Enterprise Knowledge Systems
The question is no longer "why do we need an intranet" but "how do we make our
intranet more effective?" Campbell McCracken talks through the latest issues surrounding enterprise-wide
knowledge sharing
The death knell
The death knell of the first generation intranet has already sounded. Those static
web pages that disseminate corporate information (information that is, more often than not, out
of date) are on the way out. Nobody uses the first generation intranet anymore (save the odd
employee or two who wants to read the company's latest mission statement or the canteen menu for
last month). No. The first generation intranet is soon to be consigned to the wastebin and it's not
hard to reason why, according to Intranet Focus's managing director Martin White. "It's because
there is a huge mismatch between the level of content that's on there and the level of content
that people need. Companies don't really know what their staff want on their desktops and
haven't thought about the problems of getting the material on to the intranet." Generally
companies haven't thought about the decisions that are being made by staff in the course of the
day and how an intranet can help them. The second generation intranets aim to do much better.
Second generation intranets
Although there is no clear rule that defines what a second generation intranet is,
there are a few key differentiators. One of these is that the intranet is now more integrated; it
can link to a number of different sources and the content is more aligned to the business
requirements.
Another difference is that the focus is more on knowledge than simply on data, giving rise to
the name 'Enterprise Knowledge System'. Rather than concentrating on the giving people access
to data, they are more concerned with data exchange. They have been simplified to the point
that non-technical users can publish information to the intranet from their desktops, rather
than having to get the IT department to do it - making it much easier for individuals or
departments to collaborate on projects. Essentially a second generation intranet enables everyone
to be a knowledge worker because they have access to real-time information and access to all
published documents and data within a corporation.
Search Tools
So while scientists, engineers and office workers create their own home pages
and share details of their projects with the rest of the company, they also want timely access
to accurate information to carry out their jobs effectively. This information can come from a
variety of sources. It could be corporate information from the intranet, such as data sheets,
standards, procedures from quality or HR departments, sales and marketing reports, support
centre reports, ... the list goes on. It could also be information from external sites,
such as news, competitive information, national and international standards, patents and
so on.
The two main search methods that are used today for information retrieval are the full text search
(for example, using keywords with a standard search engine) and hypertext systems (for
example browsing). With standard search engines you can end up receiving a lot of irrelevant
documents because of the inability of the engines to understand the context in which words
are being used. Using hypertext systems is not an exact science, either. It is very easy to
miss important sources because the hypertext links may not convey the full content of the
information in the document.
These problems have been recognised by several companies who have since developed products to
try to make searching more efficient. Verity's K2 integrated suite analyses documents by their
themes and contextual relationships rather than simply looking for keywords. It expands search
queries by using word stemming, linguistic analysis and a thesaurus. It is also able to perform
searches in parallel across multiple servers.
Instead of using keyword searches, Excalibur's RetrievalWare tool set uses natural language
processing. It also uses semantic networking, backed by a powerful dictionary set, to extract
the concepts and contexts from documents. The dictionary can be augmented by the integration of
additional reference works for specialist areas such as legal, medical, financial and engineering.
It also uses fuzzy search techniques that are sufficiently advanced to be able to cope with
damaged, misspelled or poorly scanned data.
As for Autonomy, it doesn't believe in searches. The software in its portal product profiles
users based on the ideas in the documents that they read or write and brings information to
them that contains similar ideas.
Hummingbird's answer to the Hypertext problem is to organise the information correctly by
getting the taxonomy right. "Users want you to structure an environment where they can follow a
logical progression through to some information," says Hummingbird's enterprise solutions
strategist Gareth Metyard. "When you go to a website looking for specific information you'll
go to the website where you believe the information is. You'll try to use the navigation bars
to find the information you're after. If you can't find that information, only then do you
look to see if they have a search button."
Hummingbird uses the Fulcrum Knowledge Server heuristic and linguistic tools behind the
scenes to train taxonomies to understand the way you want to classify your information.
Metyard suggests that in the future you won't have to think about keywords, or natural
language expressions. "This gives you the ability to browse for information, which is the
way that people normally work."
Knowledge Sharing
However, not all the information that your staff needs can be found through
conventional searching. Some of it could be locked up in other employees' heads. "Tacit knowledge
is the single biggest chunk of knowledge in an organisation," says Ernst Kallus, vice president
of global sales and marketing at Orbital. "It typically forms 40% to 50% of an enterprise's
intellectual capital." But unless this information is made available to others, it cannot be used.
Orbital's flagship product, Organik, captures tacit knowledge and organises it into virtual
communities that can be accessed either over a corporate intranet or portal. Once users have
joined a community and registered their interest preferences, they will be kept informed of
new information as it gets added. If they have a question, they can go the most appropriate
community and ask it in natural language. If the question (or a similar question) has already
been asked, the community will return the most appropriate answers.
If the community can't find an answer or if the user indicates that the answer does not meet their
requirements, e-mails will be sent to the community expert(s) and other interested members,
giving them the opportunity to supply a better answer. Any new answers are stored in a database
for future use and will also be copied to all users expressing an interest in that community.
"We're working with several corporations who are putting their research and development
departments on to Organik," says Orbital's CTO Calum Smeaton. "They want to profile their
expertise across all aspects of their organisations. They're setting up skills matrices that
allow their staff to search for a particular competence or skill and then ask questions. The
goal for them is to provide a directory of 'who knows what' within the research and
development departments."
Portals and the future
The time a user spends in an intranet is probably quite slim compared to the
amount of time he or she spends elsewhere. Where the time is spent is in the applications,
acting on the intranet information. But applications in isolation are not good enough.
"A lot of solutions have been created over the years such as CRM, ERP, document management,
knowledge management, business intelligence, reporting tools, etc," says Hummingbird's Metyard.
"But these have all been sticking plasters designed to cover the immediate pieces of pain that
the companies have. None of these tools has considered the existence of the others." This is
where the portal comes in.
The Enterprise Information Portal is an environment that enables a user to bring together all
the relevant information and technologies in the form of an application that they can use to
fulfil their daily tasks. They can use it to access structured data sets from ERP, MRP or CRM
systems, or unstructured information from e-mails, public folders, etc.
And once they have found the information and made decisions based on that information, they can
then act upon it through the appropriate applications. "There are many people who will say that
search and retrieval is key, but if you can't act upon the information you have, and you
have to go elsewhere, you're not really getting the full portal experience," says Metyard.
Application collaboration
In the evolution of portals, the first generation is the basic user interface
giving a web front-end to sources of information, and the second generation brought more accesses
to both structured and unstructured data. Hummingbird UK's marketing director Bharat Mistry says,
"The third generation is, according to the Gartner Group, application collaboration, workflow
and the ability to act on the information."
There are two facets to application collaboration. The first is the ability to integrate
your application into a portal. The second is that, once integrated, the applications expose
their functionality. This means, for example, that if you capture data in one application,
instead of cutting and pasting to a file and then saving the file on to the server, you could
move the data straight into a document management application that has exposed its
functionality.
Although the portal is more than an intranet, it doesn't replace it. The intranet has an awful
lot of content to offer the portal and the information that is held there should be an
integral part of the portal. "It is very valuable content," says Metyard. "We're trying to
make people understand that the information asset they have, the intellectual property
they've created in the intranet itself should be a part of the portal. They should rejuvenate
that information and bring it in with everything else they're using."
Intranet Focus' Martin White agrees. "Portal software is not designed to add content to an
organisation. It's designed to enable you to look at a lot of different sources on a desktop.
If you haven't got a robust intranet, databases, e-mail platforms and so on, all a portal will do
is hold up a big magnifying glass and say 'Hey! None of this fits together!'"
Scientific Generics is a technical and business innovation organisation
based in Cambridge, UK. Employing about 250 staff, the company is involved in research and
development in emerging science and technology areas across an immensely broad subject coverage,
including engineering, telecommunications, life sciences and biotechnology, applied sciences,
business management and IT software. Scientific Generics offers customised technology solutions
and strategic consultancy to a wide range of organisations internationally and has built up an
extensive portfolio of intellectual property.
Scientific Generics has a vast information requirement which is fulfilled virtually
single-handedly by Sarah Hinton, the company's information manager. Many of the staff come
straight from university and often come equipped with certain expectations about information
searching.
Consequently, rather than requiring searches to be carried out on their behalf, users doing
their own searching now require expertise and advice about information, as the question
"Can you find?" becomes "Where can I find?". As Sarah says, "Increasingly, people like to search
for themselves and will do so, whether you give them the tools or not - just not quite as well,
perhaps. As such, the information manager's function is becoming more of a training and guidance
role."
Hinton's aim is to provide an electronic library on every desktop and, to help her harness
Scientific Generic's intranet to give users a powerful information research tool, she turned
to Dialog. Dialog Intranet Toolkit's customisation capabilities, together with the powerful
Dialog search engine, give the company the ability to provide end user access - via the company
intranet - to targeted content. As Dialog Intranet Toolkit is
very user-friendly, Hinton herself manages its administration with little or no intervention
from Dialog.
One of the major benefits of Dialog Intranet Toolkit is that it gives users the choice to search
according to their level of experience. As Sarah says, "One man's command language is another
man's poison. Dialog Intranet Toolkit really helps people who aren't familiar with online
searching as it offers them a nice easy portal. You can select and pick out files in specific
subject areas so that they have a tailored search. It serves as an introduction to what
Dialog content is available without worrying them about any command language at all."
As a company with a broad research remit, breadth of content was a major consideration
when selecting a company-wide solution. As Sarah points out, "We needed something that has
the broadest scope and Dialog seems to really stand out in the industry. With Dialog, instead
of having to negotiate about 40 separate contracts with information providers on the Internet,
we can have one negotiated contract with one provider."
Roche is one of the world's largest healthcare organisations active in the discovery,
development and manufacture of pharmaceuticals and diagnostics systems. The Group is also one
of the largest producers of vitamins and carotenoids. Its pharmaceutical division focuses on
developing cost-effective drugs to combat human diseases including AIDS, diabetes and various
cancers. Having access to the most up to date pharmaceutical and scientific information is
therefore critical to Roche's scientists working at six research centres around the world.
"Our role is to provide information services that maximise the potential of scientific discovery,"
says Sue Jackson, Roche Welwyn's head of information. "A large part of this service involves
giving our R&D scientists direct access to relevant information tools that help to reduce
research time and effort."
With the help of Dialog, Roche created an enterprise-wide information management system that
delivers data to the desktop via the company intranet. By integrating Dialog's
customisable product, Datastar Web, scientists are able to search a range of global business,
technical and scientific information from Dialog's vast information banks. And, more importantly,
thanks to the Dialog solution and LitLink Pro from MDL Information Systems, they no longer have to
rely on obtaining hard copy of texts from various sources.
"We have implemented a single portal solution that provides accurate one-click access to
full text journals," said Sue Jackson. "This solution has significantly reduced the waiting time
for receiving full-text documents from up to four days right down to an average of five minutes
(subject to the availability of the document, of course)."
Other benefits include:
* Reduced man-hours involved in searching for, and obtaining, full-text documents from
dispirate sources, thanks to the automated process
* Control over document strategy - librarians have the capability to specify where an article
is to be retrieved from, whether it's an external or internal source
* Rationalisation of e-journal subscriptions - Roche can now ensure that there is no
duplication.
