Skip to Content

by Gianluigi Bagnoli, Yatsea Li, Thiago Mendes, Edward Neveux, Ralph Oliveira

In a previous blog we saw how to untangle your solution so that you are able to address more market segments, notably covering both SAP Business One and SAP Business ByDesign customers.

The goal is for SMB business owners to choose your solution amongst all the others if you offer them something different. If you stick to the usual well-known best practices of your industry, your solution will get lost in the big crowd of same solutions. What will differentiate your solution from the others, is to propose disruptive scenarios that make your solution more effective, more usable, simpler.

In a single word: intelligent.

The term Intelligent can be vague at best. For ERP solutions, we can define “intelligent” as a solution that abstracts the intricacies of internal business processes and gives the final user the possibility to do what he/she wants to do in a simpler way.

This simplicity breaks down to three main areas:

  1. The first area is clearly the User Interface (UX). Gone are the days in which an end-user has the patience to open an application on a desktop, find the right menu, open the right form, enter all the right parameters from a huge list of possibilities and click the OK button. Even the idea of installing and using an app these days becomes tedious. Technology now allows users to chat via standard messaging apps (like WhatsApp or Facebook Messenger), and these chats are usually not only text-based but very rich in content. If your solution can provide a differentiating and consistent conversational experience, you surely have an edge on older legacy based desktop
  2. The second area is on the other side of the spectrum, i.e. the back-end. The more intelligent your Digital Core is, the more intelligence your solution will be able to leverage for your In SAP we are betting significantly on what we call the Intelligent Enterprise and you can already see this in many parts of the product. Think about about Intelligent Forecast, Item Recommendation and Churn Probability.
  3. The third area sits between the areas mentioned above. the UX and the Digital Core. This is your solution. The more your solution has built-in intelligence and streamlines all the industry best practices in a seamless, simple and easy user experience for your customer, the greater edge you have against your competition. The good news is that you don’t have to reinvent all this, you can reuse all the goodies that SAP implements into its core, which are made available as a platform. All you have to do is orchestrate the right services from the SAP Cloud Platform and SAP Leonardo to bring intelligence into your solution. You can find more details about SAP Leonardo here.

If this seems too easy, let’s see a concrete example of what we are talking about here.

Shopping should be easy

We have developed a proof-of-concept that shows these concepts and illustrates how easy it is to consume business services, intelligent solutions and interchangeability of the underlying Digital Cores.

Have a look at this video:

There are a few interesting points in this proof-of-concept:

  • There is no desktop, the solution is fully mobile. There is not even an app, the solution is fully based on Facebook Messenger. The user does not have to find, download and install an app and learn how to use it. All he/she has to do is to connect to a Facebook Messenger user and start chatting.
  • The first intelligent aspect of the solution is the chat: all the underlying machines and systems and different ERP instances and ERP processes are abstracted by a natural language dialogue. This dialogue is not only textual but also heavily based on photos and geospatial information.
  • The second intelligent aspect of the solution is all the magic being done by recognising images and connecting them to the right items in the Digital Cores. The user does not deal with item codes and prices in different systems, but rather directly with images of what he/she desires to buy. All this magic is implemented in the solution itself by fully leveraging the power of SAP Leonardo services and is completely hidden to the end user.

You can see the overall architecture of this proof-of-concept below:

Let’s dig a bit deeper and see some interesting implementation details.

  • The server-side is implemented using Node.js on the SAP Cloud Platform. This solution is completely loosely-coupled with respect to the different Digital Cores, in this case SAP Business One and SAP Business ByDesign. There are no solution components installed on the Digital Cores, which are kept vanilla.
  • All the differences between the Digital Cores (in terms of connections, services, formats and semantics of the business objects, etc.) are abstracted by an adaption layer. The solution then handles business objects (in this case Items) that are abstracted from their particular implementation in the different underlying Digital Cores.
  • The sessions IDs with the Digital Cores are stored on REDIS, an in-memory standard DB, which is natively available on SAP CP (details here). This is one of the many services available out-of-the-box in SAP CP thanks to its standard nature.
  • The solution is very lightweight: there is very little duplication of item information, which is kept in the Digital Core. Only the item codes are stored in the solution to maintain the link with the Digital Cores.
  • The solution is structured in two micro-service based applications: the first one handles the dialogue with the Facebook Messenger, including the geospatial processing; the other handles the interaction with the Digital Cores and the usage of Leonardo.
  • The intelligence of recognising images is implemented by fully leveraging SAP Leonardo services (you can find extensive documentation here) :
    • At system start-up the URLs of the item images are downloaded from the Digital Cores, and a Leonardo service (feature extraction, see documentation here) is called to extract all the features vectors from them.
    • All these vectors are then stored in a Postgres DB and the images are thrown away. Postgres DB is just one example of the many persistency services available on SAP CP. You can see more details here.
    • When an image comes from Messenger, feature extraction is called again and the resulting vector is sent to similarity scoring (see documentation here), which returns the items that are closer in similarity to the image.
    • At this point, from the Digital Cores, we get all of the items information required, such as prices and so on, that are then used to carry out the dialogue on Facebook Messenger.

If you are interested to dig even deeper, you can find the code for this proof-of-concept on GitHub.

Next steps

In this blog we have seen how to bring more and more intelligence into your solution by leveraging SAP Cloud Platform and SAP Leonardo for your SMB customers. We have also seen how this intelligence can work across products and customers’ segments using a loosely-coupled approach and by abstracting the underlying Digital Cores with an adaption layer.

In the next blog we will see how to move further in this area and get even more intelligence out of your solution data.

Stay tuned !

To report this post you need to login first.

Be the first to leave a comment

You must be Logged on to comment or reply to a post.

Leave a Reply