SAP embeds Joule in entire enterprise portfolio, plans integration to other AIs

SAP has announced additional integrations for Joule, the natural language generative artificial intelligence copilot that it added to its Business AI offerings last year.

Joule, initially integrated into SAP SuccessFactors and SAP Start, and later added to other products, will become part of SAP Ariba, SAP BTP Cockpit, and SAP Analytics Cloud in the second half of this year, the company announced at its Sapphire user conference in Orlando, Florida, this week. It will also be added to the company’s supply chain management solutions; they join SAP S/4HANA Cloud public and private editions, SAP Customer Data Platform, and others that have received the functionality since launch.

SAP also revealed additional capabilities for Joule. In this round of updates, Joule is receiving document grounding capabilities that enable it to draw from documents stored in SAP and third-party repositories when formulating answers to queries about organization-specific topics such as HR policies. And it will be able to respond to queries in additional languages: German, French, Spanish, and Portuguese.

AI integrations

But Joule isn’t the only AI in the SAP house. At Sapphire, the company announced that it will use Meta’s Llama 3 to generate scripts in SAP Analytics Cloud, and has expanded its Google partnership to integrate Joule and SAP Integrated Business Planning for Supply Chain with Google’s Gemini AI assistant and Google Cloud Cortex Framework’s Data Foundation based on BigQuery to, it said, “increase forecast accuracy and allow supply chain business users to react faster to demand signals.”

Nvidia and SAP also announced that Joule will receive new capabilities through Nvidia’s AI Enterprise software, and SAP will integrate Nvidia Omniverse Cloud APIs into its Intelligent Product Recommendation solution as well, so customers can use digital twins to visualize recommended products.

Another integration is arriving in the generative AI hub capability in SAP AI Core via three large language models (LLMs) from Paris-based Mistral AI.

In addition, SAP is creating a bidirectional integration between Joule and Microsoft’s Copilot for Microsoft 365, enabling Microsoft Copilot users to leverage Joule’s capabilities and vice-versa. The companies will begin integrating the two copilots later this year, SAP said, to combine enterprise data from SAP with contextual knowledge from Microsoft 365 applications.

“That is really neat for a couple of reasons,” said Ritu Jyoti, group VP, worldwide AI and automation research at IDC. “If a customer is spending a significant amount of time in the Microsoft Office products, and from there, if you could integrate to the SAP ERP or SAP SCM, I think it’s really going to be neat. So it will be a core integration of the copilots; no copilot is going to exist in a silo. And Microsoft is heavily invested into the whole Copilot Studio and the stack, and the coexistence and cointegration is something that the customers are looking for, with the understanding that they’re going to pick the best of the breed for their tasks.”

Applying general-purpose AI to specific business problems

These integrations are important because some models are better at certain tasks, SAP’s head of AI marketing Andreas Welsch said in an interview. “The goal is really that we leverage the best general purpose AI technology in the market and contextualize it in the context of business processes to solve a very specific and defined business problem that we can measure. That’s where we see the value.”

He said that SAP benchmarks models for performance on individual tasks, as well as for bias.

“That allows us to ensure that we deliver the best quality product with the best quality outputs to our customers, irrespective of what is the large language model underneath. At the end of the day, business customers care about business outcomes,” he said.

New use cases

Welsch said that SAP has already shipped more than 50 new capabilities for Joule since January and plans to deliver more than 100 before year-end. The company enumerated some of the latest additions to its Business AI at Sapphire. Here’s a selection:

• Generative AI capabilities in SAP Advanced Financial Closing will automate error explanations as well as creating personalized video data stories for financial controllers to highlight key metrics in project data.
• SAP Billing and Revenue Innovation Management users can analyze customer behavior and correspondence to predict whether they will pay on time.
• AI-generated reports in SAP SuccessFactors will provide managers with insights to use in their compensation discussions.
• A new generative AI writing assistant in the SAP SuccessFactors HCM Suite will help employees who need to draft documents such as performance goals, feedback reports, and development goals.
• A generative AI feature in SAP Fieldglass will help those hiring external workers create job postings based on job title, description, and required qualifications. They can also use it to translate those postings into other languages.
• SAP Sales Cloud now offers predictive forecasting to predict the likelihood of deals closing, as well as recommending who within a prospect’s organization can help close the deal, and what product combinations that the prospect is likely to buy. It also generates account summaries and lead boosters.
• The CX AI Toolkit will include a series of new features, such as an AI shopping assistant to let customers ask questions in natural language and receive replies on sites using SAP Commerce Cloud.
• Data from SAP Commerce Cloud, SAP Sales Cloud, and SAP Service cloud will be used to let administrators create customized AI tools for business users.

Analyst perspective

IDC’s Jyoti liked what she saw.

“I know the world is going gaga over Joule and Copilots and gen AI. I want to stress the fact that SAP has been focused on the traditional AI embedding across their entire suite of products for a long, long time,” she said. “With AI and gen AI, specifically, they’ve been very, very agile, they’ve made some strategic investments, they created a dedicated SAP Business AI organization.”

At the same time, she noted, it is impossible for one company to do it all themselves; there are different parts of the stack that need to come together, so, she said, “They have done a very, very nice job in establishing partnerships.”

“Overall,” she said, “my impression of SAP, their agility, their due diligence, their focus on relevance and business processes and data is resonating very well with the end users, and users are happy with it.”