A Peek Under The Covers Of SAP SuccessFactors’ AI Strategy
This week I had an in-depth briefing with the SAP /SuccessFactors architecture team to dig into the company’s long term strategy for AI. I walked away very impressed, and starting to understand how AI does not “replace” an HCM platform, rather it “transforms it” in a very significant way.
There are many “surfaces” for AI to emerge in HCM. AI can assist with sourcing, recruiting, learning, succession, and most talent processes. AI can supercharge content development, employee communications, case management, and employee service. And as an analytics engine, AI can greatly enhance pay equity analysis, performance management distribution, leadership assessment, and more.
For a global HCM vendor like SuccessFactors, the opportunity space is large. And rather than look at an architectural replacement for core HR at this point, the company is focusing on pragmatic, high-value use cases that immediately make SuccessFactors more useful, productive, and valuable.
Right now the company has about ten major use case families for AI across SAP SuccessFactors modules, each of which has a multi-scenario roadmap for features. These include things like intelligent job descriptions, recommended learning and careers, talent mobility and talent marketplace, skills and capability inference and analysis, employee policies and communications, and uses cases in employee service, employee transactions, benefits, and core HR. (Intelligent Job Description and Interview Questions was already launched.)
As I looked at the entire roadmap (the details of which are still confidential), I was struck by how pragmatic and sophisticated SAP has become. Since the company has deep domain expertise in hundreds of talent processes in almost every industry, the team is clearly “applying AI” in every place it can. And rather than make a single architectural decision to use OpenAI, Microsoft Azure Services, IBM, or another tech stack, they’re finding the appropriate LLM for each use case, optimized for its purpose.
For example, one of the big use cases is language translation. (This is also a major feature of our platform Galileo). The team has hundreds of interfaces and conversations to translate, so they found a particular LLM optimized for this purpose. Yes, SAP is working with Microsoft on the Copilot, but they’re also building a copilot of their own called Joule.
Joule (a “Joule” is like a watt of energy), is a carefully designed Copilot for all the SAP applications. The team has already developed 30+ use cases (transactions or journeys) and many of these are launching in 1H 2024. It was developed with some support from IBM and it lets companies use Joule for goal management, HR policies, job changes, clock in/out, leave, and other transactions. Imagine developing, managing, and reviewing your goals with a chatbot: it’s coming soon from SAP.
SuccessFactors groups its AI projects into three types: Generative AI, Conversational AI, and Deep Learning AI. Amongst many features this means SuccessFactors customers will get various tools like writing assistants, applicant screening, career conversations, learning tutors, as well as talent intelligence features like career exploration, talent marketplace, and advanced learning recommendations. SuccessFactors is also leaning heavily into Talent Intelligence.
The way customers will see this technology is through “feature releases.” In other words, rather than try to market AI as a whole new platform, customers will see many advanced features coming in all areas of the application.
As the company builds out Joule (the central new capability in the system), the paradigm is broken into three types of workflows: Information Patterns (looking for and displaying information, like viewing your benefits or vacation balance), Navigational Patterns (navigating a user through a business process with cards and prompts, like creating or finding a career path), and Transactional Patterns (directly performing a transaction without going to the application module, like entering an expense report or opening up a new position).
Since Joule is a conversational application, it doesn’t just answer questions – it also takes the user through workflows to simplify the flow of work. SAP understands that this is a massive paradigm shift, moving from “filling out forms” to “telling the system what you want to do.” I discussed this with the head of HR Tech at a global tech company last week and he told me they are already seeing 30% reduction in call center inquiries from their use of an intelligent agent for employees. Donna Morris shared similar or even better results for the employee assistant at Walmart.
Joule works across all SAP products and it stores history as well. So as you do different things at work, the system remembers what you did last, making it easier and easier to use. If you want to give an employee a bonus, for example, you may want to look at an employee’s sales performance in CRM, put that data into the bonus documentation, and then send it off to approval. If you’re not an expert at using the CRM, Joule figures out what you need on your behalf.
The architecture under the covers is called SAP Business AI. SAP Business AI is a variety of LLM and AI services, capable of serving different functions to different parts of the system. This means that Joule is essentially an “open system,” which could access other transactional information as well. Imagine if you connected Joule to a company’s banking system: employees could use SuccessFactors to view or manage their financial accounts, see when their bonus hits their checkbook, etc. (This is not something SAP plans on doing; has not been announced, I’m just imagining many of the options.)
This open architecture is important. As companies build their own Copilots and AI agents like Galileo, they’re going to want to interface with, share data, and connect to SAP and SuccessFactors. This “open API” process is part of the Joule strategy, so in some ways Joule could play in this larger space.
Competition Is Coming
While no other HCM vendor has announced this level of depth, you can bet a lot more is coming. Workday introduced its high-level strategy a few months ago and Microsoft has already launched the Microsoft Dynamics 365 Copilot, integrated with both HCM and CRM products. And this is just the tip of the iceberg: every HR tech provider is adding Gen AI features as fast as they can.
But SAP, as the biggest provider of all, is clearly staying ahead. Their strategy is integrated, comprehensive, and focused on pragmatic needs.
Much To Come Ahead
When OpenAI hit the market last Fall we barely imagined where AI would go. Now, thanks to vendors like SAP, we will see AI features, conversational interfaces, and amazing productivity tools emerge throughout our daily lives.
It’s an exciting new world, and I’m thrilled to see SAP setting the pace.