very best applications coming out of the research thinking in the information field.

One of the planks of the rethinking of the nature and the role of records is to situate records as documents which provide evidence of transactions. This doesn't necessarily mean evidence acceptable in a court of law - for some organisations, for some types of business this may be necessary and more complex systems of managing records will be required. For others, however, the business risks will define that a lower level of evidential protection is required. What transactions require protection with a record -something which provides a trace of the action - is highly variable. There are some very clear cut cases - ministerial correspondence will always need to keep a record (see the Travel Rorts case); highly accountable transactions such as pharmaceutical trials will always need a record. The circumstances of accountability exist in both private organisations and government bodies. Some of the records required will be required only for days, some for even shorter periods and some for much longer. The business you are in dictates these things, as does the nature of the varying accountabilities which organisations face. Decisions on whether to create records are also bound up with decisions on business risk - to what extent is your organisation prepared to take the risk of not creating or keeping accountable transactions? Again, the answer to this question is only to be determined in the particular business context you find yourself in.

Records exist in many many media. What is consistent is the form and the role that the records play. Records can be paper documents or digital document like objects, photographs, video, multi-media objects and books. What distinguishes records from other types of information is the role they play in doing something - in transacting something. Protecting that transactional link is what makes records able to act as evidence. The media and format of the record is a secondary consideration to this understanding of the form and the socially constructed concept of the role that records play.

In moving to the digital environment, we have fallen victim to the temptation to treat all information as the same. The temptation has been there since the first days of information management and exists firmly in the minds of many who implement technology and whose expertise is the mechanics of computers. To them, if information can be treated in the same way, it should be and discrimination between information types is mere niceties which will not and should not survive in the new glowing digital age, where the way the computer supports information drives all. This is a dangerous view and one which throws out centuries of social and cultural development which is inherent in the forms of communication and their quite distinct roles. Thus when documents need to act as evidence of transactions, if there has been little attention to the specialised requirements of that form of recorded information, organisations will find that they cannot rely on the information as evidence. It has not been kept as an entity which maintains the links with transactions in which it took part.

Sometimes, by serendipity, the technology itself, deep in the mechanics of its operations, provides us with means of painfully recreating a record - that particular data or documents were used in particular transactions. The logging mechanisms, the back up tapes, the slips whereby the delete key only deletes pointers to the information not the information itself, have all been used in legal cases to painstakingly recreate the circumstances of use of data in particular contexts. However, reliance on the inadvertant rather than the designed uses of the technology, hardly seems adequate protection of a major corporate asset.

Records - data and documents used to transact business - are the corporate memory. Being able to define and protect this corporate memory is a major portion of managing the explicit knowledge base of an organisation. Without corporate memory knowledge management will not work - there will be no base resource on which to build the implicit knowledge which is the more people oriented side of knowledge management.

Evidence of transactions is an age old purpose of records. Embedded in this notion is the implicit assumption that records are authentic and reliable. The notions of authenticity and reliability are two things which we have needed to pull out of the old conceptions of records. We need to be able to rely on records. How authoritative is this piece of information: was it used in business, who produced it? Is the creator of this record a person with appropriate authorisation for transacting the business? Can I be sure that this is reliable?

These concepts take us into the realm of contextualising information. Increasingly this is a concern to organisations as they seek to re-use and exploit their information assets. To quote Hotwired: 'It's the context, stupid'. Contextualised information is the key to ensuring the exploitability of information resources. Information without context is the stuff which sends us into overload, into information glut, unable to discriminate between the useful and the useless. Knowing the context of the transaction and the validity of the source and resource is crucial to maintaining reliability and authenticity of records, their future exploitability and thus in enabling knowledge management to work.

Viewing records as the passive containers of past transactions is a limiting view. While they do provide that role and it will continue to be a vital role, a far more interesting way of thinking of records is as the active containers of business transactions. They can be seen as the stuff that makes business work. They are the embodyment of the business practice. Re-thought of as the agent of business, the medium through which business gets done, takes the role of records into a very different dimension. The almost instantaneous digital communications by which we transact business in this dititized, global network carry with them the authorisations and capacity to be the business. Thinking of records as the conduits of business, the transactors, moves the thinking away from records as passive residue of business transactions, into the realm of active agents doing the business. Ensuring that the business is transacted through agents invested with authority, reliability and contextualityand that those relationships are sustainable and proveable is the challenge of electronic commerce, in which records play a vital role.

To enable this world we need to see documents of business transactions as capable of carrying out multiple roles. Appropriately and inextricably embedded, encapsulated or linked to contextual meaning records can be they decontextualised - taken out of the context of their creation and threaded, linked, woven and reused for a muliplicity of future uses. They can be re-authored - the meaning of records is capable of interpretation in an infinite number of ways. The value that each individual will put on the content of records is as variable as the number of individuals who approach this source. The content of records can be exploited for other, future purposes which are unknown and unknoweable to the creators of the record. But it is not only the content of the record which is capable of interpretation: because of the contextual nature of records huge unexploited layers of meaning is conveyed through the contextual layers in which records are embedded. Some of the most interesting developments in web search engines research is the exploration of the capacity to exploit contextual information rather than content.

Records with their contextual information are also carriers of meaning. Meaning is the interpretation of the content and context in various future contexts. To ensure that records remain meaningful, we need to embed them with information which will carry that meaning across time and space. We need to envisage the depth and complexity of how meaning is created and can be exploited. It can be as simple as the keys required to translate the words - what does 'information superhighway' mean? How will outsourcing be interpreted in the future? What does competitive tendering mean? Similarly, in order to trace the authenticity and reliability of the record we need to know more than that Joe Bloggs created it. We need to know the context of who Joe Bloggs is and what he was doing and with what authority. This notion of compatability of roles and responsibilities and authority is more challanging in the fast-paced change of organisational structures. Yet, in our technological implementations there are already elements which we need to harness to capture this information. The network log ons, the access priviliges assigned to users, the networks of delegations and authorities in our software already provide some of the information that we need to capture this information. But we need it to be inextricably linked to the content of the transaction and we need it to be meaningful over time, so that when Joe Bloggs ceases to be the cleaner and becomes the chief knowledge officer, we can 'place' his communications in their appropriate framework of reference.

Implementation

If this is the environment of the rethinking of the record, its role and uses, into the new digital frameworks and organisational realities, what is the reality of implementation in organisations?

As with knowledge management, the concepts are beginning to be articulated differently, but the technology and most organisational thinking about documents as evidence, is still hampered by a paper mindset. In the same way that the first cars were seen as horseless carraiges, so most of our implementations of electronic records systems are carrying with them baggage from the paper world. Rather than being seen as one particular information architecture, paper records systems are dictating the ways that digital records systems should work. Some evolution in thinking is a necessary thing and a healthy orientation mechanism into a digital world, but the insistance that old ways of doing things should be replicated into the digital world, is condemning records systems and their proponents to obsolescance along with the horseless carraige. 

Once we understand the role that records play and the necessary elements and frameworks which need to be protected to ensure their action as evidence of transactions, we can envisage many and different ways of implementing records into organisational systems. Various options are being discussed. One, the business acceptable communications model propounded by David Bearman, proposes to capture transactions as they take part in business by crossing communications boundaries between workgroups, internal groupings and external agents. This model envisages implementation which inks in to the communications models of electronic commerce and utilises the network gateways and switches in the organisation's information architecture. Another model, put forward by John MacDonald of Canada, proposes transactions falling out of all of the diverse systems of an organisation, into an organisational wide records system which exists as a separate software layer. Records created in the process of being accepted into these systems will be linked via a variaty of implementation options with their descriptive metadata. The recordkeeping software exists as a platform under all applications catching transactions automatically.

The project described by Dagmar Parer yesterday, the Ausgils project is one such project which is identifying necessary metadata elements of documents and digital document like objects for discovery purposes. A highly collaborative project, involving many of the leading thinkers and practitioners in the Australian records community is about to embark upon specifying metadata for ensuring recordness for business, social and cultural purposes over time. The metadata elements required for transacting business and for ensuring evidence are not yet appropriately specified although some preliminary work has been undertaken in this field.

How does this relate to the existing picture of software being used in organisations? It’s a bit like the picture in the knowledge management arena. Many of the pieces of software have addressed part of the problem. Intranets are terrific for dissemination of information - but they're not repositories of time bound and context bound information. Collaborative software such as groupware and document management packages are terrific in enabling shared authorship, collaborative work on document creation and the pooling of knowledge in documentary form between defined colleagues. As these products mature, more and more of the requirements of ensuring that the product of the collaboration is more robust and more situated in its context are being added.

Document management packages are in transition. Highly powerful tools