One of the cornerstones of the Attivio AIE platform is that it affords engineers the opportunity to construct a proof of concept very quickly. This capability means that potential customers can see a live implementation of their content integrated into AIE in mere days, not weeks or months.

Image created using Wordle

Unlike legacy enterprise search solutions, AIE does not rely on pre-built taxonomies or hard-wired configurations. This allows engineers to target their content and easily start ingesting it into the AIE index. If the content is stored in files, then the engineer can just point the file connector at the target content and start ingesting the files. If it is in a database, then the database connector is simply told where the database resides, provided authentication to the repository and then told which tables contain the content of interest.

Our process does not require that the user be over-burdened with a host of up-front questions about what types of queries would you like to see run, what results are you looking for exactly or what types of content do you want to see displayed on the front-end.

AIE utilizes a dynamic schema. Our faceted navigation is built automatically based off of the content that resides in the index.

To give you an example of a recent proof of concept (POC) that was built by one of our engineers for a prospect, I'll give you an overview of the timeline of events from initial contact through delivery and the demo of the POC. This process illustrates how Attivio is agile and responsive to our prospects and customers.

Day 1 - The initial call is held with the prospect to introduce Attivio and discuss a high-level overview of their requirements. They have a lot of research documents in PDF and MSOffice formats as well as a database containing statistics. They want a search solution for their site that enables users to find exactly the reference material they are looking for, while also guiding them to other content of interest related to their search topic. By the end of the call the prospect is impressed by what they've heard about AIE and would like to see a POC using their content.

They want to be able to search across:

• All text

• Author

• Title

• Other fields

They want to be able to sort on fields by:

• Relevancy

• Subject

• Freshness

• Title Name

Day 2 - The prospect sends Attivio a sample of their content via FTP, which consists of a set of PDFs and MSOffice documents representing the research reports and appendices as well as a copy of the database tables that are relevant to the content. These tables contain metadata and statistics related to the content within the documents. This might have posed a problem except that ingesting content from PDFs and Office documents, as well as from databases, is not an issue using AIE. Once the content is uploaded to the FTP site the engineer restores the database and copies the files onto a virtual machine.

Day 3 - The engineer configures the ingestion process to pull in all of the files and database table contents into AIE. The content of interest to the prospect is author, title and all text. This content is ingested and fields are created in the index related to these requirements. Additionally, a database connector ingests all the content from the database that is relevant (IDs and statistics). A default feature in AIE also allows for using a dynamic synonym dictionary mined from the content in the prospect's database.

Day 4 - Attivio demonstrates the POC to the prospect and highlights that all of their content is present in the index. JOIN queries are used to demonstrate a search across both the PDFs (unstructured) and the database (structured) content, all of which reside in the AIE index. The automatically generated facets display navigation options by Subject, Section, Author and Associated Titles, all without any intervention from the engineer. Additionally, spelling suggestions are provided, again, with no effort needed by the engineer to construct the dictionaries.

This prototyping process is not unique to this particular prospect. This is common for many Attivio engagements.

An interesting footnote is that this particular engineer — ME — has been with Attivio for less than two weeks and had no prior experience working with the Attivio Active Intelligence Engine^TM.

This is what we mean by Rapid Prototyping.

Author Bio
Dave Voutila joined Attivio last week, literally. Prior to joining Attivio, Dave worked as an engineer for Empirix where he designed solutions in support of their sales organization.

TrackBack URI for this entry

Comments (0)