The rise of Artificial Intelligence (AI) in HR Tech, SAP SuccessFactors, and beyond

[00:00:00] Hi everyone, thanks for joining the podcast today.

[00:00:07] Joining me is a real expert in the area of AI.

[00:00:11] It's Sid.

[00:00:12] Welcome, Sid.

[00:00:13] This is the first time I'm actually trying to do something a little different with the

[00:00:16] video.

[00:00:17] So Sid is one of the smartest people I know at SAP.

[00:00:20] I genuinely mean that.

[00:00:22] Someone I love talking to because I always learn something from him just in casual conversations.

[00:00:28] He's super curious. AI is the future. I mean, especially generative AI and the direction that we have with SAP and also the other companies, it's the future of work. So super excited to be on the podcast. For sure. So let's talk about AI, right? I mean, everything changed when chat GPT launched, right? I mean, I know AI wasn't new, machine learning wasn't new.

[00:01:41] People had chat bots already.

[00:01:43] It was already integrated into technology, right?

[00:01:46] Of course you have your Google assistance and Alexa. it still needs the human intelligence. I mean, we are still running through captures and proving that we are not robots and all that. So I don't think we can do the jobs. I think the best way to look at the technology is to see how can it complement what we are doing. So that's the kind of use cases that SAP is building with success factors and with other solutions also

[00:03:01] to use AI and especially generative AI technology

[00:03:06] to augment what people do, I guess I wonder if everyone starts using Chat GPT or BARD or Microsoft Co-Pilot and they have these amazingly professional replies and everyone's creating great content. How do you differentiate between people and talent in that case? I actually even have a funny story. I decided to take a trip with my family.

[00:04:22] I used AI to create a e as original and it's very, and again, you may have seen the fake videos of people. It's extremely hard for systems and even humans to detect that. I have a song. I mean, some of the songs from artists, that was crazy. Absolutely. In fact, fun fact,

[00:05:42] I was listening to a deceased artist and it's going to be AI generated for the most part. What they are now waiting students on, and again eventually people on would be, are you able to even ask the right questions? Sometimes it's a creative aspect that comes in. As good as these systems are, lot of it as we all know is based on how you prompt it.

[00:07:01] Prompting general content comes into play.

[00:07:03] They're saying the problem statements is the same,

[00:07:06] but are you asking the right next level, how to do licensing, how to do copyright checks, and hence one and so forth. So I think we are going to this tectonic shift where some of these AI and especially generative AI is so disruptive. Like we mentioned that that concerns around private privacy, originality and copyright, especially are are in question. So so as long as those boundaries are understood and

[00:08:20] and people are and especially companies and employees are

[00:08:24] looking at it in a way knowing that people are

[00:09:22] You don't want people going and taking your data, putting it in chat GPT,

[00:09:24] which I think to some degree is happening, right?

[00:09:26] So part of tying it in that business process and using it.

[00:09:30] If we're looking at it from the vendor point of view,

[00:09:34] which both of us are working for SAP,

[00:09:36] I think that that's hugely relevant.

[00:09:38] What are some of the use cases that you see for AI and I guess

[00:09:43] SAP success factors, but maybe you want it to be more business relevant than nuanced,

[00:11:00] as opposed to being trained on a very genetic set.

[00:11:03] Of course.

[00:11:04] A good example is today,

[00:11:05] if you go to chat GPT and say, their data is protected and it doesn't leak into the public domain. It's also isolated so when you have customers in different industries, they are rest assured that anything that is copyrighted or proprietary to that is not shared with competitors. The second aspect is AI is bigger than generative AI.

[00:12:20] Also, one of the things or maybe it's

[00:12:22] people just equate generative AI to. So if you are creating job descriptions, it will, you can now enhance it with AI. So think of it as a co-pilot as opposed to something completely standalone. So there are chances that it may hallucinate or even fabricate information, but we always are conscious that the human is, the, has the, has the say in validating

[00:13:40] and curating that content so that you are assured

[00:13:43] that you're not accidentally sending something

[00:13:45] which is incorrect.

[00:13:46] So we are bringing human back into the loop. that is enterprise ready and ready for success factors. Customers also quite different and quite complex than just someone trying to work with chat GPT directly, right? So that's basically what the difference is. Yeah, I mean, you're solving a lot of the issues, right? One, the security aspects you're mentioning,

[00:15:00] but two, you know,

[00:15:01] like there are plenty of people that know chat GPT out there,

[00:15:04] but you just don't think of to use it, right?

[00:15:07] And so you want certain behavior, you can actually tweak it or train it to answer the spawn to that. So I think that's going to be key. And also, just to add on to what you said, I think the security aspect is important, but we are also

[00:16:22] expanding it to say HR is not just in an island, right? So if I'm a finance professional, yes, I do deal with the HR system and maybe I'm a manager and I have to be involved

[00:17:41] in creating a job requisition or something like that,

[00:17:44] but then in my job as a finance professional,

[00:17:46] it'd be nice to use the same on your own and bring it into your enterprise landscape. How amazing is that? Like imagine not having to. Not that we don't want to work with IT teams, but again they can do a lot more other complex things.

[00:19:01] You know very simple things can be can be now

[00:19:04] generated so you all of a sudden I'm becoming an

[00:20:03] to how I do it. What kind of responses do you actually get from people when you showed them to this that are non-techies like us?

[00:20:06] Right. And most of the stuff that I've shown for JUUL, for example, at SuccessConnect, and

[00:20:12] we were there, most of it was to customers. And I think there is this expectation that

[00:20:17] everybody is learning and really, generative AI always good. So the moment you launch the first version of it, they want 10,000 more things. And we have given them a roadmap saying, this is what's coming. But I would not call it just an agent, but it's a conversational agent. But there are underpinnings of these AI,

[00:21:41] which is more embedded in nature

[00:21:43] where you can do recommendations,

[00:21:45] you can generate things.

[00:21:46] All those are also coming. I want to be respectful of your time and where should someone wanting to learn more? I'm sure we'll have a lot of listeners that are curious. I think there might even be some AI playground that you've mentioned when we were just talking before this meeting. So can you tell us where someone should start and I can share links and stuff like that in the podcast details too. So for listener,

[00:23:01] where should they start?

[00:23:03] Absolutely, so there is. So one thing is an I

[00:23:05] actually with some folks, at use cases, specifically if you're a success factors or a PTP or a customer experience customer, you can actually go into very specific areas and see what all is activated. There are product tools that are available, there are guided experience available, and then very soon we will have certain trials where you can try these things.

[00:24:20] It's moving at a fast pace.

[00:24:22] I think the first thing is if you are

[00:24:24] just curious to learn about it,

[00:24:26] there are some blogs that we can share out on