In the business automation market there’s a lot of interest in low-code, which has acted as a spur for different vendors and analysts to position themselves around this. For no-coders, low-code still means code, which is just not a good enough answer; for pro-coders, low-code is not real developer code, which is also not a good enough answer. Of course, low-coders think it’s just right: a Goldilocks solution.
How do business leaders decide on the best solution? And where does Flowable sit in all this? For us, it’s not about the answer, it’s all about the question. What are you trying to achieve and what aspects do you have flexibility with? The best solution is automation at the level you need at the time you need it. There may be parts that are no-code, some parts low-code and others pro-code, and over the life-time of an automation project, these may change.
One of the real challenges of using no-code and low-code products is the cliff edge that you inevitably meet at some point. Suddenly, the complexity of the automation you’re trying to capture is beyond what the product can do – it’s great to get speedily to some point, but then there are sharp boundaries to taking it further. One answer may be to switch the complete implementation into deep, pro-code, but this creates a very 2-dimensional perspective of a solution.
At Flowable, we’ve taken a model-driven approach, which by its nature is no-/low-code: model more, code less. The models are what is executed by the Flowable automation engines. These models are rich, expressive, powerful and, where possible, based on open standards. The sophistication of the models means that you can express complex flows and interactions between services and people, as well as being able to define simple flows simply. From quick review and approve workflows to highly-regulated, core automation used by financial institutions, the capabilities of these models mean you can describe almost anything.
A really important feature of Flowable automation models is that they are composable: you can build more complex models out of smaller ones, or re-use fragments of models in other solutions. Create a process for validating someone’s identity and use it consistently across all your solutions. If it needs to change, then all applications get updated with the changes. You can structure your models like macros and functions, but still no need to write code.
The real power comes because these composed or nested models don’t have to be of the same type, you can have a process model starting a case model that in turn executes other process models. Users of Flowable have found this combination of process (BPMN) and case (CMMN) models to be like a superpower when trying to manage Intelligent Business Automation.
You can also find a cliff edge when you come to scale up the use of your solution if the technology you’re using hasn’t been architected with serious throughput in mind. This is the challenge with many no-/low-code products, where again you have to jump to code to make things go fast enough at scale. Having rich models with formal semantics allows us to continuously look for optimizations with the confidence that they are more testable and predictable. Our considerable investment over many years of real-world automation has been to make the engines perform ever faster and more efficiently. Whether embedding Flowable’s engines within microservices or using them elastically to run millions of processes an hour, we’re confident we’ve covered the widest spectrum of scale.
Similarly, when you find you need to customize your solution, you are ultimately limited by what your automation product allows you to extend. Having clear models to work with makes it much easier for a product to offer hooks to extend or change their behavior safely and consistently. If the product is open source, of course you can change what you wish, but you have to be careful about upgradability and maintenance. There are times when it may be more efficient to code a very specific part of your solution, regardless of how fast the automation engines are. With our open source code and mindset, we’ve always provided multiple ways to customize Flowable in a way that is fully supportable.
Another cliff edge you can discover is availability and cost of skills and experience in using a specific automation product. The more expensive the product, the more expensive the people are to employ to build your solution. Using open standards-based automation models running on open source engines allows a wide range of people to develop the skills and experience independent of an actual product. The pool of resources to help you gain the knowledge inhouse or to hire in is much larger than closed and proprietary products. Using automation models such as BPMN allows the person defining the solution to be closer to the business, such as a business analyst creating models directly, or through better understanding in fusion teams of business users and developers.
A model-driven approach is inherently no-code and low-code, but by the use of open standards, open source and open architecture, Flowable allows you to use pro-code when appropriate. For us at Flowable when working with our customers, we don’t push a particular marketing fashion or technical ideology, we listen and guide to a right code solution.
Sustainable competitive advantage in today’s business landscape requires a strategic foundation that goes beyond traditional levers of business growth. BPA lays an important foundation for embracing new technologies such as AI.
Will Generative AI replace BPA? We asked Paul Holmes-Higgin, who is co-founder of Flowable and has a PhD in Artificial Intelligence
Explore Gartner’s Hyperautomation trends for 2024! From Fusion Teams breaking IT-Business silos to AI augmenting capabilities, this comprehensive overview dives into the evolving landscape of strategic automation.