Integration mitigation

Much of what you say is true, getting the integration right is a huge issue, I have two thoughts.

A complete packaged solution does remove risk and complexity.  However if all components are tightly integrated you risk creating yet another silo where functions like WLC and AML are duplicated rather than using this opportunity to standardise on enterprise components.  A complete package with loosely coupled components which can be replaced with enterprise alternatives would be preferable.  Clients have a choice then.

This raises my next thought; you then have to integrate with these enterprise components and 3rd party systems. This has always been the problem with payment engines as they concentrate on the job of processing payments and not on the issues around ‘working with’ all the other systems a bank runs to provide a true end to end solution.  This is where payment hubs came from, a solution that helps a bank integrate payments processing from the front office to the back office and out to clearing and settlement by providing tools, frameworks, and web services around a standardise payment process.  This drastically reduced implementation time, complexity, risk and gets rid of bespoke interfaces and silos; it reduces those cost you suggest are stopping investment.  It also provides a modern infrastructure making it easier to offer innovative new services in the front office whilst supporting changes to clearing and settlement systems (such as SEPA) without major upgrades to the core processing systems. 

Hubs are therefore different from engines, but ultimately both are used to create payments solutions, the difference is in the time, approach and modern standards based tools provided to create the solution.