How ERP Giants Are Building GenAI Into Their Products
December 9, 2024 Alex Woodie
We are still near the top of Gartner’s hype curve when it comes to generative AI. Larger companies that have discipline and money, and startups with time and ambition, are finding success with GenAI, but many organizations are still in the planning and roll-out phase. The same could be said for ERP vendors, who have started their GenAI journeys but are expected to ramp up adoption significantly in the years to come.
Since the dawn of the mainframe age, computers have automated the work that was previously done by people. Large armies of file clerks were no longer needed once records could be digitized. The automation trend accelerated during the ERP era, when large centralized suites that could coordinate activities across the organization became the norm. As we enter the GenAI era, we’re on the cusp of another round of automation.
What makes large language models (LLMs) so intriguing is their demonstrated capability to automate large swaths of work that’s done with language. Knowledge workers – or folks who push numbers and words around for a living (ahem) – are squarely in GenAI’s crosshairs. The simpler the language understanding and comprehension task, the easier it is to train a GenAI model to do some of the work. (Luckily for technology reporters, GenAI isn’t self-aware enough to write intelligently about itself – yet.)
The big ERP vendors aren’t always at the cutting edge of technology adoption. But the capabilities of LLMs are so compelling that every ERP vendor is at least talking about a GenAI strategy.
Here’s a short rundown of the GenAI products launched by the Big Three in the IBM i ERP space, Infor, Oracle, and SAP.
Infor
The largest provider of IBM i-based ERP software is going all in on GenAI, which it says will “unleash hyper productivity” among its customers. It’s pushing two main GenAI products: Infor Embedded Experience and Infor GenAI Assistant.
With Embedded Experiences Infor has built GenAI capabilities directly into ERP workflows for procurement, project management, operations, financials, sales, and HR functions in Infor CloudSuite ERP products. The GenAI software automatically writes on behalf of the user and provides them with summarized insights automatically, among other capabilities.
Embedded Experiences can be used to create a project executive summary, to view contract performance at a glance, or translate emails, purchase orders, or invoices to another language, Infor director of solution marketing Benton Li wrote in an October blog post.
GenAI Assistant, which is a newer product unveiled in October, is a separate offering designed to supercharge how Infor CloudSuite customers interact with data. Infor says that it allows customers query the entirety of the data stored in Infor ERP applications. It is essentially a text-to-SQL generator that converts natural language questions, such as “List all active contracts managed by me in the geographical area ‘IND,’” into SQL that can be executed by the database.
Infor’s GenAI products run on Amazon Bedrock, the AWS service for training and serving foundation models, and utilizes Infor OS as the integration point for ERP systems. While Infor’s IBM i products can integrate with Infor OS, it’s unclear whether Infor has done the work to integrate its GenAI products into those products yet.
Embedded Experiences became generally available in October and is supported across a wide swath of CloudSuite products, while the GenAI Assistant is still in limited availability for Infor LN, Infor PLM Discrete, Infor HCM, Infor WFM, and Infor FSM products.
Oracle
Big Red has taken a comprehensive approach to integrating both classic AI as well as GenAI across its products, including databases, data warehouses, its cloud offering, and its flagship Oracle Fusion Cloud Applications.
Oracle’s wide breadth of GenAI offerings includes embedded GenAI capabilities in Oracle Fusion Cloud Applications. Oracle lets customers train GenAI models against data they have stored in their Oracle systems to deliver insights and generate content for customers, like pitches and contracts. The company is also ahead of the curve when it comes to Agentic AI, as it offers benefits analyst agents, sales automation agents, and document IO agents in various Cloud Fusion products.
Do-it-yourselfers can tap into Oracle Cloud Infrastructure (OCI) services, such as OCI Data Science, where data scientists can build and train their own custom models in Python in a JupyterLab-based environment. It also offers shrink-wrapped GenAI products, such as OCI Speech, OCI Vision, and OCI Anomaly Detection, which customers can use as they see fit. Customers can also choose OCI Generative Agents that utilize retrieval-augmented generation (RAG) techniques to narrow the context of requests and lower the odds of hallucinated responses.
Oracle has also integrated vector search capabilities into Oracle Database 23ai, giving customers a better experience than what keyword-matching can provide. HeatWave GenAI is an in-database LLM that features a vector database and is designed to enable customers to have natural language conversations with their database; it’s available in OCI and other clouds (Azure and AWS). Finally, Oracle offers Autonomous Database Select AI, which features text-to-SQL translation, RAG, and semantic similarity search.
Oracle says IT industry analysts praise its approach. Gartner IT analyst and VP Ritu Jyoti said Oracle’s common architecture “greatly simplifies the process for organizations to deploy generative AI with their existing business operations.” Dave Vellante, the chief analyst for theCUBE Research, said Oracle’s “full stack approach” to GenAI “directly connects to customer business value.”
Oracle’s out-of-the-box GenAI products appear limited to working with its Oracle Fusion Cloud Applications. The company is trying to move its JD Edwards World installed base to other products, so they will definitely not be getting the GenAI treatment (although there’s technically no reason why an LLM couldn’t drive a 5250 interface). Even JD Edwards EnterpriseOne customers will need to move to Fusion Cloud to get GenAI capabilities; there’s no mention of GenAI in EnterpriseOne Release 25, which Oracle unveiled in October.
SAP
Like other ERP bigs, SAP has fully embraced GenAI across its applications as a way to bolster automation. It has launched a variety of point GenAI products for various applications under the name SAP Business AI, as well as a GenAI co-pilot called Joule.
Examples of SAP’s GenAI point solutions include Document Information Extraction, a capability unveiled in its cloud ERP solution that automates the processing of documents. It has also adopted GenAI to automate repetitive HR tasks in its Human Capital Management (HCM) system, such as writing job descriptions.
SAP’s Spend Management & Business Network product uses GenAI to automate to categorization of spending categories, while its CRM system uses GenAI to help sales team create more personalized interactions by automating access to customer preferences and purchase history.
Business Technology Platform (BTP), SAP’s its offering for data, analytics, and AI, is a foundation for many of SAP’s GenAI capabilities. Many of these product-specific capabilities are enabled through Joule, which provides a new user interface and experience for SAP customers. Joule also functions as an intelligent search engine, enabling SAP users to explore data across their SAP applications and get faster answers to questions.
SAP has also launched a Generative AI Hub on SAP AI Core as part of BTP. This component gives SAP customers access to LLMs and also helps to manage the security and privacy settings of GenAI systems. SAP is charging access to its GenAI capabilities through a metric dubbed the “AI Unit.”
The IBM Institute for Business Value released a study earlier this year dubbed “AI in ERP” that sought to characterize GenAI implementations in SAP environments. The report concluded that organizations adopting GenAI solutions on their SAP data are experiencing greater profitability than those who don’t.
RELATED STORIES
RPG Code Assist Is The Killer App For AI-Enhanced Power Systems
GenAI Interest ‘Exploding’ for Modernization on IBM i and Z, Kyndryl Says
