We are in the midst of interplay of disparate technology solutions in an enterprise. It is always comfortable to sit back and strategise to clean up existing ones and have a compact set of solutions that make life a lot easier. This is purely academic utopia.
The reality is the old and the new will have to coexist and one must find the indifference curve that is most beneficial.
The advent of standard APIs that is solution / industry specific is something to cheer about. This has given the integrators a level of predictability that did not exist before. The true worth in the APIs will be in automation and machine learning. I am
suggesting intelligent APIs.
A Case in Point
Let us look at two solutions that need to work in tandem. A good example will be CRM and Mortgage solution. The APIs standard will outline the standard interfaces that each of the solutions must make available. The integrator will merge this together to
establish seamless communication. This again will be a project specific activity that needs to be repeated in each implementation.
The Best Case Scenario
Developingsoftware tools that automate the integration process will give true value to APIs standards. So how will this be done? The software tools will have in built 'intelligence' with extensive business rules. The tool to begin with will create a paradigm
of the solutions to be integrated. In this case a CRM-Mortgage paradigm. This will be an automated process. The specifics of the solutions will be 'Machine Learnt' by the tool and a blue print of integration touch points with business and technical data dictionary
released. This will be analysed for accuracy. Once decided the integration will be activated, tested and released to production.
The integration nightmare and associated unpredictability will to a large extent be minimised. The assumption I have made here it is a COTS solution implemented with little or no customisation.