In a previous article “How can I bring the power of cognitive services in the cloud to my personal documents, apps and devices” we discussed how end users can leverage the power of Daneel’s cognitive services to explore personal data without exposing this data to the cloud. This capacity of Daneel to act on private data can also be brought to bear on enterprise data.
Enterprise applications can be internal in nature like departmental documentation, analysis of spreadsheets, productivity analysis in real time, analysis of critical production paths in real-time as well as smart city data and event aggregation. It can also be destined to public consumption like ecommerce data (search and guidance), smart city applications for the consumer (utilities usage, traffic monitoring, incident reporting) and the list goes on.
Keeping enterprise data private and local.
Much like private data, enterprise data is usually controlled at some level for privacy protection. To be useful all data might need to be analysed but only some data can be made available to the public, or to particular enterprise user roles. Where that data is physically located is also something that is closely monitored by certain types of enterprises.
One brain is good. More brains are better.
So how can Daneel bring the power of its cognitive services to data which is hidden from the cloud? Well that depends on who is asking, and what is being asked. But before we get to that you must first understand that in concept, Daneel is like a brain. A brain stuffed full of knowledge and useful tools to act on that knowledge. Daneel can also be replicated. So I can have more than one brain loaded with the same knowledge and tools…but I can also have brains that contain particular types of knowledge and tools. This is just like humans, some have similar knowledge and skillsets while others are trained in a completely different way with different knowledge sets. Further, many Daneel instances can break apart a big problem into smaller more manageable tasks. Efficiency and speed of execution is increased while data is spread out across multiple instances but clustered by groups of subjects. This is important because it is here that enterprise data can find a home. Enterprise data and tools can be placed in its own local Daneel cognitive service (LDCS). A direct copy perhaps but most likely be loaded with only the knowledge and skillsets required by the enterprise user.
The enterprise brain is not alone, it can ask for help.
If required, the enterprise LDCS can be allowed to contact the global Daneel brain if it is unable to fill a gap in its own knowledge. For example, an internal user might ask an enterprise LDCS for a particular type of report with a particular type of graph in it. It might be that the enterprise LDCS has never constructed that type of graph before, so to solve the user’s problem it could ask the global Daneel brain if it knows how to build the graph. If the global Daneel does then it can transfer that knowledge down to the enterprise LDCS. So again just like humans, when one does not know how to do something, ask someone who might for help. Ask, learn and grow. Knowledge is not static and neither is business so enterprise solutions need to be able to adapt to constant changes.
Global searches can end up at enterprise data.
Take a case where a large retailer is using a Daneel enterprise LDCS to assist their users with ecommerce searches (search, guidance) within a large product database for their own website. This retailer could configure their LDCS to speak with the global Daneel brain to expose various concepts found within the enterprise LDCS (like types of products available). A user could ask a question to the global Daneel brain which might determine that part of the answer to the user`s question might lie within the retailer`s Daneel LDCS. It would then ask that brain to respond through the global instance to the user.
To highlight this use case better, imagine a user going to a global search engine and entering in a search: “I am looking for a 2X4”. Daneel might know that a 2X4 is a dimension that is commonly used when describing lumber. It also knows that when I say “I am looking for” that my intention within the context of this piece of lumber is to procure some. So Daneel will need to return results that are contextually meaningful with respect to a piece of lumber with the dimension 2X4 and that it also can be purchased. Thankfully attached to the global Daneel brain is an enterprise LDCS that is loaded with all the products and services of a hardware store. The global brain can ask the enterprise LDCS what to do with the request “I am looking for a 2X4”. At this point search and guidance is delivered by the enterprise LDCS which has the best knowledge to deliver the best user experience for this search.
A community of knowledge
The distribution of knowledge by region and by business interests can bring a better experience to end users who need to work with this knowledge. The ability to share new knowledge between cognitive service instances dramatically reduces the time required to propagate new data and tools across an enterprise. Looking at data as a whole can provide greater insight onto business operations as opposed to traditional models where data exists in application silos and cross pollination of data is strictly controlled not by need but by limitations of developer and systems analyst resources.
A business should be able to adapt instantaneously as their data and tools grow. A business should be able to leverage new knowledge as it becomes available. With Daneel that capacity becomes reality.