How AI Is Transforming Talent Acquisition
In this special episode of the HR Leaders Podcast, we bring you a powerful panel discussion hosted by Lynne Oldham, Chief People Officer at Dataiku, featuring talent leaders from Databricks, Omnicom Media Group, Eaton and Zapier.
Together, they explore how AI is transforming talent acquisition and talent management, from predictive hiring and internal mobility to bias reduction and personalized candidate experiences.
🤖 In this episode, you’ll learn:
Ways AI is helping reduce bias in the hiring process
What challenges teams faced during AI implementation
How predictive analytics are transforming workforce planning
How companies are using AI to match candidates and jobs faster
Real examples of AI-powered internal mobility and talent development
Deel is the all-in-one payroll and HR platform for global teams.
Deel helps companies simplify every aspect of managing a workforce, from onboarding, compliance and performance management, to global payroll, HRIS and immigration support.
Deel works for full-time employees and independent contractors in more than 150 countries, compliantly.
And getting set up takes just a few minutes.
Chris Rainey 0:16
Hey, everyone, welcome back on this one. I have the pleasure, we have the pleasure to hand you over to the safe hands of our moderator, and she's one that many of our audience knows very well. Oldham was so excited to have you back. She's a Chief People Officer of stash. Formally, when we last did the panel during the COVID, I believe last time we spoke, you were serving at zoom as Chief People Officer. So I'm so excited to have you as part of this. Lynn, how are you? I'm
Lynne Oldham 0:45
great. I'm great, and I'm now so nervous because I'm taking the reins from you, Chris. So I'm very excited and nervous at the same time.
Chris Rainey 0:55
No, well, listen, I know we're all in your safe hands. So have an incredible conversation with our amazing panelists, have fun, and we'll see you back here shortly. Okay, sounds good.
Lynne Oldham 1:05
Thank you. All right. So as you heard I am Lynn Oldham. Chris said that beautifully. I get everything from old ham to old man, so I appreciate Chris knowing how to say my name. This is how AI AI is transforming talent management and talent acquisition. I have a panelist of some old friends and some new friends. And what I love about this for you all is that these companies that they represent span the gamut. We have some that are barely teenagers, some that are teenagers, some that are in their 20s. We even have a centenarian on the panel. In terms of the age of a company. We have company sizes from under 1000 up to almost 100,000 so you've and two privates, two publics. So you've got such a gamut represented by the people who are going to speak, and I am going to let them introduce themselves, and then we'll kick it off. So let's start with Amy.
Amy Reichanadter 2:06
Well, first of all, thank you so much for having me today. I'm really excited to be here for this discussion, and I'm Amy redkane, or I'm the Chief People Officer at Databricks. We are a data and AI platform company, and we are currently about 8000 employees.
Lynne Oldham 2:21
Excellent. Diana,
Diana Blancone 2:24
I share in the joy of joining all of you talented people. Thanks for having me. I'm Diana blanc Coney. I'm the chief talent officer at Omnicom Media Group. I've been there a little over three years. For those of you who don't know, we're the biggest media holding company in the world, and we get the joy of helping the biggest brands in the world create more awareness and essentially growing some of the biggest brands across the world. So we've got a very active hiring engine. We hire 1000s of people a year, and we usually do it with a big smile on our face. So
Lynne Oldham 2:58
I love it. I love it. Thank you. Diana. Sigh. You. Side.
Sai Patel 3:06
Sorry, good morning, good evening, good day. To everybody around the world. So here from Eaton, excited to be part of this group. So, Lynn, I think we're, we're the senior citizens in this place. Eaton Corporation, been around since 1911
you know, we're in about 175 different countries, little over 92,000 employees over HR, tech, HR, enablement and iteration,
Lynne Oldham 3:36
and you've got a partner in crime with you. Wendy,
Wendy Hirsch 3:40
hi everybody. I'm Wendy Hirsch. I'm also at Eaton, and I have responsibility for HR enablement, innovation, HR shared services, workforce, analytics and planning. And then I also have an amazing continuous improvement OPEX team. And I'm excited to be here. We've done a lot in the AI space, so I'm excited to share some very practical examples. Excellent.
Lynne Oldham 4:00
And finally, last but not least, Tracy,
Tracy St.Dic 4:05
hi everyone. I just realized Zapier is 100 years younger than Eaton, so quite the spectrum indeed. So my name is Tracy stike. I'm the global head of talent at Zapier, and Zapier is a no code AI powered automation platform, and our mission is to make automation work for everyone. We're a private company. We have about 750 employees, and we're distributed across 42 countries, fully remote, and I've been at Zapier for just under three years. Excellent,
Lynne Oldham 4:32
excellent. So let's get this going. I think we'll start with a fun, fast start question. Tell me one thing that you use, you've used, personally with generative AI that has really helped you. And my example, I'll kick it off, is I've been trying to figure out where I want to do a holiday, and I've had chat GPT. Giving me multiple scenarios, pricing it out for me, figuring out the best time to go, temperatures, etc. It's been a godsend. And why do you need a travel agent anymore? Anybody else have some cool stories to tell?
Wendy Hirsch 5:15
So I have a good one, which I think I shared with you all before I uploaded a picture of my closet, and I asked it to give you recommendations and how to improve the organization of my closet. Actually thought it was going to be just sort of random, generic, but it actually interpreted elements of the visual like you have stacks of clothes that are toppling over, maybe you don't make them so high. It actually like looked and reviewed the things in the picture and gave me suggestions. That was my personal use that is
Lynne Oldham 5:46
very incredible. Who knew it could do that? I'm
going to try it right after anybody else. I would
Diana Blancone 5:54
say, for me, there's some basics, bare basics, that I've been using it for, almost as a search engine, just to get some some information that I would typically go to Google, etc, for, I have also asked it how to assemble certain lovely items that arrive in the mail that don't come with decent instructions. But also just the basics of editing some emails just for ease and for grammar and using it in
Lynne Oldham 6:23
that
Sai Patel 6:25
way cool, similar to yours. Vacation Planning. Have two young boys. We're looking at going to Colorado and South Dakota for the spring break, and it's helping plan the trip. I even tells you, if it's too cold or snow, do this or do this, and even incorporates part of the weather in there. So it's amazing. The fun one,
Tracy St.Dic 6:49
I was gonna say, the fun one that my team has been doing recently is asking chat GPT based on everything you know about me. So everything that we've asked it before create an image of what you think my life looks like right now. I really encourage people to do it. It can be a big reflection moment for you and what chat GPT thinks your life looks like, for better or for worse, you always
Wendy Hirsch 7:08
have to be prepared for the feedback, right? Yeah.
Amy Reichanadter 7:14
I've also used it a lot for communications with my team, like I wanted to write a note to them of gratitude for Thanksgiving. And I put in a prompt with some specifics, and it came up with something that was, you know, 99% on the mark. So it's going to be really, really useful in that way,
Lynne Oldham 7:30
excellent. I think we're going to learn that as we talk. We now we deep, deep dig into business, because I think the better the prompt, the the more realistic what comes out the other end. All right, so let's, let's talk now. Let's get into the work. Let's share specific examples of how our org because I really want to talk like about, you know, all this AI is fun and interesting and sometimes theoretical, but we're all using it in our day to day roles and organizations. So let's talk about specific examples where our orgs are using AI in either talent acquisition and or talent management. And let's, let's, let's bring those to life for the audience who wants to start,
Diana Blancone 8:18
can start. I think the funny thing is, our organization has actually been in the AI space for over 15 years, through our agency, an elect, and they've been going at this for a while, using their suite of tools. I think for us in the people space, we are being very thoughtful about how we're implementing and we're moving pretty slow, but it doesn't mean that we don't have a plan. So we are. We're currently piloting AI for resume screening, so we're looking at algorithms to match candidate skills and experience to our job descriptions. We're hoping that this helps with ease of process, speeding up the filtering process, taking some of the tension off of the team, and it improves the quality of life for both the talent team, but also to improve the candidate experience for the candidate, ensuring that there's all these thoughtful touch points throughout the process where they're being followed up with, they're being spoken to, they're being they're aligned to jobs that actually fit their skill set. So there's definitely a lot of room to go here, because the data coming in, you want to ensure is pure and clean and accurate and but the other place that we've been using this is within talent management. So I'm sure we all do these global listening surveys. You we all have 1000s of employees, you know, we have 80,000 employees, and we do regular listening sessions and surveys. So we've been using AI to streamline the feedback, so in minutes, we can create an analysis that used to take days and days where you were evaluating all this survey feedback globally, and you could quickly get to market nuances, theme identification. And what we are doing is then interview. Meeting with company wide initiatives, whether those are employee experience initiatives, initiatives, rather or upskilling for certain individuals, if you see that there are certain pain points or outages, there excellent
Lynne Oldham 10:15
anybody using it similarly?
Wendy Hirsch 10:18
Yeah, at Eaton, we've done a number of things like, you want to start talk about time acquisition.
Sai Patel 10:23
So in time acquisition, similar to Diana, where you were talking about is, you know, matching resumes and matching candidates to jobs. I think one of the big cumbersome issues that people face is different career sites, different hierarchies, different ways of searching in different sides. So streamlining that process with candidates have the ability of uploading their resumes, and it goes and scans the entire organization and starts finding new positions that you may have never thought of. You know, it's very common for a state Hill engineering might be under a hierarchy engineering. They could be a sub group that might have an engineering department that a candidate may never look for, and this will then parse it out for them in that level. It also helps as far as locations and where you're looking for through it, the area that we're now starting to tap into and taking to the next level is, how do we do scheduling? We're done scheduling on a one on one with a regular basis, so that, I think there's tools out there that work in a great but incorporating AI when you now start going into panel scheduling. So if you've got a group of 456, people, you're trying to schedule and work through that, how can AI help and assist with that scheduling component? So that's what's on our roadmap.
Tracy St.Dic 11:34
Yeah, yeah, that's a tough one. Sai, I would co sign that we're trying to figure that one out too. I can share a couple of examples of how we're using it at Zapier so and also, just want to emphasize what was said at the beginning of this panel is having that human in the loop is so critical. So we're really thinking about how we keep that human in the loop for accountability and oversight with all of these processes we're doing, but for talent acquisition, specifically, we're increasingly using AI to review candidate profiles like si was mentioning and specifically to compose our outreach and responses. So our sourcing team created essentially AI assistance, which we did through Zapier Central. And those can be personalized with their own writing style, their own voice. And so it can read the entire LinkedIn page or even read a transcript of a podcast or webinar that a candidate was featured in, and communicate key points in seconds, and communicate that in the voice of the recruiter or the sourcer who would be sending that email. So that one has been really interesting. We've also experimented with our interviewers and our senior leaders do this in particular, given their stretch for time is they've created a final interview assistant which reads the entire candidate profile in our ATS, it can read the score cards, it can read all of the notes, and then it helps generate questions that that senior interviewer should ask in the final interview in order to fill in the gaps or address any lingering concerns from throughout the whole process. So those are two in talent acquisition. Outside of that, we've also used chat bots to support writing performance reviews, and specifically to help managers and ICS align their evidence with our performance management framework language, so not just a summary of evidence, but adapt that to how we think about performance management at Zapier and the most recent one that's been really interesting is we started a pilot with a platform that provides summaries of meetings when we have meetings with direct reports, but it also provides coaching tips and coaching feedbacks to the manager after that meeting to say, hey, here are some things you could do better. Here are some things that you did really well in this one on one meeting you had with your direct report. And so that one has been new, but also very interesting for us.
Lynne Oldham 13:39
Amy. Amy, before you go, I just want to ask Tracy. One quick question. Did you do candidate surveys before you started all that? And if you did, are they improving? Are people feeling touched or rather than ghosted with with this outreach that you're doing?
Tracy St.Dic 13:55
Yeah, our candidate survey, we have been doing candidate surveys before and after. I would say our candidate, MPs and CSAT has steadily increased. I don't know if it's specifically due to the personalization, but candidates often note that Zapier does feel like a very personalized experience. We don't ghost candidates. We commit to getting back to them every seven days, and so the AI helps us to do that and helps us remain efficient to that point. So I think it contributes, for sure. Gotcha.
Lynne Oldham 14:22
Sorry about that. Amy, go ahead.
Amy Reichanadter 14:24
Yeah, no problem at all. We have used AI to actually try to increase the interlock between talent management and talent acquisition. So one of the use cases that we're working on now is actually a predictive model around attrition, so we can predict actually down to pretty much the person now how many exits we'll have per quarter, and then we can feed that information back on a monthly basis into the town acquisition team so that they have more insight into exactly where they're going to need to hire to in order to meet their targets for the quarter. And so that's been one area where we've seen a lot of effectiveness. Because ta used to be kind of a black box around how exactly they would hit their target, given how much movement there would be and the employee population. And so we've been able to close that gap through using this predictive analytic model around attrition. For us,
Lynne Oldham 15:14
you see that
Wendy Hirsch 15:18
just go off of something Diana said and something Tracy, that you had articulated, I know you talked about starting small, right, not going slow, but being methodical. And Tracy, your comment about human in the loop, we recently deployed an AI assistant for manager performance reviews, so it's embedded within our systems will actually pull in employee goals, their accomplishments, their self assessment. And then it allows you to be able to add in manager feedback, but you can trace it to your point, in your own voice, add the tuning instructions to have it come out. One to be direct, supportive, creative, etc. But behind the scenes, it's tied to our leadership model, so it provides the feedback relative to our leadership competencies to be able to provide that input. You also use the phrase human in the loop. We have lots of folks say, Well, why can't it automatically populate in our system with the review that we want the human in the loop, like the managers actually intervene and copy in case and engage versus kind of just having to be fully automated, because we think that human element is really important.
Lynne Oldham 16:29
Yeah, definitely, definitely. Ai, Amy, back to your your predictive pieces. I do, I do think that's fascinating. Do you think besides the idea that you can staff up your TA organization appropriately based on what's coming. How do you think about maybe interventions that you you're got to do in your teams where you see those numbers coming? Yeah,
Amy Reichanadter 16:55
I mean, we've used it on both sides. So in terms of non regretted attrition, that's something that we wanted to keep at a healthy level in our organization, because we've been growing, you know, extensively for the last couple of years. So we have over 60% year over year growth, and the employee population has been growing at about the same pace. So for us, we want to make sure that we have a healthy level of performance management, and when we see that number not at the place we think it should be, then we can work more proactively to ensure that we're helping the managers do the right thing, to ensure that they're always up leveling the talent on their team, and then on the regretted attrition side, that we can predict pretty accurately where we think we have risk in the organization, and can also work with those leaders to ensure that they are being proactive around ensuring retention of the key talent that we want to keep in the organization.
Lynne Oldham 17:48
That's cool, because if your number predicts something and you can make it lower, that's even even better, right? So let's talk about TA in specific, because I know our audience is thinking, Okay, this is all great. AI is looking at resumes. AI is summarizing your interviews. They're even giving it interview questions. All great, but how are we ensuring that we AI reduces bias in the hiring process? How are we leveraging AI to reduce bias? Because that's, I think one thing. When at stash, I announced that we were going to use a new tool, it was AI based I read on blind how nervous people were about it in general. So talk to me about how your organizations think about that.
So the way we're
Sai Patel 18:41
been addressing that is by stripping up any personal information that can help identify so talking about names, genders, you get resumes from around the world. They put a private itself in their date of birth and some of the some of the countries. So it'll go through and strip anything that's personal that you can identify that individual, and that's what then goes to the recruiter. That's what goes to the hiring manager on first screening. And then, of course, by the time they're going to do a one one, they're going to start seeing everything else. But the first part of it is to remove anything they can, bring the highest in as much as we can. That's great. Yeah.
Lynne Oldham 19:20
Good. I
Diana Blancone 19:22
agree with everything that you said. Sai, we are focused on the same things. I think that we're trying to focus on skills and qualifications and specific certifications that are required for the role clear KPIs, and scraping all of those unique identifiers that could bring bias. And then I think a lot of this is just monitoring and auditing and tracking, right, like as we talked about, starting slow and building out. A lot of this is just ensuring that the data coming in is clean and that we can, we can make any changes as as we go. Yeah,
Lynne Oldham 19:57
yeah, definitely. I think. Uh, when people were doing the resumes we backed, when I was like, 2012 we were stripping the identifiers out for the people, part of it. But now, now we've got machine looking at this. So it's, it's, it's just as important, right? Tracy, what? What were you going to add? Yeah,
Tracy St.Dic 20:20
so very similar things to what Diana shared. So I won't repeat what she shared, because we do some similar stuff as well. But to add on to that, we actually use a tool that records and transcribes all of our interviews. And so some of the ways we've been leveraging that is this tool which is which is bright, higher, it generates candidate summaries based on the transcripts with concrete evidence versus, you know, interview impressions or bias that could be creeping in to the scorecards from the interviewer. And so that's been really helpful, because we've been able, as a recruiting team to advocate for candidates, or to ensure that we're going back to the concrete evidence that is there, what the candidate actually said, and really starting to remove some of the interviewer bias that may creep in when they say, Well, I did five interviews today, and I think I remember, I feel like the candidate, or the tone that I read from the candidate is we want to strip all of that away, and we want to go back to the evidence of what was actually said. And so this candidate platform has allowed us to really speed up those summaries and to be able to go back and use their chat bot to say, tell me about this specific skill, read through all of the transcripts from all of the interviews of that candidate and give us that concrete evidence. So we're really appreciating that as a way to stay evidence based. That's
Lynne Oldham 21:34
great. That's great because I think that's what we need in the world talk about now, because some of you did, but let's hone in on the candidate experience. How have you used AI to specifically enhance the candidate experience during the recruitment process? Because, like I said, I've heard all kinds of stories out there about people feeling ghosted. It's, it's, it's an employer's market again, how do we make sure, though, that the candidate feels the love?
Wendy Hirsch 22:07
So I can speak a little bit to our internal candidate experience, which has been a really big area as for us, around being able to identify open roles that might match with internal employees, with their skills and experiences, and you get immediate feedback. So there's an element of part of our process, in addition to the AI, is what we call internal candidate care. So that specific reach out to say, hey, Tracy, your skill set match with this open role. We'd like to have a conversation. We think this might be a good next experience for you. So I think in that sense, right lot of times we do this for a lot of our external candidates, where we'll have conversations, we'll talk about where there might be fit and not fit, but we often don't think about doing that for our internal candidates. So we've been also leveraging that AI and that matching of skills and experiences to do that better touch point with our own employees.
Lynne Oldham 23:01
So let me ask you, Wendy, do people have to raise their hand, or is it always trolling for opportunities for your employees?
Wendy Hirsch 23:09
Yeah. So if you've uploaded your profile into the system, it will be always looking and identifying opportunities. So I will get emails that will be sent. Hey, Wendy, it looks like your profile matches, this open opportunity will give you an opportunity to review it so it has great positives. I'll say I lost a great employee to the IT organization. They still set match because, wow, that would be a great next experience Bucha for me, but Great Britain,
Lynne Oldham 23:37
yeah, definitely, definitely, I think that's great. Is anybody else on the team doing something like that? Sounds like you're the 100 year old pioneer. I love it.
I love it. Wendy, all right, so,
Diana, what are you doing?
Diana Blancone 23:57
I mentioned earlier, we're pretty fixated on the candidate experience. We actually survey candidates throughout the process to see how we're doing at the recruitment and we survey our hiring managers to see how we're their experience is working with us on that front. So we're pretty focused on streamlining that. And I think, you know, we're all we're all human at the end of the day. And we have, you want to get back to everybody. You want everybody to feel heard and have a good experience with the organization. That's a huge priority, but you realize you can't do it all, so that's one of the things that we're focused on with. Ai, is to have that really manage the front end, where every candidate who applies is heard from to let them know, are you a match for the job? Are you not a match for the job? If you aren't, here's some other recommendations that we would like you to explore that might be a better fit for your background. So really focused on those personalized job recommendations based on their their profiles. I think in the distant future, we'd love to have AI, like some of you have said, provide real time feedback post interview, so taking those assessments and then providing feedback, it's, you know, so that you're removing the bias, but it's. All tied to clear KPIs and how you how you would deliver in the role based on the requirements that are clear in the job description. So it's a little bit about what we're we're focused on at the moment. That's great.
Lynne Oldham 25:12
Tracy, you talked about seven days and getting back to candidates in seven days, that's a phenomenal KPI. Tell me, though I know you, there's some things that you're not yet built out that are on the road map for candidate experience. Talk about those?
Tracy St.Dic 25:28
Yes, well, we're really obsessed with candidate experience too, at Zapier, and two of the things that we saw come out from our surveys over the past year in 2024 is that one candidates had a lot of questions about the teams of Zapier, how things were set up, and so they were emailing recruiters all the time, and that was taking a lot of recruiter time to answer them, but this was information that they needed to feel prepared. That was the second thing is that candidates noted, I'd like to feel more prepared for my interviews, because it's such a rigorous process. And so one of the things on our roadmap is to build out department specific pages through Zapier where candidates can learn about a team, learn about the team that's hiring, and ideally utilize a chat bot to ask all the questions that you might ask a hiring manager or recruiter. So that would be on 24/7 they can go in anytime they can ask about what do I need to prepare for my next interview? They'll be able to see, you know, here's the general questions that we ask at each interview stage so it would allow the candidates to self serve a lot more and also, honestly, give a lot of time back to our recruiters. Yeah, that's awesome.
Lynne Oldham 26:29
I think recruiters are have to learn how to use all this too, but to give them, you know, qualitative time back, that's great. Let's turn our attention toward like there's our audiences is some of them haven't dipped their toe in the water yet, right? And probably because some of the things we're going to talk about next, but I really do want to give them a sense of how you think about the challenges that your organization has faced when you brought AI into your talent acquisition or your talent management systems and how you've overcome them. And I think this is a really important topic, because, like I said, if they folks haven't dipped their toe in the water, they're thinking through all this stuff. And maybe this is like kind of the big black box that's scaring folks. So maybe talk about what you've done in your journey to implement AI
Wendy Hirsch 27:22
so I'm happy to start because we are the big Buju in all around the world. So we had many a struggle. But what I will tell you, hopefully you all can hear me. I got a feeling that the audio was not as good. Thumbs up, everybody. Can you guys see? Well, yeah, okay, so we've had to go through a number of groups, but the positive is we got through all of them. So it is doable, although challenging, because we've had to go through whether it's the EU AI act, whether it's the Department of Justice, in your guidelines, our own AI guidelines, our own diversity and inclusion guidelines, data privacy, data security. But I would say the thing that allowed us to get over all of those hurdles was the really, really close partnership between HR, IT legal and data privacy, and coming together as a team and really working together to overcome each of those steps, and we would not have been able to do it otherwise. And even with the AI assistant I described before for manager performance reviews, I think it is approved in every country around the world, except for one, which is a phenomenal achievement. And so I would say the biggest challenges are overcoming some of those regulations and guidelines that might be country by country, but to partner with your other internal peers, and it's definitely available.
Lynne Oldham 28:52
Thank you. Wendy,
Amy Reichanadter 28:55
one of the elements we've been really thoughtful about is just the change management within our own organization, and also just resetting the expectations around openness and curiosity, around the use of AI. And so one of the things that we've done is we've actually included that now in all of our job ladders around what the expected competencies are around the use of data and AI in each particular role, just to help our team understand that this is the future, and we want them to be really comfortable with with what's coming. And so I think that this revolution that we're in, this allowed us all to come back to, like, the core basics about change management, which is having a clear vision, the communication, the expectation setting, and also helping to upskill the teams as needed in order to help them embrace the changes that are coming through the technology. Yeah,
Lynne Oldham 29:43
I think transparency is key. If you're not transparent, even with the fears that are out there, by talking about it and putting it in a job description, you're really giving people an open book, which I love. Yeah. There
Wendy Hirsch 30:00
was a statement that was made. There was a statement that was made in one of the earlier sessions today that I wrote down, that hit on it, that said people fear poorly managed change, right, that the care that is coming from a lot of these AI is not managing it properly and communicating in that degree of transparency. I think that's probably one of the easiest hurdles to overcome, because we're aware of it and just over communicate.
Lynne Oldham 30:26
Yeah, yeah. I'm not sure what the it's a it's a great point. Wendy, I'm not sure what the impediment to being more open and transparent is. Are we assuming people know and understand it, I don't know, but I do think it's a it's not what Amy says everywhere, right there. Companies aren't as good at it as some are better than others. Let's, let's put it that way, Tracy, talk about what you've done at Xavier, because I know you, you also believe in the transparency piece.
Tracy St.Dic 30:59
Yes, yeah, transparency is one of our core values. So I think that it's really important to share who has access to the data and how we're doing the prompts specifically. So giving, you know, an example of we use a manager or a performance management chat bot that helps our managers and our ICS create their performance reviews, and it's really important that we actually publish those exact prompts publicly so you can see, what are we asking the bot to do, and specifically, what are we not asking the bot to do. We would not ask the bot to help give a rating or actually make an evaluation of someone's performance. We're asking the bot to help us summarize information in alignment with our performance management framework, and then the human, the manager, or the person writing the review, is critical to ensure that things don't go out the door without human oversight. And so we're very clear that AI doesn't make decisions. Ai assists humans to make decisions. And by publishing all of that to the whole company, the company can, you know, any individual in the company can read the prompt and say, Okay, I know what this is asking it to do. I can get feedback about that. I can copy it and modify it for myself. So that transparency piece is really key. But I think that also leads into a challenge that Wendy mentioned earlier, is the implementation and setup has been one of our bigger challenges at Zapier like the implementation can be time consuming. It can take multiple attempts to get things right. I think that manager bot that I just mentioned, it took about 27 tries to get the prompt right and then be able to say, this is a place where it feels fair, it feels as unbiased as possible. Now we're going to publish it to the company and iterate it from there.
Lynne Oldham 32:37
So AI implementation is not for the faint of heart, yes. How about you? Diana,
Diana Blancone 32:46
I agree with everything everybody has said. I think one of the other things that we're concerned about is like, again, that data coming in, and the taxonomy around that data. So when we think about some of our partners, our bigger partners, let's say not to call out LinkedIn or or glass door, but if you have that easy apply feature, what is that information that we're collecting from? Is it fruitful enough? Is there enough in there to minimize bias? Do you have enough information to work with to parse them and match them to the appropriate role? So I think we all have an opportunity to work with, you know, our partners, and giving them feedback too, on what we need for this to be successful and as clean as possible at the end of the day.
Lynne Oldham 33:24
Love it. Love it. One of the things that did I miss somebody? Did somebody say something? Yeah, okay.
Sai Patel 33:33
One other item in there, and not specifically, just around town, acquisition, time management, but would highly recommend that, you know, look at it from a very macro view. There is AI everywhere in the enterprise, and it's shiny, right? There's not a vendor you're not going to talk to that doesn't have AI, because you got to keep end in mind to it. Because what's going to happen, and this is very high on our agenda, is, where and how do we implement AI and what's that final experience? But what we don't want is two years later, we've got 682 different AI is out there, and nobody's using because you don't know where, how things are moving around through it. So just as important as how are the AI, how the AI is talking to each other, how does that bring in the experience there? So you know, when having those discussions, look at it not only in that bucket, but think of where are you going to take this, a year down, two years down, three years down, and then create that road map along with it.
Lynne Oldham 34:28
Love it. Love it. Let's talk about our teams, and when we think about the roles in our own teams and how they're going to have to adapt. Obviously, we just talked about organization adaptation, but now think about our team members, right? I briefly said something about how recruiters their roles are changing. Many of our roles are changing. Given the advent of AI talk about when you brought a process in, how that's maybe. Changed the folks that were doing the role before and how you've handled it,
Amy Reichanadter 35:07
I can share an example. So when we first introduced our HR bot to answer employee questions, we used to have a team that did that work. And at first, of course, there was a lot of anxiety about the fact that their role was going to fundamentally change. But what we've seen over time is that having that self service bot available for employees allowed two changes to happen. Some of the team members that used to do that work are now answering kind of higher order questions, so their work got more interesting and more complex, and then other members of that team have moved on to project management roles, so that they've been able to elevate their role and getting out of the day to day administrative work of answering the same questions over and over again, and so now they're doing, you know, more advanced and sophisticated work than they were doing before. So we have seen a lot of positive reaction to that transformation. But at first, I think there was some anxiety about the fact that change was coming, and they didn't know what it would look like on the other side, but it's overall, it's been a very positive experience for our team.
Lynne Oldham 36:05
That's great. That's fantastic. Anybody else has had their team react, and what have you done about it?
Diana Blancone 36:16
You know, I think the general sentiment here is that this AI is going to do replace everybody, and there's going to be no jobs for people. And we all know that that's not true. We've said time and time again throughout this chat that there's the human touch is so important in all of this, I think. And we've probably all said this time and time again, is empowering our people to do less administrative tasks, more strategic conversation, so allowing our people partners to have those strategic conversations based on the information that we're collecting from AI, and think when we talk about even performance reviews, those frequent touch points. So, you know, it's a heavy lift to do these at times, but if we can utilize AI to streamline some of that and take out some of the administrative work, we're able to then have these thoughtful conversations with our employees, and hopefully at the end of the day, it helps with retention and setting people up for success, because you're able to intervene and say, Okay, well, Diane is having a challenge with x. Let's intervene swiftly through the data that we're pulling and give her the tools that she needs to succeed. You know, at Omnicom or wherever, that's
Wendy Hirsch 37:15
great. We have at Eaton, sort of a hashtag phrase we've been using for all of HR, which is hashtag make my life easier. And everything that we deploy, everything we communicate, we start our monthly deployment calls that are global with a tip or on hashtag make my life easier. So to size point about like, is it just a shiny logic, or is there actually an objective? One of the objective is it needs to make our lives easier. And I think that our HR teams are starting to see that it is, in fact, making their life easier. So I hope even when they see that on the agenda, there's a hashtag, make my life easier. Tip, I'm going to join and listen in. I want to know what that is, because it's going to help me in my job. I think part of that bringing people, you know, we talk about crawl, walk, run, we crawl, we try little nuggets, then we start to walk, and then we run. And I think the Rhr teams are starting to see that, and to see the value of each of these little bits of AI that they can take advantage of.
Lynne Oldham 38:17
It harkens me back Wendy, to that big red button from Staples. Remember, that was easy. I'm going to take us to the last question, because we really have gotten three Chris was right. This goes fast. We have three minutes left. I'm going to sort of combine the last question with asking you for a nugget. So in 60 seconds. Let's go around the horn. What future trends in AI is your org preparing for? That's the first question. And the second question is, what parting piece of advice would you leave this audience with regarding AI? So if you take more than 60 I promise I won't pull you off stage, but, but try to keep it tight. Let's, let's start with, let's see I'm going to start with Tracy. Oh
Tracy St.Dic 39:06
gosh, okay, 60 seconds. Um, so I would say one of the the trends that we're preparing for is just, how can we have aI help us with real time, anytime, support, whether it's the help desk that's always on, the chat bots that are always on, how can people self serve more to get what they need to be able to unlock the next level of productivity for themselves and do the things that humans really do best. So I think that's going to be a big trend in how we do that. And then the nugget of advice, one thing I've been thinking about lately is to stop, and we did this with my team, is to take a step back and stop thinking about what tech is currently out there. People always say AI is the worst now that it will ever be, and think about the future philosophically, of where you think AI can solve the problems that you have in your business, not so much of what's the cool tool we can implement, but what is the problem we're trying to solve, and what's the pain point you. And then could we use AI to solve that? And going from that angle, versus starting with this is a very cool tool that can help us save time. We'll actually ensure that you have a vision for how you're utilizing AI in your work.
Lynne Oldham 40:12
I love that, because tools are everywhere. We really need to start with the problem. Let's see. Let's go with Diana. Hey,
Diana Blancone 40:19
maybe I'll start with my nugget. I think my nugget, if it didn't come across clear, is that, you know, don't rush it. I think this is such a big movement for for all of us, and we do have time. I think just being thoughtful and always putting our ethics over everything, not sacrificing quality for speed, right? And just trying things out, doing a bit of a trial, maybe don't start with your entire organization. But in our case, we can start with a smaller agency and try things out. Start with a smaller pool of candidates and see how this this performs, I think, for future we're preparing for, you know, predictive analytics to support workforce planning and look at deeper personalization into our employee training, our upskilling. What tools can we use to foster that collaboration? I know we still have a decent amount of hybrid employees. We're operating in a hybrid way, but how can we function better in hybrid work environments while using the AI tools as well?
Lynne Oldham 41:13
Love it. Love it. Let's go to AI.
Amy Reichanadter 41:17
All right. I think the thing that really strikes me about AI is, I think there's an incredible opportunity for HR leaders to operate at so many different levels around this transformation, and that means thinking about the impact at the company strategy level, your leadership readiness, what talent you're going to need in your organization, and then how quickly you can move these pilots, as Diana was saying into production. So I think, you know, my advice is to really act boldly and to think outside of the specific use cases, but also to think at all the different levels of how this can impact and help grow your specific organization. Beautiful.
Lynne Oldham 41:54
And then I'll let Wendy inside you. You decide who goes first, pull up, pull up. The question, yes, I'll
Wendy Hirsch 42:04
combine it to say. What's on the horizon for us is AI conversational around HR reporting and analytics. So being able to talk, where do I have the highest turnover? How does it differ by gender, and do it in a very conversational way? Our roll out of that would be my advice, which is, and it goes to what Diana said, which is, start small, crawl, walk, run would be the biggest advice that I would give, that you don't want to do too much, too fast, start small, iterate the agile approach.
Sai Patel 42:38
Love it
Lynne Oldham 42:40
all right, Chris, I think that's a wrap for us. There you
Chris Rainey 42:42
go. I told you, it go fast. Lynn, thank you for having us. No, it's all right. The time flew by. It was an amazing conversation, and it's going to be hard for you, but I'm going to ask you one last question to answer your own question. You took away, Lynn, and then we'll say goodbye. I
Lynne Oldham 43:01
think I echo everything that the team said. Basically, start anywhere but start small, experiment. Work with your your cheerleaders in the business, to get a pilot going and then let them share. It's always better coming from the business I have, I think I have great ideas, but I love when the business is the champion of my ideas, as opposed to me.
Chris Rainey 43:28
It makes a lot easier when you when it comes to decision making and implementation, right? But listen, uh, thank you so much for moderate. You did an incredible job. Lynn, I appreciate you, and thank you to all of our amazing panelists for being with us and sharing your insights and perspectives. Thank you again to our panelists. You.
Lynne Oldham, Chief People Officer at Dataiku + 5 special guests.