For many years, businesses have been striving to create automated AI solution for SAP that can be implemented with little or no human interaction. Automation of business processes enables organizations to complete key tasks and processes while employing fewer human resources. And with recent technological advancements, more and more businesses are shifting towards low-cost automated processes that can replace the traditional labor force.
From global giants like Amazon to smaller SaaS startups, organizations of all sizes are adopting business automation and leveraging its benefits such as low recurring cost, high accuracy, fewer errors, and the ability to perform 24/7.
We’re already witnessing a sharp rise in the popularity of Business Process Automation, Generative AI capabilities, Machine Learning and Robotic Process Automation which enables AI models to do the same tasks just like humans.
Such innovative automated processes have deviated from traditional approaches and helped organizations increase throughput and ROI.
An Integration Platform powered by AI Core is one such solution that many businesses need as it seamlessly integrates multiple business applications so that organizations can have more advanced and up-to-date data between all applications. And now, even integration platforms are leveraging AI algorithms and technology to significantly reduce human effort and increase business efficiency.
The software industry has recently seen rapid commercialization of Artificial Intelligence, which is one reason why businesses are employing huge efforts to incorporate AI in their regular business processes.
AI solution providers are also offering APIs and generative AI capabilities that developers can connect with their own systems. These APIs come with a general purpose and allow developers to create the specific responsible AI solutions that they are looking for.
In this post, we are leveraging the power of generative AI capabilities to build an automated solution that can be readily implemented and is useful for the SAP business technology platform, SAP solutions and any eCommerce business, helping them to scale without compromising data.
Table of Contents
The Problem – Generative AI Solution for SAP S4hana and Shopify
In retail eCommerce, one of the essential but time-consuming activities is to correctly define and maintain products in the storefront. For any industry, the products are the core of the business, and the popularity of any product depends on how well it is maintained in the ERP such as SAP and described in the eCommerce listing pages. We want to ensure that the listing provides enough info about the product using AI and SAP Data Intelligence, and also how it is written in an SEO-friendly way.
The product description within the SAP applications should also be proofread, free from grammatical errors, and highlight the key features and benefits of the product. Writing the perfect description by incorporating all these factors requires sufficient knowledge about the product and human intelligence to craft the right words. These things take a lot of time, and that can delay the overall chain of operations.
In fact, for bigger organizations that deal with a large number of SKUs, delays in creating bulk product descriptions can further increase the time to market. The cost of delay is huge and businesses end up spending more money just to gain a competitive advantage over others.
To highlight this problem and develop a suitable AI solution, we have created a hypothetical eCommerce business called ‘Dressify’ which is a fashion brand running a storefront through Shopify. In the backend, all product SKUs have been entered, but in this case, the store manager cannot manually ensure that the data is perfect and complete.
Some SKUs might not have descriptions or images, some may have grammatical issues, and others may not be properly optimized for search engines and target keywords. Unless this backend data is proof-checked within the SAP applications, the business owner will be reluctant to publish the product pages online even though their Shopify store is ready and correctly working.
So, the problem is to find an efficient and automatic process for validating the product data, which is securely stored in SAP S/4HANA on-premise.
The Solution – eCommerce Product Metadata Generation and Validation for SAP S4 HANA and Shopify
APPSeCONNECT, being an intelligent iPaaS solution, aims to address this issue by integrating the SAP business suite with Shopify, so that data can be seamlessly transferred from the SAP ERP backend to the Shopify online stores and be published automatically.
While doing so, we will leverage AI and use generative AI to proofread and check the business data for every product, and perform the following actions:
- Fix grammatical errors, spelling mistakes, etc.
- If a product description is not present, the AI will write an appropriate description.
- The AI will also check for images and if any product is lacking images, it will be generated by the AI and uploaded with the product.
The Pre-requisites
Before we start configuring the integration using our low-code platform, let us take a look at the prerequisites for developing this automated process:
1. Access to an on-premise S/4HANA environment.
2. Administrative access to configure IDOC URLs
3. A Shopify account
4. An account with OpenAI
5. An APPSeCONNECT account
Once you have managed these aspects, you can start implementing the process.
Configuring the platform
We will now guide you through a step-by-step process of configuring the platform, which includes two parts.
Part 1: Integrating SAP with Shopify
- Navigate to https://portal.appseconnect.com and log in using your credentials.
- After logging in, go to “Design” from the menu and click on “New”.
- Once you open the designer, you will be asked to name the ProcessFlow. In this example, we will be naming it “Creating Product Metadata Using ChatGPT”, but you can name it as you see fit, as long as it aptly describes the flow.
- Choose “Event” from the section before you click on “Proceed”. Choosing “Event” will provide you with a URL that can be configured into your own application for Webhook, which is what we will use to configure SAP S/4HANA.
- Now choose the “SAP ECC and All in One” App from the list. This app will allow you to connect S/4HANA on-premise applications and fetch data from it. We use “MATMAS” in Schema and “Get Products IDoc Data” from Action.
- Now create “Mapper” from the SAP node and connect “Shopify”. For Shopify, we use “Product” as a Schema and “Adding Items to Shopify” as an Action.
- Configure the mapping of data by choosing appropriate data from the Source Schema to the Target Schema in our mapper.
- Now deploy the ProcessFlow using any one of two methods. You can either choose “Hosted” in which case, both your applications must be accessible online. Or you can choose “On Premise” where you need to install a small tool in your local server to sync data.
In our case, we have chosen “On Premise”. Here, the name of the computer where we have installed the APPSeCONNECT Agent is shown. Click “Next” to continue. - Now you need to go to your on-premise tool to configure “Credentials”. Provide Credentials for SAP S/4HANA and Shopify.
- Finally, once your deployment is complete, go to Deploy from the menu and get the ProcessFlow URL. This URL must be configured in SAP so that you can send the IDocs (Intermediate Documents) from SAP directly to your server.
The steps to configure an IDoc is mentioned in this post. - Now execute the process by sending an IDoc from SAP, and you will see that it is posted directly into your Shopify Store.
However, this is just one part of our solution, where we connect SAP with Shopify. For the second part, we still need to connect OpenAI APIs to automate the process of checking and adding product descriptions and images.
Part 2 – Configuring OpenAI APIs
- To create an OpenAI-powered solution, let us first create an App. Open the APPSeCONNECT portal again and start creating a Custom App. Choose Manage – > Custom Apps from the menu. The App registration wizard will start.
- Upload the OpenAI logo to the app. We strictly follow size limits, so if the logo doesn’t have the appropriate size, please crop the image using ImageResizer and then upload it to the screen.
- Describe your OpenAI App by providing the App Name and App Description. Once done, click on “Save and Proceed”.
- Now configure the API authentication details. OpenAI provides Bearer Token Authentication. Choose it from the list. Provide the API Path as https://api.openai.com/v1/. The Token can be found on the OpenAI portal.To get OpenAI Token, go to https://openai.com. Login to your account (in case you don’t already have an account, create one). Choose API from the home screen. Now click on the User link and go to the View API Keys link.
You can create an API key from the screen and add it in the Token section of authentication.
- Finally, add APIs for OpenAI. The ones that we are going to use are:
You can go on adding the data to the API details.
Once you are done adding the API – Request and Response structures, you are done with building the Custom App.
- Finally, go to the ProcessFlow again and start adding the OpenAI app.
- Click on the + sign just beside SAP ECC and All in One node and create another mapper.
- Choose the OpenAI app from the list.
- Inside the mapper, we use the MAXTX field coming from SAP ECC IDoc to generate the documentation. We configure the whole body which needs to be sent to OpenAI for output.
You can see the content, as we have passed our customized message which will generate the output in HTML.
- In the OpenAI node, we create a Variable for storing the description that we want to generate. We use a variable desc to gather content from the response.
- Similarly, we added two more mappers from the SAP ECC node to generate images and moderate the text. We use the imageurls and mod variables to store the images and modified data.
- Finally, we configure the main mapper to add the content stored in these variables.
- We also need to change the order of branch execution. Open the Link Ordering section and drag and drop the branches to change the execution order.
Here, you can see that the OpenAI mappers are executed before the actual execution in Shopify. If you don’t see the same execution order in your case, drag and drop the mapper nodes to configure the order properly.
- Finally, deploy the ProcessFlow again.
The final ProcessFlow should look like the image above.
Understanding the Available Options in OpenAI
OpenAI provides many API options. We are mostly aware of the ‘chat completion’ API, and we saw a few additional APIs that were used in our automated solution. Other than those, here are a few other APIs which can also be useful in many cases:
- Text Edits – Used to fix spelling mistakes, grammatical errors, etc.
- Create Transcription – Used to convert Voice to Text.
- Chat Classifications – Used to categorize data into different groups.
- Create Embeddings – Creating Vector data to pass into other APIs as context.
Executing the Integration
Finally, when you are confident that the integration is ready, you can start executing the solution.
To execute, we go to the SAP S/4HANA environment and follow these steps.
- Open the “Send Material” form by typing the command “/obd10”. The ‘/o’ part is to open a form, and ‘bd10’ is the form number for Material Master.
In the above screen, we are trying to send the Material created inside SAP to APPSeCONNECT. - If you click on the “Display Material” screen, you will open the screen with the data present in SAP.
Here, you can see the material description is not written properly. - Now send this data as IDoc. After the data is sent, you can open the Snapshot section to view the data sent to Shopify.
As you can see, the actual data that is sent to Shopify is different than the original data we had in SAP. - If you check this Product in Shopify, you will see that it is published with all the details.
Hence, you can see that even though the product had only the name in the description, our AI-powered process has automatically populated the whole data along with images in Shopify and made the product immediately available to be listed for purchase.
Conclusion
In today’s fast-paced eCommerce landscape, the most successful businesses are those smart enough to adopt new technologies quickly and efficiently. The agility of adopting novel technologies is a crucial factor that gives businesses a competitive edge as well as financial returns.
With APPSeCONNECT, we use the latest products and services to not only enhance the seamless integration between enterprise applications and enable complete Business Process Automation to help business owners save money by reducing human efforts.
The AI solution we have developed and explained in this post will be incredibly useful for eCommerce businesses to automate the generation and validation of product metadata. This will save significant time and allow retailers and sellers to launch their products in the market faster.
To learn more about eCommerce integration and AI solutions, check out the resources below.
- AI in eCommerce: 11 Use Cases You Should Know
- IPaaS meets AI, a Case Study
- OpenAI Integration for SAP and Shopify – Webinar Recap
APPSeCONNECT is an intelligent, low-code, user-friendly integration platform that seamlessly connects enterprise applications like SAP with eCommerce platforms like Shopify to help businesses streamline and automate core operations.
We offer readymade integration packages as well as custom automated solutions if you have specific processes and requirements. Contact us to know more or try our demo version to explore the benefits of API integration and business process automation.
Looking to get started with SAP and eCommerce integration to automate the business operations in your organization? Seamlessly integrate and automate your ERP with other applications under a single, intelligent, and secure Business Process Automation platform through APPSeCONNECT’s robust integration packages and achieve exponential business growth.
