The State of the Art in Search
If you are interested in search technologies, you might have come across an excessive amount of articles regarding the use of graph search in the social media environment and how much insight it can deliver. I have noticed that people are very enthusiastic about it and more and more are adopting this new way of searching. Connections between people, their actions – such as opinions, likes, shares and analysis of entities involved – are all part of the universe that can now be surfaced using a modern search platform.
When discussing the subject with customers and peers, it seems search is still viewed as a complex technical tool or activity with its own set of specialized queries and mysterious algorithms. But when thinking of modern graph search use cases, users can browse or search through a universe of connected people, items and associated topics; and through the searches themselves, ultimately express their own interests through their actions.
Once users have the ability to browse for topics of interest and combine that with natural, easy to understand queries, search technology ultimately moves beyond the classic keyword search dogma and gets perceived as a more natural and easy to understand tool.
RAVN and Enterprise Graph Search
Navigating social data is definitely exciting. At RAVN, however, we found that searching graph data can be even more useful and interesting in the enterprise world where social bit is just one of the dimensions.
Real-time search based on a person’s graph within the enterprise is something I always expected to see at places employing knowledge workers. It makes sense and I believe search technology should evolve and be employed to get you to your answer with as little effort as possible.
That said – in real-life I am still on a daily basis confronted by many basic systems/search engines that are probably utilising very little of available knowledge from the “structured world” within enterprise and even less from the unstructured one. But this simply means we have to try harder, people should give new technology a chance; and adding knowledge, structure and relationships to unstructured information is key.
I believe that technology that can accurately analyse and, most importantly, connect pieces of information into a coherent, scalable and navigable way has the potential to add immense value to any organisation. It should therefore become a critical component of business intelligence, powering many organisational processes. There is tremendous potential for business improvement, if you put engine, such as RAVN Core, at the heart of your organisation.
At RAVN we realised early on that to become the best at providing graph search for enterprises, we had to be very smart from the ingestion process on. Analysing data cannot simply be resolved by throwing masses of resources at it. For instance, the fashionable solution of simply using Hadoop is not the answer in the environment we are operating in, although it too has its uses in some contexts.
Therefore we built our own engine that delivers what we were looking for: a hybrid of best of breed object storage, scalable full text search and an innovative linking engine. Stemming from the belief that knowledge has to be at the core of everything we do, the RAVN Linking Engine has evolved into a crucial component of our Core engine and our own graph search technology. The Knowledge Graph ® – was born.
How does it help enterprises?
There are many types of “graphs” within an enterprise. RAVN’s enterprise graph search doesn’t limit the types of relationships that can be handled, searched for or exposed. Sometimes social links within an enterprise are not the most efficient way of linking people and knowledge – sometimes it is hierarchy and interest areas that matter most. Those are the types of inter-relationships that get highlighted when applying a graph view to the knowledge corpus of a modern enterprise.
From expert communication pools to a team members exchange; from before unknown connections to opportunity discussions, interest suggestions or activity stream entries – it’s all within an enterprise, but it has been so far much underexploited. An enterprise’s knowledge should not be limited to social interactions or connections only as in classic social graphs – there is plethora of valuable data that we can analyse, enrich or simply surface using our linking engine.
I realise this blog entry is quite academic and involved, so the following posts will describe some of the practical uses of our technology, how we can provide business intelligence and help organisations make right decisions and see the unseen.