In my first post on the IS blog, I was claiming that “Search was Yesterday”. In my view, this is a true statement, but I think that also the opposite is true – at least with regard to retrieval engines.

Let’s have a look at contemporary, modern retrieval engines. They present a very solid and proven technology, they can handle a high volume of (textual) data and there are ways to solve the challenge around change both from the angle of data (new data comes in very quickly) as well as from the angle of the user (different users want different things and one user also evolves over time).

If you google retrieval engines, there are sites which promote “semantic retrieval engines” or “image and video retrieval engines”. And as you probably know, Wolfram Alpha is regarded as a “knowledge engine”.

All this is going in the direction of semantics and context from a search perspective, which is in my opinion absolutely correct and necessary. These roles will certainly have their share in the market in the coming years. But there is an additional dimension, which I rarely see reflected in the discussions on the web: context from a software perspective!

One major aspect where an information provider like Wolters Kluwer brings value to its customers is by delivering the information to the point-of-need, which is very often accomplished through software applications for professionals. The information needed could be existing textual content that we have been publishing for decades. But the user should not have to search for it, he or she would only need a snippet which includes the information and hints where to find similar stuff or more detailed things, etc.

The question now is how to create an intelligent interface between the information and the software application, which takes into account all the context information coming from the software and also knows everything about the content that contains the answers. So the software itself is asking the retrieval automatically on a more or less permanent basis for hints that help the user in his current situation. And these hints and answers are displayed as context information within the professional workflow the user is currently executing.

I am pretty sure that for building a solution that really works, we could highly benefit from the strengths of modern retrieval engines described above. Still the gap is huge.

But where are the alternatives? Creating all the distinct answers from scratch and include them hard coded in the software application? Inventing a completely new intelligent interface? Simply give up and not support the customer adequately?

Let’s have a closer look at what we already have: technology and domain expertise. There we will find the answer.