AskJeeves Inc., Recommind Inc. and Google Inc. are each attempting to solve the vexing problem of enterprise data search.
AskJeeves last week announced the latest version of its natural language search product, JeevesOne Enterprise, which can tap into data from enterprise applications to find answers to users natural language queries. This gives users more complete answers to questions, said company officials in Emeryville, Calif.
In addition to searching enterprise applications, JeevesOne Enterprise can crawl data marts, legacy systems, and corporate and partner intranets. The software generates results from structured and unstructured data.
MindServer 2.0, which was introduced this month by Recommind, of Berkeley, Calif., takes a different tack, searching on words or phrases and grouping results into topics and categories within those topics based on their relationship to one another. For instance, a search on "Java" could lead to separate results for Java the programming language and Java the Indonesian island.
Like JeevesOne Enterprise, MindServer can search unstructured and structured information.
Meanwhile, search engine developer Google last week said it had upgraded its search appliance, which the Mountain View, Calif., company introduced this spring. The Google Search Appliance bundles Googles patented search engine software with a dedicated piece of hardware that can be plugged directly into a corporate network.
In the past month, Google has added administrative enhancements that make it easier to customize the user interface to match a corporate intranets look and feel.
Google has also doubled the capacity of its departmental appliance, the GB-1001, so that it can search through 300,000 documents, officials said.