Making changes to key business systems is always highly disruptive, expensive, and risky. So much so that most institutions can make the case to swap out critical software packages only once a decade or so. In a rapidly evolving world, this restriction seriously hampers innovation and agility. In particular:

• Integrating new capabilities, as these become available in products from different vendors, or from outside the core system, can be expensive or even impossible.
• Opportunities for digital transformation, workflow streamlining and other benefits cannot be easily implemented.

For example, deploying rating model changes to business lending operations in large financial institutions can take months.

Our Model Injection™ technology reduces deployment effort to mouse clicks, and time-to-market from months to overnight

RFM also includes a powerful machine learning engine. Model Injection™ automatically serves the functionality and metadata defined by models and data sources to enterprise systems, making both functional and presentation changes available instantly.

Model Injection™ starts with a Services Oriented Architecture (SOA). SOA greatly simplifies enterprise information architectures by unambiguously defining “authoritative services” that are solely responsible for providing a nominated set of capabilities throughout the enterprise. At the foundation of Model Injection™ is the RFM Services Core, which becomes the authoritative source of models throughout the enterprise. The RFM Services Core offers a rich and stable set of data objects and methods for defining models, populating them with data from almost any source, accessing their metadata, and executing the functions they define. A comprehensive and granular security layer with integrated session management and transaction control protects all Model Injection™ services. It includes configurable access control and logging down to the individual transaction. The RFM Services Core exposes both traditional SOAP 1.1, and also REST interfaces.

As with all SOA services, “implementation hiding” ensures that any internal changes made to the RFM Services Core do not affect the client systems that consume those services. While this is one of SOA’s much-touted benefits, it hides a critical but rarely mentioned blind spot:

Implementation hiding ensures that any internal changes made to services never have any effects on the operation of client applications. However, clients are affected when updates are made that require user interfaces to be re-programmed. Model Injection™ eliminates this by extending SOA to include user interface abstraction. To take advantage of this, applications need only “plug in” an automation client (called the RFM Code Stub). This plays a role analogous to a platform-specific SOAP client: it invokes the RFM Model Injection server to generate specific areas of the UI in real time, along with the associated logic. The RFM Code Stub translates between the Model Injection API and the application’s presentation environment (e.g. a web page or mobile app). This is illustrated in the diagram above.

When a model update requires user interface changes, Model Injection™ propagates these, along with functionality changes, to client applications in real time. Function and UI are consistently updated as all updates are controlled by common information held in the Model Repository. Model Injection™ can be implemented to various levels as appropriate to an organization’s needs and existing infrastructure, as shown in the table below:

	Description
Level 0	Standard SOA architecture without user interface extensions.
Level 1	SOA architecture with abstract UI metadata from Model Injection server. Clients are responsible for rendering this in their implementation language.
Level 2	SOA with client-language UI generated by Model Injection server and configuration templates. No work required on client to manage UI.

Request further information.