We recently gave a joint webinar with IBM on Reimagining User Experience with AI-Powered Automation. We talked about the Business Automation Open Editions (BAMOE) and how AI tools like IBM Bob and watsonx Orchestrate can be used to complement it with modern user interfaces.
BAMOE is great – it is powerful and easy to manage, delivering a cloud-native process/state engine and decision engine. Both the process engine and the decision engine are driven in the latest release by industry standard models: The process engine uses Business Process Model and Notation or BPMN and the decision engine uses Decision Model and Notation or DMN. This combination of enterprise support from IBM and standards-based, open sourced, cloud native software is a very powerful one and Blue Polaris has a number of customers in production with it today.
One challenge has always been that BAMOE is a “bring your own user interface” environment – it has basic UIs for process tasks and administration, but most serious users end up building their own. This meant there was a lot of coding to do if you wanted modern, nice-looking UIs.
Now, with AI coding tools like IBM Bob and agentic interfaces like watsonx Orchestrate, there’s a whole new set of options to build these interfaces.
Basic outline of the approach we showed in the webinar
- First, you need to define the backbone of your application. Using BAMOE Canvas and VSCode, you populate a git repository with the core process and decision models for your solution. BAMOE then builds cloud-native applications based on these models, leveraging the full set of open source capabilities.
- With these models in hand, IBM Bob and other AI coding tools can build end user and. These UIs are driven by the tasks in the process models and the decision services defined in the decision models. These provide the key data elements that must be captured as well as information on how those UIs drive the process forward.
- You can apply your own style sheet, colors and logo so that Bob will build a suitable look and feel into the interface. Our experience is that you have to iterate quite a lot initially, but once you have something you like, it’s easy to save a summary of your instructions. This allows Bob to generate future UIs more rapidly and get you to a “nearly finished” look and feel that is easy to clean up and finalize.
- Because BAMOE exposes all its administrative tasks as APIs, you can use the same approach to generate modern UIs for process administration, claiming processes, checking on status, tracking results and more.
- You can use the new MCP service to integrate AI agents, built with watsonx Orchestrate for instance, to provide a chat or natural language interface for the process. Instead of traditional form-based UIs, or as a complement to them, these chat interfaces can be very effective. This is especially true for more casual users who aren’t going to use a UI enough to learn it.
- With that in hand, you can go one step further and link task completion activities to things like Slack or Teams so that occasional users don’t have to leave an environment they are familiar with to interact with the process. They can simply ask their favorite environment what tasks they have open, it can query the process engine, and then the user can answer questions in natural language to complete those tasks.
Taken together, these AI tools make it easier than ever to bring your own UI, letting you deliver a great user experience and driving up the value of the BAMOE back-end.
If you want to see a detailed demo or have questions, get in touch here.
You can get also access to the recording and the slides here and learn more about DMN, in this great book I wrote with Jan Purchase – Real World Decision Modeling with DMN.