Recruiting Rewired w/ Shiran Yaroslavsky

Is the future of HR tech just more features, or is it finally less friction?

In this episode, Chad Sowash sits down with Shiran Yaroslavsky, SVP of Product and a leading voice in agentic AI, to dismantle the hype and get real about enterprise-scale hiring. Shiran isn’t just talking about chatbots; she’s architecting a world where "agent orchestration" and "agent-to-agent communication" do the heavy lifting so recruiters can finally get back to the human side of the job.

Deep Dive:

• The Death of the Dashboard: Shiran explains why forcing recruiters into complex systems is a failing strategy and how to move workflows into the tools they already live in—Slack, WhatsApp, and voice.

  
  

• Waking the "Zombie" Databases: How to use AI agents to activate years of idle candidate data that most companies have completely forgotten they own.

  
  

• The Trust Architecture: A masterclass in AI auditing and bias monitoring. Shiran breaks down how to build transparent systems that candidates and recruiters actually trust.

  
  

• Frictionless Migration: Hear how Shiran’s team is using AI to turn 18-month ATS migrations into a project that takes just weeks.

  
  

• The Psychology of Adoption: Why "gradual automation" beats "all-at-once change" every single time in an enterprise environment.

https://www.youtube.com/watch?v=JnKqXlRfkgk

Episode #2 - Talent Autopilot

Episode #3 - Trust First

PODCAST TRANSCRIPTION

Chad Sowash: Welcome back, Shiran.

Shiran: Good to be here. Thank you for having me.

Chad Sowash: Last year at this time, we were in Spain together, just a different part of Spain. [laughter] Uh, and things have changed dramatically since then. Last, last year, uh, you were giving... Given a new mission, product-wise. Winston.

Shiran: Right.

Chad Sowash: Right? Winston was launched. You and Michal made that happen.

Shiran: Yep.

0:00:60.0 Chad Sowash: Uh, and then, and then there's [laughter] an acquisition.

Shiran: Yeah, that's, you know, I think part of, I think, to be honest, it's part of the impact of Winston. Uh, I think we made some really a lot of impact for our customers. Also, in the market, took it very well, and the acquisition is, you know, part of the outcome. So that, that was the true success story, I think.

Chad Sowash: Talk, talk a little bit about the change. So there's gotta be a transformation of talent within your organization going from the old-style applicant tracking system form-based model to a conversational model, like a Winston model, right? Because those building UI for one is not building conversational UI. They're two entirely different things. Talk a little bit about that. What, what kind of growing pains did you have? How did you, how did you overcome that? Because obviously you met the goal, right? What, what happened? Where were the bumps?

Shiran: Yeah, I think, you know, it's, you never ship AI-based product and then, you know, that's it, right? It's like magic. You always need to go and iterate, and you learn so fast as a, as a product leader. Also, some of the intuition that we have about product, because it's, it's very different, like as you mentioned here, it's conversational. Like, what's, how people react to AI, their expectations is very different. So you need to keep iterating and improve all the time. So I can't say, you know, we delivered that, we're done. Uh, we always aim higher to get it better as well.

Shiran: I think to your point, how, how we're thinking about the interaction with the product is that, and you've heard me, I'm, I'm talking a lot about, umm, seamless personalization, right? And in order to create a truly unique experience for each and every individual, you need to also, I would say, expand the concept of UI. So we're talking a lot internally about adaptive or composable UI. So how you're taking your UI and breaking it so you can basically surface it, you know, no one really wants to log into ATS, you know that.

Chad Sowash: Oh, God, no. Yeah, nobody wants to log into all the different tabs that they have open on their browser, but yet they have to. How do you, how do you push past that? Because again, you're one system, they have all these other systems. What, what's the, what's the future look like for SmartRecruiters to be able to get away from multiple tabs and literally just providing the user what they need in the time that they need it?

Shiran: It's a lot about tapping into the flow of work, right, and where users are. And so what I think you would see a lot in the system and what we are building, umm, toward is the ability to basically serve all those tasks that users need to do wherever you are. So if you're spending more time, you know, using Teams or WhatsApp or SMS, this is where you should, you know, take your task.

Chad Sowash: Do work. Okay.

Shiran: Exactly. And, and, and provide feedback about the candidate, you know, you don't need to log into the system. You can do that maybe with the, uh, voice, you know, using your WhatsApp.

Chad Sowash: So a hiring manager doesn't have to go into SmartRecruiters. Winston just shows up in Slack or WhatsApp or something. Is that, is that what you're talking about?

Shiran: Yeah, exactly. And I think the other piece of it, umm, there is two more aspects to it. One is the agent-to-agent. So I think in the future, what you will see, and you're talking to many vendors in this space, umm, everyone are building the agents. So how the, all this swarm of agents can collaborate with one another, this is where we believe the true, true value will be unlocked. So this is one aspect of it. And I think in the future, what we call zero UI is that you will see more and more use cases where people interact with product without even screens. So they will use voice and body gestures. And how you building product for this type of reality, right? Umm, so this is also top of mind.

Chad Sowash: That's gonna give you more time back because you're not building all these frameworks and all this UI. Literally, you're just feeding new features into a conversational system.

Shiran: Exactly.

Chad Sowash: Right. So that's gotta be exciting. I mean, you know, maybe not for UI designers, but I mean, that's gotta be exciting because the, the discussion is always about ship, ship, ship, ship, ship, right? And if a lot of that time is now taken out of the equation, and I'm sure there's some transformation that you're gonna have going from forms to, you know, fully conversational, that's gotta be exciting for, for, for you. What about the team? Because obviously the team, uh, you have those UI people and you've got the people that are just used to, "Hey, this is, this is what we, it's always worked this way."

Shiran: Yeah.

Chad Sowash: How, how has it, how's it been trying to actually pivot them out of that mindset into a, uh, tomorrow mindset?

Shiran: Yeah. I think, and this is something that you hear, you're gonna hear from, you know, all tech companies these days, how you upskill the people. How you're changing the mindset. A lot of, I think in the space of product management is also changing. So for example, if it used to be very strict like functions of engineers, product and design, there is kind of a mesh and you can use for example cursor or other kind of tools to, for as product manager or designer that don't know how to code.

Chad Sowash: Sure.

Shiran: Basically go and create code in the, in the product, and use Figma Make to create a whole UI. So I think there's a lot of upskilling that we are doing a lot with in our team. But also in terms of mindset we, we are bringing people that are experts for conversational design, for example and there is a whole, you know, skill set around that and how you're making it really, really better. It's an, it's an art.

Chad Sowash: Right.

Shiran: Umm, so there is a lot of, I would say new knowledge that you need to bring to the team. A lot of upskilling. Umm, and we always try to challenge and what we think find best. Especially if you're working on AI based product, as I mentioned, many of our assumptions or expectations of the users is very different. So what we're trying to do is to ship and iterate. We have really strong design partners program getting their feedback on every step that we're having, making sure that we're building, you know, based on the right expectation, meeting their needs. This is how you, you know, ship paths. I'm sure that Michal will talk to you about it.

Chad Sowash: Oh yeah, well it's, it's ever hard to meet expectations of a user, right? Especially because there are so many different users and they have different expectations. Umm, and for years we've sent them into a black hole. They weren't happy about that and we understand why they weren't happy.

Shiran: Right.

Chad Sowash: Umm, but then you're hearing about job uh, seekers, candidates who don't wanna talk to AI, they don't wanna talk to chat bots. So it's like the expectations like you have to constantly meet the next expectation and when you're talking about Conversational AI. How in, in the iterative process, how do you make it even though it's not human, how do you make it more human? Is it literally just time or what training? What it, what, what's the secret sauce?

Shiran: It's terrific question. Umm, I think it's, it's a few things. One, philosophically, I would say, umm, we're designing the product in a way that there is a few crucial decision points that we are keeping human in the loop. So not everything, we believe, need to be fully autonomous and agentic. We don't want to completely delegate the decision making to our agents. Umm, we truly believe that in order to keep the, the process human, you need to decide very intentionally in some decision, umm, points where you need to have the human decide. And so I think this is one aspect of it.

Shiran: Umm, second, there is different things. For example, how the product react when, umm, you know, you're not... They're not getting the right answer, and how it kind of correct and recover and how, you know, how the voice and tone sounds more human. How you understand better context, how you better understand the user goals. Umm, so there, there's a lot into it, you know, really to, umm, we call it for you to feel like you have a partner, not chatting with a chatbot, right? Umm, and so there is a lot of, I would say, intelligence coming to it, a lot of design. Umm, so, yeah, it's not an easy task, I must say.

Chad Sowash: Well, I, I feel like there's a double standard though, because there are so many people that are out there now who have these AI assistants, and we're even hearing about AI partners, boyfriends, girlfriends. So it's like the expectation is I, I need somebody there. But when I'm looking for a job, I gotta have a human. To me, to some extent, it kind of feels overblown. I, I think, I think that because of the black hole, decades of going into the black hole, any type of connection right now is good. But I also believe what you're doing and what you're talking about is trying to make it as human as, as possible. Now that can't be easy when you're, I mean, uh, a, a tone even from the west coast of the US to the east coast, you live in, in New York, right?

Shiran: Yeah.

Chad Sowash: It's... That's different. Those tones just in that one country is very different, let alone going to, you know, the UK or, or Asia PAC or something like that. So how do you start to build for those types of culture differences in tone and conversation?

Shiran: Yeah, it's, it's a great question. So first you need to provide a lot of flexibility. So we, for example, we have a product that we just delivered for early access called Winston Interview. It's the ability for, uh, Winston to, umm, send it to do the pre screening, uh, interview for you using video or a call. Umm, and what you, you see there is that you can also tailor as a company what kind of voice and tone it will have, what it will know about your company so it can answer the right, you know, answer the right questions in the right way. Umm, and so this is, uh, this is part of it, right? How you really give control to the, to the... Also to the customer because they will have different nuances. They want to represent their brand in a different way.

Chad Sowash: Yes. It's a great point. Yeah. Yeah. Yeah. Yeah.

Shiran: And I think what we also learned specifically, also in the context of this product is that expectation shift very quickly these days, the pace of AI And I think in the, in the future, in the, in the past, I would say if people were, you know, seeing chatbots, they're saying, I'm not talking to a chatbot. I'm, I'm going to, you know, wait for a call into an agent, you know. Umm, but, uh, what we see is that a, like, younger generation is really more open to using, umm, agents, umm, to receive, you know, different services.

Chad Sowash: Yeah.

Shiran: Umm, we also see that people are more inclined, inclined to use it when the quality improve and technology is just getting better. So if it feels more human, to your point. Then they're more open to use it. Okay. It's not that bad. It's actually saving me time. I don't need to wait for scheduling an interview with a recruiter in two weeks from now. I could just do it now. Right.

Chad Sowash: Yeah.

Shiran: Umm, so...

Chad Sowash: Which is always painful.

Shiran: Yeah. Always painful. Umm, there's huge bottlenecks. So why, why, you know, getting into those bottlenecks if you can do it differently? So I think it's a matter of maturity, people expectations in openness, technology, different, umm, you know, changing, you know, pretty fast. So I feel, I feel like we are getting there much faster than we thought we would.

Chad Sowash: Let's talk about a couple of these products that I saw you talk about today.

Shiran: Yeah.

Chad Sowash: And they were sexy, by the way. Now you think one is more sexy than I do. I'm, I'm on the... In the other camp. We'll talk about...

Shiran: I'm, I'm glad that you're thinking that.

Chad Sowash: We'll talk about, we'll talk about mine second because we're... You're, you're. You're the interviewee. So. Agentic CRM.

Shiran: Yeah. Mm-hmm.

Chad Sowash: This to me is, uh, exciting from the standpoint of so many companies had CRMs for so long and literally, much like the candidate database that they had their candidates in their applicant tracking system. Two entirely different databases. They both literally just atrophied and, and died. They didn't do anything with them. They might have had some drip campaigns here and there, but generally no, right? So this, to me, and you correct me if I'm wrong, this to me almost gives you somewhat of, uh, an autopilot to set some rules and allow agents to, to go out and engage and make those people feel special as opposed to feel nothing at all, which is a very lonely [laughter] place. So talk about that. Am I, am I, am I at least in the right lane?

Shiran: I, I agree. And I would say, uh, what we're hearing these days is, and what's different than, I don't know, three years, umm, you know, ago is that companies these days are getting so many applications. You know, it was a very different market a few years ago. Umm, also talent teams are getting smaller. And so instead of just publishing and advertising another job and then getting more applications, they probably. You have, you already have the applications that you need in your system, right?

Chad Sowash: You do. [laughter] Yes.

Shiran: But you're missing the, the, the tool or the capability to unlock it and lock the potential. Like be able to search, you know, in a conversational way, as we mentioned, across all your talent pools. You know, just describe, just describe what you need right now and be able to find it with, you know, a few seconds. Umm, and it sounds so intuitive, so needed, but yet, like, no one really doing it. Not well. Umm, and we have our own CRM, so we are also part of the problem right today. Umm, and we thought, well, that's not good enough. This is not what the customer really needs. Again, about expectation, where the market is heading. So instead of just, you know, throwing more advertisement kind of budget, umm, let's...

Shiran: They probably have all the potential that they need, all the talent they have, so why not helping them to unlock it? So, uh, I think this is a big part of it. And then the candidate side, as you mentioned, be able to personalize it, meet the user where they are. If you're using TikTok, you saw today, you know, you can run TikTok campaigns if you, you know, more like a, you know, you know, using WhatsApp or whatever you... Wherever you are, you can, you know, basically tailor it to the voice and tone, the, the people expectations, umm, their region. Umm, so it also create a much more human, personalized experience than this cold, you know, emails that people are usually getting.

Chad Sowash: Yes, yes. Well. And as you know, Rebecca has talked about consumption models and those types of things, being able to look at like different tiers of, of cost and saying, look, the system is working for you. And what happens is we're going to engage these individuals who we think are qualified, but we're going to reach out and do a gap analysis and then ask them if they do meet the requirements. If they do, then we push them forward. If they are qualified, umm, then you pay consumption wise.

Chad Sowash: If we have to go outside the system, that's an entire... That's obviously more money because there's, there's more, more, more, umm, uh, vendors that you have to actually engage with. Uh, that to me just sounds like the next evolution. You, you've already paid for most of the talent that's out there if you're a big enterprise company, or at least the ones that you want. Umm, so taking that model even further, going to internal mobility and having a segment of your CRM that's literally dedicated to trying to get your people that you already have in place, it's a hell of a lot cheaper to keep somebody and move them to either laterally or up to another position. Umm, are you guys working on doing that as well?

Shiran: So, I think it's a, it's a terrific point. And we see also companies, these days, many of them are downsizing or you need to, umm, hire for new people, umm, at new roles. And you, you already, to your point, you already have a really good talent that maybe you can, you know, leverage in a different way. And so internal mobility, we see it like the, the potential there and the interest is just, umm, increasing, I would say. And a big part, you, you saw our vision around the integration, umm, it's also to integrate much better with the kind of what we call within SAP, it's like, uh, Employee Central, other parts, talent management part of the system to be able to do just that.

Shiran: So it's but just connecting the dots. So if you have the information from the rest of the suite basically about your talent, what they're interested in, what their goals are, you know, what their performance reviews are, what's their inspiration, you would be able to target much better, you know, basically giving some recommendations, saying, "Hey, you know, I know that you need this new role. There is two folks that are really good fit for you that are already working in other teams." And large enterprise, usually they have this talent, but it's not something you can manage easily.

Shiran: It's like huge corporations, you know, so it's like tons of blind spots. To be able to use the Agentic CRM, as you mentioned, together with the right data, this is really unlock a lot of potential.

Chad Sowash: Uh, just the opportunity to be able to, hopefully, as you, you'd said, umm, talent teams are smaller. Uh, they need effective processes and tech to be able to help them do their job more, faster. Uh, and with that consolidation, you're starting to see, you've been seeing for years, many companies consolidate their talent acquisition team and their talent management team into just a talent organization where they're literally just focused on all of talent. Which means, obviously, this whole infinity loop that Allyn's been talking about for about 10 years now actually goes into place and a system is designed to actually facilitate that, that whole infinity loop.

Chad Sowash: And you showed today on this very sexy video that I hope you will actually share with everybody because that was, that was probably the best marketing, umm, was that you had alumni in there as well. So you can have your internal database, do campaigns for TikTok, internal database, right, of people that we've already paid for, and then also alumni and employees. It's like that type of connectivity, uh, to me is magical. So talk a little bit about that. Is that, how long do you see it taking for us to actually get there, especially in the connectivity of systems? Because a lot of TA systems aren't connected to, to talent management systems.

Shiran: So, I would say about us, I think it's not gonna take long. So this is why we're also investing in integration. AI at the end is...

Chad Sowash: Did you hear that? It's not gonna take long. [laughter]

Shiran: So, and this is where also prioritizing, umm, some of the effort that we have on, on integration. Because AI is as good as the data, that data that is getting access to, right? And so if you get the right data, your AI is much more powerful and impactful. And so for us, we are just building the really strong integration to Employee Central, to onboarding. This is exactly where that this data is actually, you know, lies. And so for us strategically, in order to get to this point is exactly about it, to be able to get the right data, be able to tap to where users are in the system, in the right place, in the right time. So definitely it's on our roadmap. So not too far away.

Chad Sowash: You heard that. You heard it here first. Umm, so we're gonna go ahead, since you've already mentioned it, my favorite, uh, was the integration piece. We, we both know that the biggest cost of change management for these huge systems is literally the integration.

Shiran: Yeah, that's true.

Chad Sowash: It's integration, right? It takes months, six months, 18 months, depending on how big the system is. You've got humans actually matching fields and doing all of these things and it takes forever. What I saw, and you got me all excited, umm, was an integration by the touch of a button. Now, I'm sure it... There's more to do with it than that, but if I can go from 18 months, six months to six weeks...

Shiran: Mm-hmm.

Chad Sowash: That saves a lot of money, saves a lot of time, and my cost for change goes down dramatically. And that gives you guys a market edge. So, talk a little bit about how deep that actually goes. Obviously, applicant tracking system, that's where the start is, but, uh, where, where do you guys go with this type of technology? And when can we prospectively see it in operation?

Shiran: Yeah. So, I would say, and, and we ran discovery, and you can't believe people's reaction, like, gut reaction when they're hearing the word migration.

Chad Sowash: Oh, yeah.

Shiran: Right? Like, they're like, "Okay, it's gonna be so painful."

Chad Sowash: Oh, data migration? Uh.

Shiran: Yeah, exactly. So, this is the...

Chad Sowash: The shivers. Yes.

Shiran: I know. And so, uh, we, we asked ourselves, "Well, we can, you know, do and, you know, move data from one place to another. It will work. That's fine." But, you know, you know us already by now. We said, "Well, is it, is it good enough?" And the question, the, the answer is no, right? And, and we thought, "Well, how we can reimagine the way people integrate and migrate to new system? And how we can leverage that to be a differentiator for us?" as, as you just mentioned.

Shiran: And so, we are having, like, huge groups, a lot of people that are also experts in, in AI, thinking about how we can design it in a way that it will be productized, that we can also create, like, really powerful AI-based tools to help to save a lot of time. So, I think we, we talked about it briefly, but, umm, you mentioned field mapping. This is a huge pain point. If you're a large enterprise, you have hundreds of custom fields. Be able to map those to another, you know, field in a different system takes sometimes weeks or months.

Shiran: Umm, so be able to have an agent is just, you know, doing that for you in a few seconds, you know, and providing you the visibility, umm, this is, you know, changing the game really for companies. And so, I think this is how we're thinking about how we can make it delightful, how we can actually use that for a moment in time to actually also making sure they're integrating right. You can't believe how many customers that we speak with are integrating in the wrong rate, and then they are suffering years later.

Chad Sowash: Oh, yeah.

Shiran: So, how you can make it smarter, so you make sure that you're actually doing it right, so after that you have the best product experience. This is our interest as, as product organization, right? Umm, and the way we're thinking about it is also strategic, as you mentioned. So, because if we're doing that right, we can expand it, so you can use that for new customers that are coming to us from different, you know, suite or different tools and want to go and opt in to use SmartRecruiters. Uh, we can do that for, umm, other modules within SAP, like we also, uh, own onboarding. It's a whole new product. Also, they need to migrate from older, older version to new version. So, why not expand it into it? So, we're thinking about it in a way that how we can leverage the technology to really be a game changer for our customers.

Chad Sowash: Is that, is that the first phase? The first phase is, "Okay, we have all these SAP, you know, success, Success Factors customers, umm, we've gotta migrate them over into SmartRecruiters, some aspects of their, their, their business over into SmartRecruiters." Is that phase one? We're gonna, we're gonna eat our own dog food, we're gonna build this ourselves. Is that phase one, and then you start pushing out to others later?

Shiran: Exactly. So, we, we actually, the way we, we did it internally, we told customers, right, uh, "We're SmartRecruiters. If you're interested in, you know, being early adopter for us, umm, raise your hand if you want to go and migrate in the next few months." We got, let's say, much more demand that we could ever consume.

Chad Sowash: I'm sure. I'm sure, yes.

Shiran: Uh, we were very surprised in a good way. Umm, so what, how we're doing that, we, we definitely have, we have different programs. So, we're using those tools to different... We have a design partner program, we have a larger, more, a broader program, but still very curated. Umm, so we can work closely with customers to understand, again, customer expectations. What's the experience like? What's the, are we meeting their needs? How we can make it better and iterate on that?

Shiran: So, first, as you mentioned, umm, there is so much demand and interest from SAP customers that are today using the, uh, recruiting... Umm, the Success Factors recruiting system. They want to move. How we can make them successful and do that in a few weeks versus, you know, a few years? Umm, so this is, you know, probably phase one. You know, we're gonna iterate, we're gonna make it amazing, and then we can expand it to new, more use cases.

Chad Sowash: Well, and the faster that you can do that, the more revenue you can push through because you don't have a bunch of companies that are waiting in line for integration or migration or just to get the, the, the ball rolling, right? So, this is... I mean, to me, this has to be probably one of the most innovative, and it's boring, I know. It's, it's, it's...

Shiran: It's not the sexiest.

Chad Sowash: It's data, it's data migration, it's integration, it's boring. But it's the biggest problem that most companies have and why they don't change. I've had so many companies tell me that, "I would get off of this platform if it wouldn't take me 18 months."

Shiran: Yeah. They are basically locked, right?

Chad Sowash: Yeah. I... And, and they feel like they're, yeah, they're, they're locked in a cage, right? Whatever database cage that is. Do you see this being the new standard where it's like, "Look, systems should be more fluid, they should be easier to access," umm, and then you'll see your competitors doing the same thing. They're gonna have to, to be able to match. Uh, because if that is the biggest go-to-market obstacle, if I am right, for many of them, they're gonna follow your lead.

Shiran: Exactly. And I think, you know, many companies are choosing to use AI for the day-to-day tasks. But as you mentioned, if you're a large enterprise, adoption of new systems or moving between different systems, it's a huge pain point that today no one really targets, not really. And be able to remove this kind of blocker really allows much more freedom for, for companies, uh, these days to choose the right solution for them, umm, not being locked into something just because it's, you know, legacy, someone chose it like 10 years ago. Umm, so this is how we're thinking about it.

Chad Sowash: Well, and it's an unlock for either further down, downstream into SAP's business suite because obviously if you guys lock into this integration product, you can start to do, you know, obviously other point solutions into SmartRecruiters, into SAP, so on and so forth. So it becomes almost like the, you know, you had the Apple iPhone, then you've got the AirPods. The AirPods could be its own company, the amount of money that it makes. I mean, that could actually be the, the unlock for, uh, a go-to-market in many different areas of the business.

Shiran: Exactly, exactly it. And there is a, umm, also selfishly, umm, there is a huge power of being part of the suite. So if you make it easier to go and basically buy more modules, you can basically... It just becomes easier to be part of it and, and benefit from the suite-like power, right? And this is part of also SAP's, to be honest, like strategy, like how you leverage the power of the suite because all those point solutions creating all this fragmentation, right? And all the problems we're talking about in, in recruiting and hiring these days.

Shiran: So how you make, removing friction, it's a lot about removing friction so they can basically have better connectivity, umm, have AI access all the data they need to really unlock the value. So it's about it, right?

Chad Sowash: And if I don't have to spend money on 18 months of Deloitte consulting, I can actually put that back into new tech earlier, which, I mean, this is another unlock where it's like, look, you're gonna be saving this much time, this much, this much money. Now you can really double down on building a stack quicker, an agentic stack. I, I just thought of that and that's just brilliant. So when it comes to, when it comes to one of the things that, that, uh, Rebecca says is that, you know, pretty much for, like, recruiters, 80% of their job is admin.

Shiran: Mm-hmm.

Chad Sowash: Right? Uh, and I think what we've tried to do is we've tried to say, "Hey, here are the tools. Use them," as opposed to just bake the tools in, much like interview, uh, scheduling, right? If somebody goes and they, they, they meet the requirements, then boom, they're automatically scheduled. Nobody has to touch them. It happens behind the scenes and it's magic. How do we get to that level in more of the tasks throughout without scaring the living hell out of recruiters and hiring managers and, and so on and so forth? How, how do we get there? Is it literally just kind of like piece by piece or do we have to force-feed them the whole, the whole thing? [laughter]

Shiran: So in my view, and I analyze a lot of, umm, products that were successful in being more agentic because there is no many products like that across industries that you're looking at. There was a few that, you know, kind of, I would say, hacked something that, you know, making it work and it's really more agentic, more autonomous. And what you could see, the way they did it, they did a few things well. One, they tap into the flow of work. They don't force people to come to them, they're going to where they already are.

Chad Sowash: Yeah.

Shiran: Second, they build trust gradually. So you don't start with the most risky task that, you know, recruiter has and saying, "You know what? I will take this decision from you. I'm just gonna do it for you," right? This is not how you build trust. And you start from places like scheduling that maybe the, the cost of error is not that bad. Umm, you're building this trust, you know, as you go and getting, getting there, I would, together with the user, and user feel more comfortable to go and, and, and try new things because they see the value, they see the impact, it's measurable.

Shiran: So it's about a lot about building this trust with users, especially large enterprises. So this is how you're thinking about it. But also you need to have the broader vision about how it's all connected because you also don't want to deploy different assistants and different co-pilots without, like, the full vision of how it's all connected. So if you truly will get to, like, flows that are much more agentic, how it will look like, right, as you're gaining more trust and how it's all connected as part of one holistic experience.

Chad Sowash: So, uh, that being said, uh, the thoughts of all these agents out there, you're gonna have agent-to-agent communication. Obviously, there's gonna be some redundancies and whatnot. So, in the orchestration, one of the words had... That hasn't come up yet, uh, the orchestration of how all of these things come together, is this gonna be literally like a, a new job that a lot of companies are going to have to create is work orchestration? With all of these agents, you've gotta know where the human fits in. Here's, here's the process.

Chad Sowash: Here are the tasks, here are the agent tasks, here are the human tasks. It... Are you guys looking at that? Or comp... Have you seen companies do that? Because to me, orchestration mapping isn't gonna be just something where you do it upfront and then you're done. It's... It almost seems like it's gonna be a constant.

Shiran: It's, it's a really good point. So what we see is also within the talent team that it, it becomes much more, you know, there is more automation, more AI being added. And part of the... I would say, what their new job is being like is also to be much more of a, you know, a designer of the process, of the experience, and thinking how everything fits together, what we need to be connected, how it need to be connected. And part of it is exactly what you mentioned about the agent-to-agent interaction.

Shiran: And it's, I would say, for most of this, you know, in this industry, in HR tech, it's pretty early stages, but we do have, we launched our first agent-to-agent last year, umm, use case. We have a few more that they are coming in H1. The first one was, uh, with one of our close, uh, partners, Visier.

Chad Sowash: Okay, yeah.

Shiran: Around analytics. So you can ask Winston, "Hey, how many, you know, people in how many regions I have, you know, in this type of skills?" And you can, you know, call to Visier agent, get the, the feedback there in a very conversational way to segment the data, follow up. So it's really cool use case. We're adding a few more. Umm, but the ability to really understand, look, how the orchestration would look like and how also govern it in a way that will make my organization secure and safe, which because agent-to-agent is also adding all this layer of complexity around security and access and permissions and who get the chance to work with which type of agent.

Shiran: So we have something called, umm, Winston Hub, AI Hub. Yeah. And so you would be able to see her there, basically all the agents that you have connected to your system, and you can very granularly control who got access to what. So this is part of it, right? And if you're thinking about it more holistically, it's, it's gonna be big, big part of about trust and keeping your also organization safe as part of this, all these new things that are popping up.

Chad Sowash: Yeah. Right. It's... When we're talking about product and, and obviously there are, especially in this space, uh, every space, uh, tech companies are incredibly competitive. And they're, they're, you know, on the razor's edge in many, many... Some are, some are not. Uh, and so that, that word trust means a lot because there are a lot of lawsuits that are out there right now. That are specific to agents enriching profiles, you know, uh, age discrimination. When are we going to get to the point, and, uh, where we have...

Shiran: You mentioned Visier, and that's why, that's why I'm thinking of it. Where we are going to have on-demand analytics and on-demand auditing so that we know every decision that's being made, it's being racked up. There's no reason why we should have to wait 30 days for that information, right? We should have a dashboard. We should be able to talk to Winston. We should be able to get reports not just on analytics, not just on, you know, what job's doing well or what have you, but also the audit process. Where are we falling down? Where, you know, where, where are the hotspots that we need to go take care of? How long is it gonna take for us to actually get to that point?

Shiran: I think it's, it's a must-have. Uh, I think as, as companies are using more AI, umm, also the expectation and, and with the new regulation is also, you know, getting higher and higher. Uh, we're using a third party to basically audit every AI-based product that we have. And so you can as a customer just get a report on every AI-based product that you, that you use within the system and just see scoring on, on... So many different...

Chad Sowash: So do I ask Winston for that and then, then it pulls back now? Or is that like you...

Shiran: So, because it's a, it's a third party, we want it to be objective. We could have built something in-house. We felt like, you know, it will make our customer much more, umm, I would say, comfortable to use a third party, that this is what they do, this is their expertise, it's not biased. Umm, and so you basically will get... Umm, you're getting, like, all these reports on a, a different cadence based on what you need. Umm, and you see all these different metrics on, you know, we're, uh, auditing for bias and hallucinations and discrimination, different type of, of categories there.

Shiran: Umm, so you can see it there. Beyond that, and, and you mentioned it as well, the conversation that people had with your AI-based product, you need to provide, you know, access to that. So, for example, the Winston chat, the candidate chatbot that we have today, you can see all that from your admin experience, right? Then you can...

Chad Sowash: So like the transcripts of conversations and... Yeah.

Shiran: You need to have this transparency. You need to lead with transparency to get to trust. And I think, this is, this is a big part of it.

Chad Sowash: Does the candidate have that transparency too? Do they have access to those?

Shiran: So I think we're, we're getting there. So there is more and more elements that we are, umm, looking to expose more to candidates. And I think it's a big part about, not only leading with transparency, not only with hiring teams, but also with candidates, but also how you're providing them added value as part of it, right? They're using AI, they're given in some, in some way, the access to their data. Umm, so how are you giving them also added value? So they feel good about also being, participating in this, in this process.

Shiran: And so, for example, one thing that we are, uh, thinking about is that right now we're using, umm, skills-based matching. We have a product called Winston Match, right? How are you exposing also to the candidate their match score and how Winston thinking about them as, you know, what's their strengths, their weaknesses, what they should work on? This is, uh, amazing added value to a candidate as well. So what, what do you do also to provide some added value and transparency to candidates? So this is some of the things we are working on.

Chad Sowash: And that's a lot to work on.

Shiran: Yes, that's true.

Chad Sowash: Phew. So I'm gonna let you get back to it. [laughter] Thanks for, thanks for joining me, Shiran. Thanks for having me here in wonderful Mallorca. Uh...

Shiran: Great view that we have.

Chad Sowash: It is beautiful. And, uh, hopefully we'll do this again.

Shiran: Yeah, looking forward to it. Thank you.

Chad Sowash: Take care.

Outro: Wow, look at you. You made it through an entire episode of the Chad and Cheese podcast. Or maybe you cheated and fast-forwarded to the end. Either way, there's no doubt you wish you had that time back. Valuable time you could have used to buy a nutritious meal at Taco Bell, enjoy a pour of your favorite whiskey, or just watch big booty Latinas and bug fights on TikTok. No, you hung out with these two chuckleheads instead. Now go take a shower and wash off all the guilt. But save some soap, because you'll be back. Like an awful train wreck, you can't look away. And like Chad's favorite Western, you can't quit them either. We out.