 |
|
|
Abstract
Dynamic publication has become a strong requirement in environments where content flow is important and need for customized organization of information is critical. Financial organizations use both expert produced documents (tax, legal, financial) and commercial documents about their products. The overall documentation is used in different contexts, for different purposes, using several formats.
Managing "content structures" in a repository independent from the content itself can significantly improve content publication process, reusability of content itself, and accurate, customized dynamic publication.
We'll show on an use case of assets management documentation how a clear separation from "Content" and "Content Structures" can be used to allow every actor in the information chain to create new structure patterns, as well as new documents, and finally deliver completely customized documentation to the end users.
We'll show how RDF and Topic Maps standards enable this crucial segmentation between reusable content and content structures.
We'll show how "content structure management" tools feet into content management architecture with portals, content management systems, and content authoring tools.
We'll show how separate management of "content structure" open the way to a new family of content management tools such as "content intelligence tools", "content organization authoring tools".
Keywords
Table of Contents
A large financial organization wish to improve its client services and sale teams productivity by providing the employees an efficient access to all useful business documentation. The bank distributes a large panel of products and services from classical banks services to insurance policies, retirement plans, saving plans and address both individuals and corporations. The bank has more than a thousand implantations all over France.
Today sale teams find information on the products in product datasheets and over the bank web server. For legal and tax information each branch receive one copy every year of a "bible" on the subjects. Giving information to the client takes a lot a research, copies of documents, phone calls to the products experts... Changes in product offers or legal rules are difficult to provide. Insecurity of the sales team about some products may prevent them to sell the product or create a bottleneck at the product expert level who are systematically solicited by sale teams.
User needs are the following :
use an unique documentation service on the intranet to access all documentation useful for the sales and client services process,
get smart query services over the documentation to find the useful documentation depending on the client needs and client profile,
personalize the documentation information level depending the profile of the employee (bank teller, product expert, back office employee...),
reuse the documentation to answer the client by mail, attached PDF documentation or with a formal and personalized document.
For each client case the documentation should give a set of organized documents about :
legal context,
tax informations,
product offer,
description of the financial product and its various components,
business processes documentation about product opening, life and closing,
some informations on the main competitors offers for the same client needs.
Content manager need to organize content for the intranet uses, but also for the web publication and the paper publications. Documentation should be available in HTML, PDF format, text format and XML format depending on its final utilization. Several sources of information have to be mixed : content bought from legal and fiscal editors, content produced by the marketing, content from the product experts, business processes from the back offices, content produced by external actors such as regulation authorities or competitors...
The same content should be reusable in different context of documentation publication.
The bank content management team has already built its new documentation system using a content management system for the content produced internally and a intranet portal for content delivery. But none of those solution gives a useful answer to organize the business content and to give an efficient access to it. Several solutions have been studied :
The use of a search tool based on text mining is not a good answer a) the noise on the answer is very important, the result does not provide the documentation in its context but as a collection of independent documents. Linking of documents at the document level is extremely costly with poor result and a impossible management of broken links.
Automatic classification tools aren't the good answer for two reasons a) internal business documentation has already an organization in the actual paper publications. b) automatic classification can aggregate documents but only humans can build an organization of the documents that match the client cases or describe business processes.
Each content management tools in the bank manages a document organization, useful for authoring, validation process and documents metadata management. Those organization cannot be used for content organization in the delivery process : the content organization in content management tools reflects the authors vision; tax, legal, business process, it reflect a functional organization build by experts. On the other hand the documentation delivered to final users must be organized for problem solving, client needs merging documentation produced by various experts, on various field to answer a client need.
Both for content aggregation purpose and to be able to manage a content organization focus on clients needs, it has been decided to build and manage a content structure independent of content itself. The content structure becoming the backbone of all the content delivery process in the intranet portal.
Content structure models already exist in book publication, databases, knowledge management systems. For this case we will implement generic content structure models and specific content structure modes build for the client needs.
The content structure management system is customized with a content structure model. This model describe various types of topics and of semantic links between topics.
Three types of topics will feed the system :
- conceptual-topics used to identify thesaurus and classes topics
- business subject-topics to identify real world objects such as products, offers, financial products, processes
- document-topics to identify available reusable documents.
Five families of semantic relationships will be used to organize the topics :
- semantic relationships for thesaurus organization
- semantic relationships for classification organization
- semantic relationships to organized business subjects as knowledge structures
- semantic relationships to organized documents in publications
- semantic relationships to index documents using thesaurus and business subjects following RDF DC
Except for the semantic relationships organizing business subjects as knowledge structures, all other semantic relationships are generic and can be reused in various contexts.
Semantic relationships used to organized knowledge bases will enable to describe relationships between products, decomposition of products in different components, links with competitors offers and any other semantic relationships that will help to describe useful relationships between business subjects….
Thesaurus is feed with existing indexes of publications, corporate thesaurus and standard financial thesaurus from specialized editors.
Business subjects and their relationships are mainly retrieved from existing databases.
Document organization are retrieved from existing publications of the bank or from financial editors (legal documentation, tax documentation).
All those elements of the content structure are transformed into XTM files before to be input in the content structure management system.
How content structure is used to retrieve the useful informations to answer client requests :
- Filtering using content structure : filtering using classification of topics and user profiles enables to automatically reduce accessible documentation in the further steps,
- Efficient search using topics instead of full text access : thesaurus, classification, business subjects can be searched on. Using this limited set of topics labels gives a limited amount of pertinent answers, using synonyms links in the thesaurus help the user to find the information without too much constraints on how to formulate the query,
- Smart navigation using content structure : Each answer is the starting point to navigation in related subjects, in related documents, related business processes description...
How to deliver the content structure to the users in the portal using the content management architecture ? As for content management solution, the portal users don't access directly the content repository but access to published HTML, PDF documents.
Content structure is published as XML files giving full properties of the topics, their relations with other topics in the content structure and links to documents.
Those content structure XML files are declared to the content management system. The content management publishing process merge the reusable XML content and the XML content structure files into a final document containing content and content structure. The final result in HTML and/or PDF is made available to portal to be delivered to final users.
We get a final architecture where :
- authoring tools manage reusable XML content without elements of organization except a minimum set of metadata on the document enabling to index them automatically with the content structure management tool,
- content structure authoring tools manage all the content structure and the links to validated content,
- document management system manage inventory, versioning, validation of content files and content structure files going all the way to their merge and publication in various formats,
- portal give access to business documentation made of internal and external documents with rich search and navigation tools
This user case demonstrate a very practical and still ambitious use of content structure management system using XML Topic Maps and RDF standards in a state of the art Enterprise Content Management architecture.
A intelligent query of the content structure, accessible as a Web Service would be the next piece to add to the architecture but this very interesting and promising feature will be the subject of a next paper.
  |
Design & Development by deepX Ltd. 2002 |
