Bold prediction #1: Data streams will eat Batches for breakfast

Data will be distributed based on a central nerve system of data streams and only old legacy systems in Accounting will consume data in a monthly batch mode out of the never-ending stream of data within the company.

Nobody remembers what the words “frontoffice” and “backoffice” mean because there is a full integration of operational and analytics processes and data (see bold prediction #2).

End of 2025 IT is finally able to serve business with implementation of changes & new services within max 1-3 months due to a clean & decoupled architecture with 50% less interfaces (multi-subscriptions instead of point2point connections). Development teams are able to scale and work independently in their defined domain using domain driven design, which leads in total to more than 10.000 releases/go-lives per year.

Innovation backlog was yesterday, now IT teams are waiting for business requests (see bold prediction #3) and in their spare time they are learning how to build chatbots and attending the online-lessons “how do I build my own quantum computer”.

Reality check 2025

Obviously no …. AND yes -> some companies are not sooo far away from at least some parts of that prediction:

Obstacles

Yes, the transformation especially of legacy consumer will take longer (especially in complex architectures of banks) and in worst case will be done with the next end of life, which is normally for IT applications after 4-6 years. There will be as well software vendors resisting the change, basically because they can’t effort the complete cloud native redesign of their software, but at the end they will as well have no choice or will be beaten by (new) competitors.

Silver lining on the horizon

But we have implemented such integration platforms and patterns already in various industries, even a couple of banks. And in each and every case the modern data streaming stack is continuously extended by new use cases, applications and microservices. Some of the companies scale the streaming platform in different speed, some take more advantage by a smart governance that others, but all are continuously migrating legacy or/and invent new processes.

We have even a client who started the journey already a few years ago with 1 yearly release (mainframe architecture) and offloading to a streaming platform. Combining ongoing migration/innovation based on a communicated strategy and with a smart focus on automation reached the status-quo of >1.000 releases per year! Thanks to decoupling of services with a central nervous system, domain driven design (which gained massiv popularity in our software development projects) and high test automation, real CI/CD.

Even a rare number of Accounting departments I talk with are going to daily PnL, from where it is not so far to a “continuous Accounting” -> if someone is interested, I have a concept including benefits of continuous accounting in my drawer waiting for the right timing

There are banks where data streaming & reuse of data is part of the IT strategy, on the same level as “cloud first”. So “Streaming first” will in near future be a paradigm of most IT data strategies and from there on: why should any project want to consume "outdated" data from the complex batch-world instead of subscribing to data streams?