Technology Blogs by SAP
Learn how to extend and personalize SAP applications. Follow the SAP technology blog for insights into SAP BTP, ABAP, SAP Analytics Cloud, SAP HANA, and more.
Showing results for 
Search instead for 
Did you mean: 


SAP Cloud Integration version 3.30.**/ 4.17.** comes with two enhancements/features

  1. Flow Step Recommendations

  2. Quick access to flow steps palette list and search

The first enhancement is named as Flow Step Recommendations. As the feature name suggests, for a given flow step - while you are modelling/designing an integration flow - it recommends you a next matching flow step. The recommendations are provided by a machine learning model.

The second one gives quick access to flow steps list/palette to add a flow step and also it addresses current issue of missing search support in SAP Integration Flow editor’s left hand side flow steps palette.

Below is a sample screenshot highlights these enhancements.

Note: You have to be in the Edit mode to add a flow step.

Introduced (+) Icon/Speed button on the flow step selection, which open up a pop-over view


Recommended steps and All steps sections



Flow Step Recommendations

As initially described, flow step recommendations are made by machine learning model, which captures the flow steps combination available in large collection of SAP pre-shipped/standard integration packages. Here, SAP standard integration packages and scenarios act as ‘data set’ for machine learning model. This machine learning model doesn’t read existing integration scenarios of your tenant (from SAP Cloud Integration perspective, we don’t read customer integration scenarios for this machine learning capability, we use our own standard integration content for machine learning).

The recommendations start appearing when you start modelling your integration flow. When you click on a flow step, and press on the (+) icon, the machine learning model behind the screen takes the selected flow step as input and finds the next matching flow steps from its analysis result and proposes 3 flow steps. These 3 flows steps will be sorted based on the best matching ranks as per the machine learning model.

Below are the sample screenshots which provides a different recommendations for different flow steps.




Note : Receiver flow step for Request-Reply recommendation

For a Request-Reply flow step, the recommendation will be a Receiver flow step. But on click on Receiver step, the flow step is not dropped/created on the canvas (because it has be created outside of integration process rectangle) rather, the mouse pointer gets receiver icon, and when you drop the icon outside of the integration process rectangle, then the receiver step is created on the canvas. This will be same behaviour, i.e. mouse pointer icon changing to receiver icon, if you search and select “Receiver” step from the flow steps list.

Sample screenshots below

Select Receiver flow step


Mouse pointer gets Receiver flow step icon


Drop the Receiver icon outside Integration Process rectangle/box


Receiver participant box created



Planned Enhancements

Below are the topics on which we will be working on with "Flow Step Recommendations" feature

Machine Learning Model Fine Tuning

As machine learning is a continuous and iterative method, we will also be enriching/fine tuning our exiting machine learning model to recommend as accurate flow steps as possible. Also, there are some suggestions which can be corrected, e.g. the “End Message” flow step sometimes may get ranked higher, but if there is already an “End Message” flow step available in the integration flow, the appearance and rank of this can be fine-tuned. We will be working on such fine tuning activities.

UI Rendering performance

Sometimes, you may feel a very little UI lag between click on (+) icon and appearing of pop-over view with flows steps, we are working on this fix this UI performance.


Quick access to flow steps list and search

If you find no matching flow steps under the “Recommendations” section, you can use the below section, titled as “All Steps” in the pop-over view, which represents all the flow steps, which are similar in entries of the left-hand side palette.

You can search for a step, click on the desired flow step, which puts this flow step onto the sequence connector line. Below is the sample screenshot

Search for a flow step and use it


Note: For some flow steps, like Groovy Script and Message Mapping, there was a (+) icon previously and used to create a groovy script resource or message mapping resource and used to open the corresponding editor. Now, since the (+) is used to create flow steps, there is a new icon introduced to create a resource. Below screenshot describes the same.



Planned Enhancements

Below are the topics on which we will be working on with "Quick access to flow steps list and search" feature.

(+) Icon on Sequence Connector

As of now, the (+) icon is only available on the selection of flow step. We will be providing this (+) icon on the sequence connector line as well. This becomes a must have for the flow steps which are multi-outgoing connectors in nature, e.g. parallel/sequence multicast, router etc., wherein, currently clicking on such flow steps will create a new branch and then adds the flow step. But if you want to add a flow step on an existing branch, then a (+) icon on the sequence connector/line will come handy.


(+) Icon on sequence connector


Name Corrections

There are some name corrections which we need to do, e.g. there is only entry of “JSON to XML” in “Converter” section, and there is “XML to CSV Converter”, we will be fixing them.

Auto Resize

As of now, if you add a flow step, either from “Recommendations” section or from “All Steps” section, it will add only the flow step, but the tool doesn’t resize, e.g. push the existing flow steps to right hand side by some pixels, increase the width of “Integration Process” pool/box etc., that is in one word Auto Resize after adding a flow step is not available. You need to increase the size of those shapes manually as of now. We are working on Auto Resize feature.



With introduction of these two enhancements, we plan to increase the integration flow developer’s modelling efficiency.