SAP Spend Connect – See the bigger picture

SAP held its annual Spend Connect conference (previously known as SAP Ariba Live) in Vienna, Austria from October 9 to 11, where they graciously hosted the Spend Matters team to get the latest on product announcements and catch up with clients and SAP executives (you can catch the “best of” follow-up here). The event illuminated both the firm’s strategic intents and tactical efforts (e.g., GenAI). We’ll cover the announcements made at the event and publish some follow-on insights to address “under the hood” progress on integrating the firm’s app/platform portfolio and integration with customer tech stacks and the broader tech partner ecosystem (e.g., EcoVadis, Fairmarkit, Scoutbee, ZIP and others).

Spend Connect’s themes were generally focused on driving more value into SAP’s app stack and business network(s) individually. But collectively they focused on the UX (CEO Christian Klein said that the UX should “look like a consistent solution whether it’s S/4HANA, Ariba, Fieldglass or Concur”) integration (which is a story unto itself), ESG, business network synergy and of course, Generative AI.

Here are some specific highlights:

Rebranding and re-tooling

SAP has rebranded and broadened ‘SAP Ariba’ to now include all of SAP’s core procurement products (i.e., SAP Ariba and SAP S/4HANA) such that the SAP S/4HANA procurement apps shall now be part of the ‘SAP Ariba’ brand and product family.

This means that a CPO who wants to manage all spend will need not just these core SAP Ariba procurement apps within the broader Intelligent Spend [Management] and Business Network (ISBN) stack, and can also leverage the capabilities of SAP Business Network to go beyond their four walls and collaborate with suppliers (which includes all connectivity/integration and app-centric aspects of the network: core B2B, supply chain collaboration, logistics and asset management) and manage services and contingent labor spend within SAP Fieldglass. Travel and Expense management leveraging SAP Concur is an easy plug in for travel spend management and all spend flowing through expense reporting.

This announcement isn’t just branding, though, because SAP Ariba Buying, base edition and SAP Ariba Central Invoice Management solutions are now embedded in SAP S/4HANA Cloud Public Edition that uses the expanded S/4HANA data model (based on the harmonized SAP One Domain Model) rather than the one from legacy SAP Ariba. This could also be an interesting starting point for going downmarket (beyond SAP Ariba SNAP) via ‘GROW with SAP’ and ‘RISE with SAP’ solutions that also have SAP S/4HANA Cloud ERP Public Edition as their ERP core.

Similarly, the SAP Business Network Supplier Portal solution is also now embedded within SAP S/4HANA RISE Premium Plus Edition, providing access to key Network capabilities for PO and invoice automation.

AI, GenAI and Joule

SAP made some announcements about SAP Joule, its GenAI-powered copilot solution. As they stated: “SAP will also embed Joule, its new generative AI copilot, throughout its cloud enterprise portfolio, with availability in SAP spend management solutions planned for 2024. SAP is also developing additional use cases in collaboration with customers with availability planned for the first half of 2024.”

Understandably, there was a lot of excitement and curiosity surrounding Joule (and broader GenAI), but there is also some healthy skepticism regarding SAP’s announcements in AI given its mixed track record with AI that includes SAP Leonardo (discontinued with a few pieces migrated to BTP), SAP Conversational AI (planned for sunset/support) and an SAP Ariba Extension Studio chatbot (for approval management). We do think much of this should improve because of the broader GenAI strategy of working with the mega hyperscalers (whereas other app vendors generally pick a single hyperscale ecosystem) and with SAP’s more modular approach to working with the AI provider ecosystem (e.g., IBM and SapienceS2P were demonstrating their chatbots at the conference).

Joule is slated for release in early 2024 in terms of integration to SAP S/4HANA Cloud Public Edition, and later in 2024 for SAP Ariba solutions and the SAP Business Technology Platform (BTP). SAP BTP already has a tech stack for developing and deploying AI tools: SAP AI Core, SAP AI Launchpad and the resultant SAP AI Business Services that enhance apps with AI models pre-trained on business-relevant data. And regarding that data, SAP’s head of Product Engineering Thomas Saueressig recently said … “we have more than 20,000 customers who gave us consent that we use that data for showing in an anonymized fashion, that we use that data to train models — and that’s part of our contract. That’s the reason we can build up this huge SAP foundational model — most probably 200,000 database tables.” In the Joule press release, SAP CEO Christian Klein also commented that “Joule draws on SAP’s unique position at the nexus of business and technology and builds on our relevant, reliable, responsible approach to Business AI,” and that “Joule will know what you mean, not just what you say.”

There is a lot to unpack in those statements, but more tactically, SAP is planning to release 12 identified use cases for GenAI (beyond the Joule co-pilot) across the spend management landscape. Muhammad Alam, the President & Chief Product Officer of ISBN, outlined them below:

For example, one SAP Ariba product area specifically targeted for GenAI is the newly released SAP Ariba Category Management solution. At the end of 2023, the solution is planned to have GenAI functionality that will use large language models (LLM) prompt engineering in concert with ChatGPT to pre-populate market dynamics such as Porter Five Forces analysis within the CatMan module. While very eye-catching, the results seemed generic, but we can certainly see how the ‘outside-in’ content from the evolving LLM landscape can mesh with third-party partner content (e.g., Beroe for supply market intelligence) that’s woven into the solutions to enhance better decision making. This same outside-in approach is also apparent in other source-to-contract areas such as:

• A Spend Control Tower solution in Q2 2024 that expands the SAP Analytics Cloud (SAC) procurement analytics solution (which itself has improvements for spend classification, opportunity analytics and embedded process analytics) to now include data from SAP Ariba, SAP Fieldglass and SAP Concur. All are built to be integrated with the new SAP Datasphere platform (part of SAP BTP) that will offer better native supplier data enrichment, modular pre-integrated dashboards and spend benchmarking (which we’ll be diving into later). We were also happy to learn about the network-intelligence powered opportunity analytics from SAP’s incubator venture Source Agent that is slated for inclusion into the base edition of the planned Spend Control Tower solution.
• Supplier discovery beyond the SAP Business Network using Scoutbee integrated into the SAP Ariba Sourcing UX rather than just a provisioned partner application tab (available Q1 2024).
• Supplier evaluation within sourcing and supplier management applications that ingest various ESG and risk feeds from internal S2P data, adjacent SAP apps (e.g., CO2 estimates from SAP’s Sustainability Footprint Management solution), partner data (e.g., EcoVadis), the SAP Business Network and/or GenAI.
• The evolution of SAP Ariba Supplier Lifecycle and Performance continues with supplier 360 capabilities now released for early adopters. New enhancements broaden functionality beyond the solution scope (and underlying platform) and include GenAI for assistance with supplier scoring. SAP Ariba Supplier Management capabilities expanding use cases providing supplier and risk information across the source-to-pay process will continue through 2024.

Better guidance and support within P2P

‘Guided Buying’ is a concept introduced in 2005 (actually, by one of the authors of this piece, Pierre Mitchell) many years before SAP launched a capability by the same name that helped with basic user guidance, search and ‘smart forms’ to go beyond free-text requisitions. Interestingly though, a consumerized end-user experience is perfectly suited to AI-enabled personalization (e.g., based on role, predicted intent, historical behavior, and so on) and for appropriate chat-enabled UX features that work in concert with more purposeful intake management functionality to better predict/understand intent and then triage and prescribe an orchestrated workflow into different apps, process paths and data access.

SAP’s advancements in guided buying capability were generally a continuation of previous efforts and SAP currently partners with Zip for intake management. The 2024 roadmap does however have some interesting focus beyond GenAI to also include spend channel optimization, exception handling (with feedback/learning) and other areas.

For SAP Fieldglass, the ongoing integration work continues regarding process integration, business network integration, analytics and of course GenAI. For GenAI, there was some low-hanging fruit related to auto-generating job descriptions based on skill maps (and there’s a lot of cool stuff happening in this area over in the SuccessFactors world) and auto-translating job descriptions to local languages. The most interesting area however is the ability to auto-create SOW descriptions (basically a lite version of what a best-of-breed vendor like Globality, Zivio, and a few others can do).

SAP Concur has been making numerous changes related to UX (especially in travel planning where GenAI chatbot can be helpful) and across-the-board enhancements regarding platform, integration (with budgets/policies on front-end and global payments on the backend) and other areas.

SAP Business Network is still a work-in-progress … and a real potential differentiator

On the ‘network’ front, the existing solutions[HT2] (core B2B, direct-focused supply chain collaboration, logistics and asset management) are making functionality (e.g., unified supplier profiles), UX (improved onboarding process) and integration enhancements on both the buy-side and sell-side to enhance the individual product areas that they belong to. Enhancements include integrating with third parties to enable partners on the network to build richer profiles by pulling [verified] data from EcoVadis, CatenaX and World Business Council for Sustainable Development (WBCSD) for supply-level sustainability data as part of SAP’s sustainability data exchange [IE3] available since September. See below …

One use case that is very real and present in the procurement technology space is related to how vendors/suppliers respond to RFx documents. SAP is obviously working with buyers to auto-generate RFx documents, but also with suppliers on generating responses (rather like what responsive.io [the new name of RFPIO, a provider of response management software that acquired RFP360 in 2021] and Loopio do). Although there will likely be an eventual ‘battle bots’ arms race of suppliers and buyers (some of the advanced autonomous sourcing tools that we cover already do some of this), it does provide an interesting opportunity for SAP to build an open ecosystem of data formats (similar to some of the win-win best practice contract formats/tools in the CLM space) to reduce friction with B2B interchange beyond POs/invoices/etc.

We’re also excited about the move to improve the openness of the network such as the 1-to-many supplier-managed public catalogs that can be more easily discovered and also linked to other ecosystem data (e.g., carbon footprint data) from SAP-provided content and partner content alike. And we continue to be strong advocates for more native network-derived outside-in intelligence that creates strong value (e.g., baking in benchmarking opportunity analytics more deeply and formally into the analytics, sourcing and category management applications). Finally, we would still like to see a common knowledge graph ‘backplane’ to model n-tier supply chains that ties the various networks (and applications) together better. Of course, this isn’t easy — it’s much harder to ‘run simple’ than just automate complexity.

Looking forward

In terms of GenAI strategy, SAP, like others, is pursuing a pragmatic approach to getting POCs and use cases built out with customers to augment current solutions while simultaneously planning out future strategy. The firm is partnering with the three big hyperscalers to integrate to their LLMs through appropriate APIs and associated deployment models (ChatGPT Enterprise as an example for ChatGPT) and guard rails, but SAP has also taken equity positions with Aleph Alpha, Cohere and Anthropic in order to better control its destiny around the creation of the LLMs themselves and how they are tuned and integrated into the SAP-specific knowledge models/graphs (e.g., using RAG). This is key to make the LLMs more efficient and effective to understand user persona, intent and domain area that then drives the optimal use of those LLMs to augment the apps and the humans.

Let’s face it, this is an extremely complex challenge! SAP needs to co-exist and coordinate with hyperscale frenemies and their AI stacks (e.g., Microsoft Azure and its embedding of OpenAI’s ChatGPT tech stack), best of breed platforms/tools, customer-specific environments, etc., while also using BTP (see here for more on BTP within the broader context of digital platform strategies/architecture) to inject AI into the other platform elements …

… and ultimately into the applications and data/knowledge models of those 200K database tables. Then it has to make sure everything works, and is secure, and is usable, and not unaffordable (which can be a challenge with the compute costs of many LLMs that need to be paid by someone). That’s a big job to tackle, but SAP does have over 5,000 engineers in the spend management space (not even including BTP) working on all this. And luckily, GenAI also helps to ‘speak’/generate the language of code, including ABAP!

We’ll be writing follow-up deeper-dive information on SAP and we are also about to release an upcoming AI series for our Spend Matters Insider readers that gets into all these issues in gory detail. For non-members, please subscribe to our newsletter to stay up to date. And as always, if you have any questions on all this please don’t hesitate to reach out and talk to us humans … we’re not bots (yet).