Faster Time to Value with Pervasive Integration
SaaS offerings has seen a surge in the IT world primarily owing to the economy of scale in making a Configurable COTS product that gives value out of the box. But todays SaaS solutions are never standalone implementation, adding to the platter - AI/ML, Internet of Things, Robotics, Cloud systems, On premise and niche applications all bring one common requirement - the need to talk to each other in an efficient way. This exposes a problem Customers face that many stalwarts have tried to solve it, time and again - Standardizing how applications interact with each other.
Historically we have seen consortium of experts produce standards like EDI X12, EDIFACT, xCBL, Rosettanet etc for the very purpose of standardizing integration. Though some of these standards continue to adapt and stay relevant, technology companies that embrase change don't see them to be nimble enough to adapt to their growing needs - as a result more and more cloud solution providers adopt a Self imposed standard - with the perspective that customization can be handled at the integration layer.
Moving away from Centralized Integration Team
A centralized integration team was a critical success factor for an organization to do electronic data exchange in the erst-while world where various standards needed to be adopted. An organization's ease of doing electronic business was judged by the number of industry standards it supports. This meant, much of the time spent by the central team was to build a standard map that can be reused to scale it to other implementation who also followed one of supported standard. These exchanges were Server (provider) heavy for the client (consumer) to integrate with minimal effort. As a result the centralized B2B integration team played a key role in handling the entire architecture, development activies and ensuring scalability thus shaping the offering by the organization.
- Internal - Custom standard <- Producer Integration -> Industry standard - External application
- Internal - Industry standard <- Producer Integration -> Industry standard - External application
- Internal - Industry standard <- Producer Integration -> Custom standard - External application
Open API standards
With the advent of SaaS products, this changed with each product adopting a light weight integration approach by using REST APIs, leaving behind the server heavy integration techniques for obvious reasons - to be nimble and flexible to make changes without the need to have consumers change everytime. A tradeoff to this adoption is the fact that there is no established standards in the world of APIs and as a result every product started to come up with their own way of exposing their data, allowing the client to discover their schemas, structure and expected responses. Leading us into the Open API standards where the only standard is the openess of information to help the consumer understand the product behaviour better.
- Product (Open API) <- Consumer Integration -> Business need - External application
Drift towards Pervasive Integration
Over time the centralized integration team starts to see a decline in Producer defined format in favor or Rest API based integration given the move from a Stateful integration product to stateless integration products that can also be harnessed from the cloud. Combining integration platform as a service (iPaaS) and REST APIs consumers are more forthcoming to build product specific integration basis business need (not necessarily industry standard). This is now the new "configurable" COTS allowing integration to be the mediator.
AI Enabled Integration for Business Users
Modern integration technology providers have already AI-enabled their platforms to simplify the development of integration flows by giving suggestion to the user on what is the next logical step as a result shorten the learning curve, lower skills requirements and assist in the execution of integration processes. This is achieved by AI/ML which constantly gathers usage and patterns to better predict and suggest user guiding them through the integration build
This results in
- Pre-packaged artifacts for business understanding (like Swagger UI) to make learning more intuitive and guided development
- Easy to use (example: Drag drop - UI based integrations) resulting in empowering less-technical roles to solve integration problems. Users can build integration with minimal support from professional developers or integration specialists.
- Faster Value creation with connect systems basis business need
This will now empower many users to start taking up repeatitive integration task clearing the backlog of integration activities. This essentially make integration pervasive across all business line instead of being centralized.
Integration no longer needs a focused development team. Instead, business users can, with AI to aid them, start to visually create integration. This means that a significant reduction in developers time to build integration and the developers now either take on the business side of the work or end up isolating themselves from the world of lightweight webservices.
A move from a Development team centric Integration -> Business User empowered integration giving a lot of control and agility to the project team
Integration Center of Excellence
Does this mean the integration team becomes irrelevant? May be not completely, there is still a need to experts for supporting business users and also to ensure the solution is built right given the process, operational overhead and architecture, making the Integration developer/specialist a solution architect. This becomes the Integration Center of Excellence.
Examples of AI powered integration products in the industry - Dell Boomi's Boomi Suggest, SnapLogic's Iris Artificial Intelligence, Informatica's CLAIRE and Workato's Workbot.