People working with #SAP #SuccessFactors the past five years will have known that SAP has had some attempts to introduce conversational navigation – also known as a digital assistant. The time has finally come where SAP has introduced Joule. Now, in the upcoming 2H 2023 release, SAP will finally be releasing a digital assistant with the energy laden name Joule. This blog goes close and introduces my first impressions of Joule in the context of SuccessFactors and provides a small historical perspective.
Read on to understand why!
Everyone who has tried to work with a more or less helpless chatbot for support on a website knows that nothing can be more annoying than a digital companion that has no empathy, no contextual knowledge, does not seem to learn or improve and who goes around in circles asking the same question again, forcing your blood pressure up where it should not be. And the first think you want to write is “I want to speak with a human being”.
In HR, digital assistants is a computer program that help users by answering questions and performing basic tasks within many HR domains such as Recruitment, Leave management, Answering FAQs, Employee onboarding, Reimbursement processing, Payroll management, Employee analytics, Pre-screening applicants, Hiring and Scheduling meetings.
Digital assistants can answer questions in real time and in a conversational way. They can provide quick, accurate answers about topics. SAP’s vision with Joule is to offer conversational interactions between humans and computers to simplify access to information and automate business processes - improving both employee and customer satisfaction. Joule supports three conversational patterns:
My colleagues at Effective People and I have been part of the Joule beta programme, known under the name SuccessFactors Digital Assistant. The program has run for a few months, providing us insights into how the solution is configured and integrated with SuccessFactors as well as how it functions for end users. In the next section I will take a look at each of these use cases.
The first use case I will explore is navigational support. If a SuccessFactors user is a frequent user of a certain function, the need for navigational support is typically not relevant. But even frequent users sometimes use less-used features, and for everyone else even the most userfriendly and well-designed system may sometimes be overwhelming. For example, Susan can be a line manager who has given lot’s of feedback and guidance to her direct reports with continuance performance feedback and development plans, but if it’s the first time in a long time she needs to perform a certain task it might be hard to find out where to start. Of course the system navigation helps or the already existing Action Search, but asking a friendly chatbot is an alternative. Navigational use cases can be seen as a simple version of a transactional use case (see next section). In below example, Susan is unsure where to find a specific function in the system and thus asks Joule for help, in order to fond out where to start a certain process.
Although just a few steps are shown in the above use case, you can see that Joule knows that the person has a few direct reports already. Joule is forgiving and understands the request even though a key word has a typo mistake (“requsition”) in it and suggests next steps. Following selection of the person the requisition can be initially created and sent for approval.
The second case I will explore is the transactional use case. This is where a person that is initiating a transaction is unsure of how to get started or to support in completing a specific task (transactional is a more in-app specific version of navigational use case). For example John might be used to creating job requisitions as a recruiter, but since he was recently promoted to first time manager, he is less sure how to provide quality ad hoc feedback to his direct employees. Here a transactional use case comes in.
Although limited to certain use cases right now, SAP has promised a longer list of transactions that will be supported in the upcoming releases. “I need to give feedback to John about the failed project, can you help me”, “I need to reclassify 15 positions”, “Can you route all performance forms to calibration for me, if they pass validation”. This will be interesting to see in the upcoming release news (check out various announcements from SAP on the What's new viewer or join the Effective People SuccessFactors Customer Community.
Trying out the informational use case
The third case I will explore is the informational use case. This is a more basic system interaction where a person want’s to retrieve information stored in the system, but is unsure where to find it. This is a more simple and entry level use case, intended for infrequent users who “just need to find something but are unsure where to find it”.
Although quite basic right now, informational support can be useful for some users. Future direction spells many areas where informational use case can increase knowledge and improve decision making. Questions such as “do I have any pending tasks to complete before I go on vacation”, “what is the diversity profile of my team”, “Susan just requested vacation first two weeks of August, are there other team members who have planned vacation those weeks”.
This question is harder to answer, because we are still in the early days. There is no question that the underlying technical architecture behind Joule supports machine learning and also can be fed with information and learn from user patterns. But the current solution is a closed box that customers do not have access to see the inside of. So for example if you want to build your own vocabulary or add your own navigational pattern this is not yet possible.
Will Joule learn from end user interactions to improve future user interactions? This question is also hard to answer, because we are in an early stage of the release. My hope it that the answer will be “yes”, and I know that usage insights and analytics capabilities is on the road map for 2024 so we hope to see this of course.
Will Joule be able to provide proactive guidance and for example tell the logged in user Susan “Hi Susan, you have not provided feedback to John for 6 weeks – would you like help to give him some constructive feedback?”. This is a feature for the future, but also something that needs to be done carefully in order not to give users “AI fatigue”, just like many are suffering from “survey fatigue”.
SuccessFactors has had a history of including various end user support features in the system for many years.
For example, SuccessFactors has had the Action Search feature for quite many years. This allows a company to add their own words and vocabulary to the search engine, and when users type in these keywords the system proposes navigational options. This is not AI but pure text based algorithms. In SuccessFactors Recruiting there is a built in feature called Job Analyzer that will analyze job description texts used for the candidate portal for unwanted terminology. In Goal Management we have had the writing assistant and coaching advisor capabilities for more than 15 years (and still going strong).
There is no doubt for me that SAP will be investing heavily in Joule and other solutions for improving productivity, increasing user adopting, easing system usage and ultimately improving employee experience and customer satisfaction as a result. That is the vision and the promise that I will measure the success of Joule. The foundation is being built with the 2H 2023 release, but the real test will be to see what happens in 2024 where this will need to take off and support every module across the suite, including also administrative use cases for system administrators and configuration specialists.
End users get’s a new method of accessing system content and getting navigational guidance. Support and application management teams are also likely to benefit through fewer support related queries that will instead be handled by Joule. SAP – assuming they mine the data collected from user journeys – also benefits because they have better understanding of end user interactions and where processes break.
A digital HR assistant is not the same as a virtual HR assistant. A virtual HR assistant is a remote physical worker who helps the HR department of an organization with various tasks within recruitment, payroll management, background checks, screening tests and similar administrative activities. I think that companies offering virtual assistants are the categories that risk the biggest business impact of digital assistants.
The first initial plan was to build a conversational tool native for SuccessFactors, which started in 2018. After initial development was started, this program was eventually pulled back to create a tool based on SAP Business Technology Platform with the aim of supporting the entire SAP business suite. In early 2023 this initiative was also pulled back, and replaced with the current program that we are now seeing materialise as Joule. So it has been a long time coming, but since the decision was made (public) things have moved fast – and rightfully also should move fast.
At high school many years ago, I learnt that a joule (pronounced /ˈdʒuːl/, JOOL) is the unit of energy in the International System of Units. It is both equal to the amount of work done when a force of 1 newton displaces a mass through a distance of 1 meter in the direction of the force applied, and also the energy dissipated as heat when an electric current of one ampere passes through a resistance of one ohm for one second. Joule got it’s name after the famed English physicist James Prescott Joule who lived from 1818 to 1889. In the context of SAP Digital Assistant you can say that Joule will give the user an extra force and energy to get things done and to get it done fast ... I like the name! It also makes sense to give a tool - that mimics a person - a real name just like other large tech providers have done, for example Siri, Echo, Cortana, Nina, Viv, Jibo, Silvia, Bixby, Lucida, and Aido. So this strategy puts SAP Joule in good company with a catchy short name that works in multiple language zones (I know at least it can be spoken in at least English, German, French and Nordic languages without issues, although in France they may mix it up with another gentleman with last name Vernes).
I believe that Artificial Intelligence, Machine Learning and Digital Assistant are topics that most companies will be utilizing. I don’t know if we are at the dawn of a new technical revolution, but I am sure that my colleagues at Effective People and me will be working much more with AI in 2024 than we have done previously. We also see this materialized through topics that our clients raise with us: they include more often now topics such as “how can we improve productivity through automation / technology support / process alignment / end user support”. Joule comes in very handy here because it provides support for all of these areas.
SAP has provided a short infographic that summarises their mission and priorities of generative AI in different functions, including HR. Check out this link: https://www.sap.com/documents/2023/07/c814592c-7e7e-0010-bca6-c68f7e60039b.html
SAP has gone a long route to get to this point. I think it makes sense to utilise a proven technology from a partner company (IBM). Although technology is proven, and that SAP does a great job of protecting end user from even seeing the tech layer, I am afraid of the extra technology component and what it will mean in terms of complexity. I am also unclear on licensing terms and under what conditions Joule will be made available. I am super curious to see how customers can make Joule their own, and add company specific knowledge, intellectual property and capabilities. Because a company is better off with an assistant that knows the company that a range of assistants that knows parts of the company. This is to be proven, but with the speed that SAP moves I am not afraid of the future – rather contrary.
Thank you for reading this and I hope it was of value to you. If you would like to discuss particular elements of Joule, how to use AI in HR and SuccessFactors specifically, feel free to reach out!
Erik Ebert, SuccessFactors Solution Advisor and Tech Geek, Effective People
With over 25 years of experience in the tech industry, I am a Senior Director at Effective People, a leading provider of digital transformation solutions for workforce improvement. As a SAP SuccessFactors Confidant, I also advise on the SAP HR cloud leadership team on market development and innovation. My mission is to help organizations across industries improve their people operations and business outcomes through a consultative and collaborative approach, utilizing new technology for the benefit of humans. Feel free to connect with me on LinkedIn https://www.linkedin.com/in/ebert X https://twitter.com/ErikEbertDK as well as on SAP erik.ebert2
There is significant value in end user system interaction and adopting by using Joule – the new Digital Assistant from SAP
Read on to understand why!
What is a digital assistant in HR, anyway?
Everyone who has tried to work with a more or less helpless chatbot for support on a website knows that nothing can be more annoying than a digital companion that has no empathy, no contextual knowledge, does not seem to learn or improve and who goes around in circles asking the same question again, forcing your blood pressure up where it should not be. And the first think you want to write is “I want to speak with a human being”.
In HR, digital assistants is a computer program that help users by answering questions and performing basic tasks within many HR domains such as Recruitment, Leave management, Answering FAQs, Employee onboarding, Reimbursement processing, Payroll management, Employee analytics, Pre-screening applicants, Hiring and Scheduling meetings.
Digital assistants can answer questions in real time and in a conversational way. They can provide quick, accurate answers about topics. SAP’s vision with Joule is to offer conversational interactions between humans and computers to simplify access to information and automate business processes - improving both employee and customer satisfaction. Joule supports three conversational patterns:
- Helps users navigate to the functionality they are looking for
- Assists users in efficient completion of their tasks
- Helps users retrieve the information from existing documents
My colleagues at Effective People and I have been part of the Joule beta programme, known under the name SuccessFactors Digital Assistant. The program has run for a few months, providing us insights into how the solution is configured and integrated with SuccessFactors as well as how it functions for end users. In the next section I will take a look at each of these use cases.
Trying out a navigational use case
The first use case I will explore is navigational support. If a SuccessFactors user is a frequent user of a certain function, the need for navigational support is typically not relevant. But even frequent users sometimes use less-used features, and for everyone else even the most userfriendly and well-designed system may sometimes be overwhelming. For example, Susan can be a line manager who has given lot’s of feedback and guidance to her direct reports with continuance performance feedback and development plans, but if it’s the first time in a long time she needs to perform a certain task it might be hard to find out where to start. Of course the system navigation helps or the already existing Action Search, but asking a friendly chatbot is an alternative. Navigational use cases can be seen as a simple version of a transactional use case (see next section). In below example, Susan is unsure where to find a specific function in the system and thus asks Joule for help, in order to fond out where to start a certain process.
SuccessFactors Joule Use Case for Navigational Support (borderlining Transactional Support)
Although just a few steps are shown in the above use case, you can see that Joule knows that the person has a few direct reports already. Joule is forgiving and understands the request even though a key word has a typo mistake (“requsition”) in it and suggests next steps. Following selection of the person the requisition can be initially created and sent for approval.
Trying out a transactional use case
The second case I will explore is the transactional use case. This is where a person that is initiating a transaction is unsure of how to get started or to support in completing a specific task (transactional is a more in-app specific version of navigational use case). For example John might be used to creating job requisitions as a recruiter, but since he was recently promoted to first time manager, he is less sure how to provide quality ad hoc feedback to his direct employees. Here a transactional use case comes in.
SuccessFactors Joule Use Case for Transactional Support
Although limited to certain use cases right now, SAP has promised a longer list of transactions that will be supported in the upcoming releases. “I need to give feedback to John about the failed project, can you help me”, “I need to reclassify 15 positions”, “Can you route all performance forms to calibration for me, if they pass validation”. This will be interesting to see in the upcoming release news (check out various announcements from SAP on the What's new viewer or join the Effective People SuccessFactors Customer Community.
Trying out the informational use case
The third case I will explore is the informational use case. This is a more basic system interaction where a person want’s to retrieve information stored in the system, but is unsure where to find it. This is a more simple and entry level use case, intended for infrequent users who “just need to find something but are unsure where to find it”.
SuccessFactors Joule Informational Use Case
Although quite basic right now, informational support can be useful for some users. Future direction spells many areas where informational use case can increase knowledge and improve decision making. Questions such as “do I have any pending tasks to complete before I go on vacation”, “what is the diversity profile of my team”, “Susan just requested vacation first two weeks of August, are there other team members who have planned vacation those weeks”.
Is Joule intelligent or just a static algorithm
This question is harder to answer, because we are still in the early days. There is no question that the underlying technical architecture behind Joule supports machine learning and also can be fed with information and learn from user patterns. But the current solution is a closed box that customers do not have access to see the inside of. So for example if you want to build your own vocabulary or add your own navigational pattern this is not yet possible.
Will Joule learn from end user interactions to improve future user interactions? This question is also hard to answer, because we are in an early stage of the release. My hope it that the answer will be “yes”, and I know that usage insights and analytics capabilities is on the road map for 2024 so we hope to see this of course.
Will Joule be able to provide proactive guidance and for example tell the logged in user Susan “Hi Susan, you have not provided feedback to John for 6 weeks – would you like help to give him some constructive feedback?”. This is a feature for the future, but also something that needs to be done carefully in order not to give users “AI fatigue”, just like many are suffering from “survey fatigue”.
Other "AI" capabilities with SuccessFactors
SuccessFactors has had a history of including various end user support features in the system for many years.
For example, SuccessFactors has had the Action Search feature for quite many years. This allows a company to add their own words and vocabulary to the search engine, and when users type in these keywords the system proposes navigational options. This is not AI but pure text based algorithms. In SuccessFactors Recruiting there is a built in feature called Job Analyzer that will analyze job description texts used for the candidate portal for unwanted terminology. In Goal Management we have had the writing assistant and coaching advisor capabilities for more than 15 years (and still going strong).
What about the future
There is no doubt for me that SAP will be investing heavily in Joule and other solutions for improving productivity, increasing user adopting, easing system usage and ultimately improving employee experience and customer satisfaction as a result. That is the vision and the promise that I will measure the success of Joule. The foundation is being built with the 2H 2023 release, but the real test will be to see what happens in 2024 where this will need to take off and support every module across the suite, including also administrative use cases for system administrators and configuration specialists.
Who benefits from Joule
End users get’s a new method of accessing system content and getting navigational guidance. Support and application management teams are also likely to benefit through fewer support related queries that will instead be handled by Joule. SAP – assuming they mine the data collected from user journeys – also benefits because they have better understanding of end user interactions and where processes break.
Who should maybe be more worried
A digital HR assistant is not the same as a virtual HR assistant. A virtual HR assistant is a remote physical worker who helps the HR department of an organization with various tasks within recruitment, payroll management, background checks, screening tests and similar administrative activities. I think that companies offering virtual assistants are the categories that risk the biggest business impact of digital assistants.
Historical attempts from SAP
The first initial plan was to build a conversational tool native for SuccessFactors, which started in 2018. After initial development was started, this program was eventually pulled back to create a tool based on SAP Business Technology Platform with the aim of supporting the entire SAP business suite. In early 2023 this initiative was also pulled back, and replaced with the current program that we are now seeing materialise as Joule. So it has been a long time coming, but since the decision was made (public) things have moved fast – and rightfully also should move fast.
Oh, Why the name Joule you may ask?
At high school many years ago, I learnt that a joule (pronounced /ˈdʒuːl/, JOOL) is the unit of energy in the International System of Units. It is both equal to the amount of work done when a force of 1 newton displaces a mass through a distance of 1 meter in the direction of the force applied, and also the energy dissipated as heat when an electric current of one ampere passes through a resistance of one ohm for one second. Joule got it’s name after the famed English physicist James Prescott Joule who lived from 1818 to 1889. In the context of SAP Digital Assistant you can say that Joule will give the user an extra force and energy to get things done and to get it done fast ... I like the name! It also makes sense to give a tool - that mimics a person - a real name just like other large tech providers have done, for example Siri, Echo, Cortana, Nina, Viv, Jibo, Silvia, Bixby, Lucida, and Aido. So this strategy puts SAP Joule in good company with a catchy short name that works in multiple language zones (I know at least it can be spoken in at least English, German, French and Nordic languages without issues, although in France they may mix it up with another gentleman with last name Vernes).
Concluding Remarks
I believe that Artificial Intelligence, Machine Learning and Digital Assistant are topics that most companies will be utilizing. I don’t know if we are at the dawn of a new technical revolution, but I am sure that my colleagues at Effective People and me will be working much more with AI in 2024 than we have done previously. We also see this materialized through topics that our clients raise with us: they include more often now topics such as “how can we improve productivity through automation / technology support / process alignment / end user support”. Joule comes in very handy here because it provides support for all of these areas.
SAP has provided a short infographic that summarises their mission and priorities of generative AI in different functions, including HR. Check out this link: https://www.sap.com/documents/2023/07/c814592c-7e7e-0010-bca6-c68f7e60039b.html
SAP has gone a long route to get to this point. I think it makes sense to utilise a proven technology from a partner company (IBM). Although technology is proven, and that SAP does a great job of protecting end user from even seeing the tech layer, I am afraid of the extra technology component and what it will mean in terms of complexity. I am also unclear on licensing terms and under what conditions Joule will be made available. I am super curious to see how customers can make Joule their own, and add company specific knowledge, intellectual property and capabilities. Because a company is better off with an assistant that knows the company that a range of assistants that knows parts of the company. This is to be proven, but with the speed that SAP moves I am not afraid of the future – rather contrary.
Thank you for reading this and I hope it was of value to you. If you would like to discuss particular elements of Joule, how to use AI in HR and SuccessFactors specifically, feel free to reach out!
Erik Ebert, SuccessFactors Solution Advisor and Tech Geek, Effective People
With over 25 years of experience in the tech industry, I am a Senior Director at Effective People, a leading provider of digital transformation solutions for workforce improvement. As a SAP SuccessFactors Confidant, I also advise on the SAP HR cloud leadership team on market development and innovation. My mission is to help organizations across industries improve their people operations and business outcomes through a consultative and collaborative approach, utilizing new technology for the benefit of humans. Feel free to connect with me on LinkedIn https://www.linkedin.com/in/ebert X https://twitter.com/ErikEbertDK as well as on SAP erik.ebert2
- SAP Managed Tags:
14 Comments
| User | Count |
|---|---|
| 15 | |
| 12 | |
| 4 | |
| 3 | |
| 3 | |
| 1 | |
| 1 | |
| 1 | |
| 1 | |
| 1 |
