Skip to Content
Business Trends
Author's profile photo Kishore Kumar Vivekanandan

Infusing GenAI in BTP Cockpit

Disclaimer: SAP notes that posts about potential uses of generative AI and large language models are merely the individual poster’s ideas and opinions, and do not represent SAP’s official position or future development roadmap. SAP has no legal obligation or other commitment to pursue any course of business, or develop or release any functionality, mentioned in any post or related content on this website

In the realm of artificial intelligence (AI), Generative AI stands as a groundbreaking advancement that has captivated the world’s attention. Over the past few decades, AI has witnessed remarkable milestones, from the introduction of neural networks to the utilization of GPUs and Large Language Models (LLMs). OpenAI’s creation, ChatGPT, has played a pivotal role in democratizing the usage of Machine Learning, making AI discussions accessible to a wider audience. In the past, AI was perceived as an exclusive domain for scientists and researchers, but now, even non-experts engage in conversations about the transformative effects of AI.

Our Executive Board Members, Juergen Mueller and Thomas Saueressig had emphasized the fact that Generative AI is a disruptive technology which brings immense potential to add new value to our customers. Infusing AI into SAP solutions to achieve Business AI is becoming the key priority. While SAP has already established a strategic partnership with Microsoft to leverage Generative AI capabilities as a service via the Azure OpenAI service, lately SAP has also announced a multi-vendor strategy by partnering with niche players in the market like Aleph Alpha, Anthropic, and Cohere.

From BTP, we have worked on this revolutionary technology to bring in visionary use case in BTP cockpit. This effort has resonated very well with some of our customers and their excitement towards the future potential.

Let’s look at this visionary use case – BTP Agent. For the users of BTP, we already offer different flavours for administration and operation like UI based approach over BTP cockpit and terminal based approach over btp CLI. Now we are additionally exploring AI based interaction to make the platform experience even simpler and better. BTP users must be familiar with the below UIs or the following commands which lead to BTP set-up including creation of subaccount, adding new users, assigning different user permissions, and adding service entitlement, etc.

While the above UI fairly works nice for the administration, it can get tedious if the customer must do any of those steps 10x (for ex: onboarding multiple users to the platform). For such repetitive use cases, customer could use btp CLI, using which customers can create an automation script either directly or using Terraform, infrastructure as a code (IAC) script to perform automation tasks. We recently released the Terraform provider for SAP BTP as a means to simplify the platform administration and operation tasks, as a next step building on that efforts we are enabling AI capabilities to even simplify the platform experience.

Imagine, now the BTP cockpit is infused with GenAI capabilities, and users no longer need to navigate through multiple UIs or write any automation scripts but simply ask AI (BTP Agent) what they want in natural language (prompts) and get back the desired BTP set-up in a few seconds.

Some of the prompts could look like

  • Create a subaccount in Europe region.
  • Onboard my team with the following users with the relevant permissions
  • Can you recommend which BTP service I can use to leverage SAP’s Intelligent capabilities and is there a free-tier (or) a free plan offering on that service?
  • I would like to try out SAP Build Apps, prepare my BTP account for the same.

SAP is committed to building relevant, reliable, and responsible AI. To make this agent more reliable and relevant, we have augmented the Large Language Model (LLM) with SAP’s own knowledge base as embeddings following the Retrieval Augmented Generation (RAG) technique, which mitigates the general problem of LLM around hallucination and not having up-to-date information.

We also designed the agent keeping users in the center, which means that the response from the Agent is not applied automatically but available for user’s review, modifications and only with the action (confirmation) of the users, the set-up is applied.

As you can see in the above demo, BTP Agent is also capable of sharing recommendations about BTP services to the platform users. BTP users have full flexibility to either use BTP Agent as a fully autonomous Automation engine or as an Intelligent Assistant to get recommendations towards platform services. The backend of the agent is designed to be a pluggable interface to other systems like SAP Digital Assistant to provide a consistent user experience across broader SAP solutions and ecosystem. In a real productised scenario, SAP Digital Assistant could become the first class UX for such interactions and have it integrated to the BTP cockpit.

What you have seen here is only a demo of what is possible when we infuse GenAI capabilities into our platform and the direction in which we are exploring the GenAI space. Stay tuned to learn more.

#btpcockpit #btpcli #generativeai #sapbtp #businessai  

Assigned Tags

      You must be Logged on to comment or reply to a post.
      Author's profile photo Dario Fracassetti
      Dario Fracassetti

      Thanks for sharing these insights! Very nice blog, I can't wait to see it in action on my subaccounts, and why not, on other SAP Products

      Author's profile photo Morten Wittrock
      Morten Wittrock

      Very cool, Kishore. Looking forward to seeing more!

      Author's profile photo Ashish Jain
      Ashish Jain

      Good one, is this already available in BTP?

      Author's profile photo Kishore Kumar Vivekanandan
      Kishore Kumar Vivekanandan
      Blog Post Author

      @Ashish Jain: Not yet, this is a visionary showcase at this moment. Stay tuned to learn more on this regard.