Many consulting organizations and individuals have a high degree of competence in specifying and designing knowledge systems. Unfortunately, however, there are many more consultants whose capabilities fall far short of claims. Various arrangements can be worked out with consultants.
Some prefer to take total responsibility for the project. They will participate in selecting the project, elicit the knowledge, then design and test the complete system. Their preference may be to take an assignment for a single large and complex system rather than a number of smaller systems.
An organization might select such a consulting arrangement if it feels there would be little need for developing additional systems or if it feels that the time and expense of developing in-house competence is not warranted.
Another approach is to have the consultant only support the development of a knowledge system rather than take full responsibility. The consultant participates in selecting and specifying the system and, perhaps, carries the responsibility for the initial prototype.
During this early phase, the consultant might also provide training for the organization's own development personnel. Subsequent to acceptance of the prototype, further arrangements would be made allocating responsibilities for the balance of the project.
A disadvantage of this approach is that development time and cost would probably be somewhat greater since the consultant must take time out to train the organization's personnel. An obvious benefit, however, is that the organization eventually acquires in-house competence.
There are a number of trade-offs in working with consultants. The cost of consultants can be quite high, but this must be balanced against the cost and time in developing in-house competence. Good consulting organizations may have a broader range of experience than that available in-house. They may have better access to skilled personnel and, significantly, be better able to retain the services of these skilled people since they can offer a challenging range of assignments.
Consultants may also be more objective about the organization's requirements or problems as well as the market potential or user interest for such a system. (There is possibility of a conflict of interest here.) A disadvantage arises if the consultant does not know much about the organization's procedures and business environment. The organization would, therefore, be paying to educate the consultant in the specifics of its own operation. Some destructive antagonisms may also develop. If reliance is placed upon the consultant for a complete package, in-house personnel will gain only limited experience in adapting the package to the organization's infrastructure.
Several measures should be taken to improve the probability of having a successful project. Initially, of course, the consultant's competence in developing the type of system required should be evaluated. Secondly, a careful understanding of the responsibilities of both the organization and the consultant should be worked out.
There may be many unknowns which complicate the writing of an appropriate work statement for an Al system. Provision should be included in the work statement for further definition as work proceeds and more information is obtained. The work statement should cover such other matters as the extent of training to be provided; requirements for accessibility of domain experts; acceptance criteria for both prototype and complete system; provision for maintenance.
Finally, costs and schedules must be defined, although there should be sufficient flexibility to accommodate the uncertainties in this type of project. Contractual stipulations should cover such matters as eventual ownership and marketing rights for the system, protection of proprietary information, and competitive constraints.
Sourced from ToolBox - written by Craig Borysowich