Most modern data platforms will now incorporate some capability for basic artificial intelligence, although this area is still awash with buzzwords and over-promised imagineering! A highly innovative organisation looking to make a significant investment may want to consider a full enterprise-wide AI platform. There are three basic capabilities however that we would recommend most organisations consider when implementing a new data platform:
- The ability to directly control a business process from the data platform. This could be data sent directly to an internal CRM system to show customer behaviour, data sent to a marketing platform to determine which offers a customer receives or data sent to a front line system to allow certain self-serve actions based on risk profile data. This is known as data driven processes and is a capability in addition to the usual analytic capabilities.
- The ability to explore data (or big data as larger datasets are sometimes referred to) with the sole purpose of innovation. This is to look for new business opportunities or commercial advantage within your data as opposed to traditional analytics to run the current business as efficiently as possible. The skills needed here will vary, depending on the complexity of your dataset and business challenge and may well justify an individual or team of data scientists to fully exploit the data set. This will require additional tools such as R and Python and some sort of “lab” area of your data platform and depending on the size and type of additional datasets required for this might need modern data ingestion capabilities, for example to read stream based data and modern data storage technologies to store larger structured or unstructured/semi-structured data sets, in what becomes known as a data lake.
- The ability to produce a model that then runs continuously against a data set. This could be a machine learning algorithm or a much simpler piece of SQL or other code. This is about the ability to productionise the innovative data science that as been developed and exploit it either in analytics or a direct data driven process.
Kellerton can help you understand what level of business value you can really achieve in this space and help introduce an appropriate level of skills and technology. We can also help implement and develop these AI models for your specific needs.