Ep 23: The Impact of AI on Contingent Workforce and Talent Transformation with Jeff Mike

[00:00:09] [SPEAKER_03]: Hey everyone, it's Bob Pover.

[00:00:11] [SPEAKER_03]: In this episode of LVature AIQ, I talked with Jeff Mike, managing director at FlexTrack about AI in

[00:00:17] [SPEAKER_03]: Talent Acquisition using a broad lens that encompasses all talent pools.

[00:00:21] [SPEAKER_03]: It is because the practical applications of AI, how AI tools are being integrated into

[00:00:26] [SPEAKER_03]: FlexTrack's next-generation vendor management system and the role of data privacy and trust

[00:00:31] [SPEAKER_03]: in AI's future.

[00:00:33] [SPEAKER_03]: To explore how AI is augmenting human capabilities in Talent Acquisition and beyond,

[00:00:37] [SPEAKER_03]: I can leverage these advancements to elevate your own AIQ.

[00:00:41] [SPEAKER_03]: I really enjoyed this chat with Jeff as I do all my conversations with him, and I hope you

[00:00:45] [SPEAKER_03]: find it insightful.

[00:00:46] [SPEAKER_03]: Thanks for listening.

[00:00:48] [SPEAKER_03]: Hi everyone, welcome to another edition of Elevate Your AIQ.

[00:00:52] [SPEAKER_03]: I'm your host Bob Pover.

[00:00:53] [SPEAKER_03]: With me today is my friend Jeff Mike.

[00:00:56] [SPEAKER_03]: I do not Jeff.

[00:00:57] [SPEAKER_03]: Do well Bob, thanks.

[00:00:58] [SPEAKER_03]: How are you?

[00:00:59] [SPEAKER_03]: Good.

[00:01:00] [SPEAKER_03]: Thanks for being here.

[00:01:01] [SPEAKER_03]: Just to start things off, don't you give the listeners a little bit of background

[00:01:06] [SPEAKER_03]: to that.

[00:01:07] [SPEAKER_03]: How you got to your current role at FlexTrack?

[00:01:10] [SPEAKER_04]: Sure, thanks Bob.

[00:01:11] [SPEAKER_04]: So my role is managing director of customer strategy and value at FlexTrack.

[00:01:16] [SPEAKER_04]: FlexTrack is an ex-generation vendor management system.

[00:01:20] [SPEAKER_04]: My role is focused on working with our sales teams, our product teams, and our customer

[00:01:25] [SPEAKER_04]: success teams to ensure that we're identifying and tracking towards customer business outcomes.

[00:01:32] [SPEAKER_04]: We deliver the required capabilities.

[00:01:33] [SPEAKER_04]: And then this is important, making sure we're measuring progress or achievement of those

[00:01:39] [SPEAKER_04]: business outcomes.

[00:01:40] [SPEAKER_04]: By background, I have a long history in talent acquisition, interestingly enough, started

[00:01:45] [SPEAKER_04]: in staffing, went over to HR for about 17 or 18 years, focused mainly on talent acquisition,

[00:01:51] [SPEAKER_04]: but also HR leadership.

[00:01:53] [SPEAKER_04]: And then I spent about five years at person and the loites doing HR research and leading

[00:02:00] [SPEAKER_04]: the research team for Deloitte's Human Capital Practice.

[00:02:02] [SPEAKER_04]: I've been with FlexTrack for about two and a half years.

[00:02:06] [SPEAKER_04]: In addition to the description that I gave previously, I also focused on partnerships

[00:02:11] [SPEAKER_04]: with our global systems implementers and then also our technology alliance partners.

[00:02:18] [SPEAKER_03]: Thanks, Lord.

[00:02:19] [SPEAKER_03]: So the way you describe which your buildings found impressive and the opposite of a lot

[00:02:28] [SPEAKER_03]: of shiny objects that we see today in talent acquisition, especially as it relates

[00:02:34] [SPEAKER_03]: to AI and looking at real value creation, as opposed to productivity and some of the things

[00:02:42] [SPEAKER_03]: we see that are labeled as AI, but they're really automation.

[00:02:46] [SPEAKER_03]: And some of that stuff has been around for, as you know, probably a decade.

[00:02:52] [SPEAKER_03]: As you think about the contingent workforce space, how AI helps that and perhaps connect it

[00:03:01] [SPEAKER_03]: to some of the other talent acquisition strategies.

[00:03:06] [SPEAKER_03]: How do you think about that space and how FlexTrack is kind of incorporate that?

[00:03:11] [SPEAKER_04]: Yeah, that's a great question and it's a good place to start because I think with all

[00:03:19] [SPEAKER_04]: what we call AI, the community interesting comment.

[00:03:22] [SPEAKER_04]: We AI that we use is very much machine learning.

[00:03:26] [SPEAKER_04]: AI in terms of true intelligence is still a few years off, but since that's what the industry

[00:03:30] [SPEAKER_04]: is using, that's what everyone's talking about, that's where we go.

[00:03:33] [SPEAKER_04]: And so really what's important is to start with the use cases.

[00:03:37] [SPEAKER_04]: So getting through this kind of hype, all the noise around it, you know, what's been important

[00:03:42] [SPEAKER_04]: for us and what's been important for our customers is that we start with practically

[00:03:46] [SPEAKER_04]: use cases.

[00:03:47] [SPEAKER_04]: What is in need that the organizations or the customers have and how can the AI tools both

[00:03:52] [SPEAKER_04]: predictive and generative address those use cases?

[00:03:56] [SPEAKER_03]: And so that's typically where we start.

[00:03:58] [SPEAKER_03]: I was going to ask like, I guess maybe just expand a little bit about or how you're

[00:04:02] [SPEAKER_03]: integrated on a built on a Salesforce platform.

[00:04:05] [SPEAKER_03]: Obviously they've been raised, you know, AI as well.

[00:04:09] [SPEAKER_04]: So FlexTrack is built on the Salesforce Lightning platform.

[00:04:11] [SPEAKER_04]: And what that allows us to do is become much more of a platform as a service rather

[00:04:16] [SPEAKER_04]: than the software as a service model, that the industry has known for a couple of decades.

[00:04:22] [SPEAKER_04]: And specifically that allows us to escape some of the technical debt that the legacy

[00:04:26] [SPEAKER_04]: systems have accumulated over the years and allows us to bring Salesforce investment in

[00:04:33] [SPEAKER_04]: those platform tools directly to our customers.

[00:04:35] [SPEAKER_04]: So Salesforce invest billions of dollars a year and things like user experience, security,

[00:04:40] [SPEAKER_04]: global region of course, AI.

[00:04:43] [SPEAKER_04]: And specifically what our customers seem to get most excited about in terms of Salesforce

[00:04:47] [SPEAKER_04]: tools or the capabilities that being built on Salesforce allows us to bring is an integration

[00:04:53] [SPEAKER_04]: studio.

[00:04:53] [SPEAKER_04]: So it's a very robust studio of tools designed to make integration and opportunity

[00:04:58] [SPEAKER_04]: rather than a hassle.

[00:05:00] [SPEAKER_04]: So putting FlexTrack into your tech stack is really much more simple and streamlined

[00:05:06] [SPEAKER_04]: than it has been before because of, you know, different types of APIs.

[00:05:11] [SPEAKER_04]: We have certified connectors and then there's the app store with 7,000 or more business

[00:05:16] [SPEAKER_04]: enabling apps that are available right on the platform.

[00:05:19] [SPEAKER_04]: A couple of other things include Flow Builder which is a graphic or drag and drop workflow

[00:05:24] [SPEAKER_04]: tool which allows for a lot more flexibility and agility.

[00:05:28] [SPEAKER_04]: We offer on platform analytics and AI which is really exciting as well.

[00:05:33] [SPEAKER_04]: So Salesforce says you may know purchase Tableau a few years ago and has integrated

[00:05:39] [SPEAKER_04]: data visualization and data management tools of Tableau into the platform analytics capabilities

[00:05:46] [SPEAKER_04]: and AI capabilities.

[00:05:48] [SPEAKER_04]: So real briefly that allows for machine learning.

[00:05:51] [SPEAKER_04]: It's called Einstein AI.

[00:05:53] [SPEAKER_04]: It has data modeling tools, data cleaning tools and really allows us to build, you know,

[00:05:58] [SPEAKER_04]: data models on really complex sets of data from the program without meeting data science

[00:06:03] [SPEAKER_04]: expertise.

[00:06:04] [SPEAKER_04]: You can actually ask questions and set up models in natural language questions and once you

[00:06:10] [SPEAKER_04]: have the data identified, the platform data tools and AI will execute the data model

[00:06:17] [SPEAKER_04]: and the analysis and give you suggestions.

[00:06:20] [SPEAKER_04]: So one quick use case there is for example onboarding data.

[00:06:24] [SPEAKER_04]: You might have a large complex program because of our case functionality again which comes

[00:06:28] [SPEAKER_04]: with the Salesforce platform, we create a case for every onboarding situation.

[00:06:34] [SPEAKER_04]: And that case is workflow through that flow builder for specifics of whatever segment geographically

[00:06:40] [SPEAKER_04]: labor category or even business unit is required.

[00:06:44] [SPEAKER_04]: Each step can then be automated, audited and tracked and reported on to keep track of every

[00:06:50] [SPEAKER_04]: step in the onboarding process.

[00:06:52] [SPEAKER_04]: That allows the Einstein AI to create a model of where are some of the bottlenecks for

[00:06:57] [SPEAKER_04]: example that you might experience in onboarding.

[00:07:01] [SPEAKER_04]: How can we identify those and what can we do to address some of those bottlenecks?

[00:07:06] [SPEAKER_04]: So that's one of the use cases that we have.

[00:07:08] [SPEAKER_03]: You've basically got embedded for a process mining, baked into that platform with that

[00:07:12] [SPEAKER_03]: particular use case which sounds pretty powerful, right?

[00:07:15] [SPEAKER_03]: Like where are the points of friction that may not be evident if you look at it from

[00:07:21] [SPEAKER_03]: to reverse I view but to be able to go in at individual points, you know if you just sort

[00:07:28] [SPEAKER_03]: of follow it along that process map or that workflow, be able to have the analytics

[00:07:33] [SPEAKER_03]: step by step.

[00:07:34] [SPEAKER_03]: You're talking about improving experiences at the end of the day and identifying other opportunities

[00:07:40] [SPEAKER_03]: for automation or where to apply AI.

[00:07:43] [SPEAKER_03]: So I think that's an excellent use case and a lot of robust capability because otherwise

[00:07:49] [SPEAKER_03]: you got to bring in another vendor or maybe updating myself but I remember what process

[00:07:56] [SPEAKER_03]: mining had to be a separate and discrete implementation to go and basically be your

[00:08:02] [SPEAKER_03]: pick an MRI of your process.

[00:08:04] [SPEAKER_03]: The fact that that's embedded now, I wasn't even aware of that so that's interesting.

[00:08:10] [SPEAKER_04]: As you may know there's predictive and there's generative AI which are the most common,

[00:08:13] [SPEAKER_04]: predictive is using the copious amounts of data and programs or whatever the application

[00:08:19] [SPEAKER_04]: is to identify patterns and trends and make predictions and prescriptions.

[00:08:23] [SPEAKER_04]: Generative is really more about creating new content and since the release of chat, GPT,

[00:08:30] [SPEAKER_04]: year and half or so ago there's been a lot of hype around generative AI.

[00:08:33] [SPEAKER_04]: There is a lot of investment in generative AI from Salesforce but we take a little bit

[00:08:38] [SPEAKER_04]: different approach because the customers have different preferences and are at different stages

[00:08:43] [SPEAKER_04]: of their AI journey.

[00:08:45] [SPEAKER_04]: So we provide access to generative AI through our integrations.

[00:08:50] [SPEAKER_04]: So if a customer is using chat, GPT, you know they're using copilot if they're using Google's

[00:08:57] [SPEAKER_04]: generative AI solution, we can integrate those instances directly into the platform

[00:09:02] [SPEAKER_04]: and maintain the relationship between the customer and maintain the relationship between

[00:09:07] [SPEAKER_04]: the provider with regards to access with regards to data protection, we simply integrate

[00:09:12] [SPEAKER_04]: it into the platform and into the program for the customers.

[00:09:17] [SPEAKER_02]: Hi there, I'm Peter Zolman.

[00:09:19] [SPEAKER_02]: I'm a co-host of the inside job boards and recruitment marketplaces podcast.

[00:09:23] [SPEAKER_01]: And I'm Stephen Rothberg and I guess that makes me the other co-host.

[00:09:26] [SPEAKER_02]: Every other week we're joined by guests from the world's leading job sites.

[00:09:30] [SPEAKER_01]: Together we analyze news about general niche and aggregator job board and recruitment

[00:09:35] [SPEAKER_01]: marketplaces sites.

[00:09:37] [SPEAKER_02]: Make sure you sign up and subscribe today.

[00:09:41] [SPEAKER_04]: So that's really designed to be part of a larger, maybe enterprise-wide generative AI strategy

[00:09:47] [SPEAKER_04]: versus the HCM or the HRIS or the ATS has one type of AI, the VMS has another type

[00:09:55] [SPEAKER_04]: of generative AI.

[00:09:56] [SPEAKER_04]: The analytics team, the finance team might have different types of AI.

[00:10:00] [SPEAKER_04]: We're designed to really fit where the customer is going, where they are and where they're

[00:10:03] [SPEAKER_04]: going with their generative AI approach.

[00:10:04] [SPEAKER_04]: And then from there we'll develop similar use cases.

[00:10:08] [SPEAKER_03]: A couple of things come to mind as you describe that because you have those specialized

[00:10:13] [SPEAKER_03]: use cases for generative AI and I'll just call them each of those, you know, an agent

[00:10:18] [SPEAKER_03]: for this sake of simplicity would you potentially provide user interface that access sort

[00:10:26] [SPEAKER_03]: of the, I guess I described it as like the general contractor which then goes and takes

[00:10:32] [SPEAKER_03]: that command or that request and then goes and knows which of those other agents to contact

[00:10:39] [SPEAKER_03]: based on the request and what it knows about, you know, the context and how that work

[00:10:44] [SPEAKER_03]: though exists.

[00:10:46] [SPEAKER_04]: Yeah, we could do that.

[00:10:47] [SPEAKER_04]: It really depends on, again, the broader corporate strategy but we've already done

[00:10:51] [SPEAKER_04]: some experimentation in our, in our call it our design lab.

[00:10:55] [SPEAKER_04]: We've done some voice interface to get started.

[00:10:59] [SPEAKER_04]: For example, we've had customers ask us about, you know, our managers sometimes are on a

[00:11:05] [SPEAKER_04]: shot floor or they're on a rig.

[00:11:06] [SPEAKER_04]: They're out in the field.

[00:11:08] [SPEAKER_04]: They're not able to access even a, even a keyboard so how can we have them start to do

[00:11:12] [SPEAKER_04]: some of their, you know, contingent workforce management responsibilities?

[00:11:16] [SPEAKER_04]: How can we have them perform them with a, with a chat interface or with a verbal command

[00:11:21] [SPEAKER_04]: interface?

[00:11:21] [SPEAKER_04]: So we've started experimenting with that.

[00:11:24] [SPEAKER_04]: We've done a couple of use cases there and we're going to continue to build that out.

[00:11:28] [SPEAKER_04]: Similar to the experience, this may be a little bit different but we also are creating

[00:11:33] [SPEAKER_04]: kind of like a Netflix or like an Amazon experience that kind of personalizes the experience

[00:11:37] [SPEAKER_04]: based on the persona and then ultimately the individual when they're working in the

[00:11:43] [SPEAKER_04]: extended workforce.

[00:11:44] [SPEAKER_04]: That might start with a requisition or a different type of, different type of contingent

[00:11:49] [SPEAKER_04]: worker that similar managers are accessing or have had similar success within the past

[00:11:54] [SPEAKER_04]: or based on the history of the, the individual hiring manager just to kind of help them

[00:12:00] [SPEAKER_04]: accelerate their ability to get things done.

[00:12:02] [SPEAKER_04]: And eventually that could lead into a more generalized interface either with the rest of

[00:12:06] [SPEAKER_04]: the BMS which is kind of already there or you know a broader interaction with, with maybe

[00:12:12] [SPEAKER_04]: HR systems or with broader corporate systems.

[00:12:15] [SPEAKER_04]: We haven't built those yet but we have, we have the connectors and we have the capabilities

[00:12:20] [SPEAKER_04]: to do it.

[00:12:21] [SPEAKER_03]: Nice.

[00:12:22] [SPEAKER_03]: The other thing I was going to bring up, but this came up in conversation with one of my

[00:12:26] [SPEAKER_03]: talent intelligence community.

[00:12:28] [SPEAKER_03]: We were talking about dashboards and so you mentioned tablo, which is part of Salesforce

[00:12:33] [SPEAKER_03]: now but one of the questions or the discussions was if you can ask a contextual question

[00:12:44] [SPEAKER_03]: about your data through that, you know, conversational interface with a device or text.

[00:12:51] [SPEAKER_03]: And it can generate you a graph, a chart, whatever on the fly.

[00:12:58] [SPEAKER_03]: We need all these dashboards because I feel like there's a lot of people that's all

[00:13:03] [SPEAKER_03]: they do.

[00:13:04] [SPEAKER_03]: Their job is to create new, request them dashboards for people but I mean it's like why do

[00:13:11] [SPEAKER_03]: you need a cockpit if all I want to know is how fast I'm going.

[00:13:15] [SPEAKER_03]: Seems like much, much more efficient, you know, fit for purpose approach to getting

[00:13:22] [SPEAKER_03]: a visual of what you're after.

[00:13:26] [SPEAKER_03]: So I'm not sure if any of your clients have talked about that or have you done any experimentation

[00:13:32] [SPEAKER_03]: with that just want to get your thoughts more generally as well.

[00:13:35] [SPEAKER_04]: We have the conversational interface connection, again there's lots of them out there

[00:13:41] [SPEAKER_04]: you can integrate them very quickly into flex track because of our architecture.

[00:13:46] [SPEAKER_04]: I think what you're describing though is we're in a transition period so take it in a couple

[00:13:50] [SPEAKER_04]: of steps having tablo on the platform, having essentially drag and drop report builders

[00:13:55] [SPEAKER_04]: having these data tools that really it accelerates the ability right now to be able to get

[00:14:02] [SPEAKER_04]: those graphs and charts and that contextual data more quickly, you know, you can do it

[00:14:07] [SPEAKER_04]: through, say point and click or drag and drop or whatever, you know, whatever works.

[00:14:13] [SPEAKER_04]: The challenge right now is you have a lot of people who are used to seeing dashboards

[00:14:16] [SPEAKER_04]: and aren't used to applying the capabilities that you're describing.

[00:14:22] [SPEAKER_04]: So the first step is really we're taking that process of building dashboards

[00:14:27] [SPEAKER_04]: and data visualization from let's say having a team of analysts in, you know, offshore team

[00:14:37] [SPEAKER_04]: monthly reports or from having to go to a third party system like you suggested to being

[00:14:43] [SPEAKER_04]: able to do it right on the platform. So someone in the PMO or the VMO or the MSP rather than

[00:14:49] [SPEAKER_04]: having to take a day or two to create the reports, those reports and charts and visualizations

[00:14:53] [SPEAKER_04]: are going to be available more quickly. So as people get used to seeing those more quickly

[00:14:58] [SPEAKER_04]: and get used to the more powerful tools that are available there will be an evolution towards

[00:15:03] [SPEAKER_04]: I think what you're talking about is being a little bit more specific. Well, let me use the

[00:15:08] [SPEAKER_04]: example that we've had. I got to decrease my onboarding time for this particular segment

[00:15:13] [SPEAKER_04]: I would say it's professional services. I'm losing revenue because I'm not able to onboard

[00:15:17] [SPEAKER_04]: my consultants and my contingent workers fast enough to deploy to client work. You could ask

[00:15:23] [SPEAKER_04]: the question you could drill through kind of the steps that we're taking, you know, use onboarding

[00:15:28] [SPEAKER_04]: data for this segment to help me understand or to point out how can I improve my onboarding

[00:15:33] [SPEAKER_04]: where are the bottlenecks and what should I do. So essentially those are the questions that we're

[00:15:38] [SPEAKER_04]: asking, we have to take a few steps you have to identify the data set, you have to make sure that

[00:15:43] [SPEAKER_04]: the natural language processing understands your question and can push that out. So kind of

[00:15:48] [SPEAKER_04]: a long way of saying I think we're taking steps there and the technology while the technology might be

[00:15:53] [SPEAKER_04]: capable now, you have also if the users up to where the technology can bring them and sometimes

[00:16:00] [SPEAKER_04]: that takes a little more time or it may take transitional steps to get there. So I think your vision is

[00:16:05] [SPEAKER_04]: correct, I think over time will get there and over time people will start to build their skills

[00:16:14] [SPEAKER_04]: in AI and all things data. Again, I think one of the real benefits of AI is it kind of

[00:16:19] [SPEAKER_04]: it reduces that data literacy challenge that has been so prevalent in terms of analytics

[00:16:25] [SPEAKER_04]: and particularly people analytics and HR and contingent workforce. It moves that forward very quickly.

[00:16:32] [SPEAKER_04]: So then the people working can be deployed to do other activities versus, you know, putting

[00:16:37] [SPEAKER_03]: reports together and creating visualizations. Now that makes sense and it will take time and it

[00:16:43] [SPEAKER_03]: will be sort of an evolution like you said. And you made an important point just a second ago around

[00:16:49] [SPEAKER_03]: the data literacy piece when we talk about, you know, upscaling and, you know, getting people

[00:16:57] [SPEAKER_03]: comfortable with AI in it's Bob Pover, host to you podcast human centric AI, AI driven

[00:17:06] [SPEAKER_03]: transformation hiring for skills of potential, dynamic workforce ecosystems responsible innovation.

[00:17:13] [SPEAKER_03]: These are some of the themes my expert guests and I chat about and we certainly geek out on the

[00:17:18] [SPEAKER_05]: thing to do technical. I hope you check it out. I want to take a break real quick just to let

[00:17:24] [SPEAKER_05]: you know about a new show we've just added to the network up next at work hosted by Jean

[00:17:31] [SPEAKER_05]: and Kate a keel of the Devin group and tastic show. If you're looking for something that pushes

[00:17:37] [SPEAKER_05]: the norm pushes the boundaries has some really spirited conversations, Google up next at work,

[00:17:45] [SPEAKER_05]: Jean and Kate a keel from the Devin group. In the current state,

[00:17:52] [SPEAKER_03]: much easier to upscale on generative AI because it's sort of its own, the tool itself can guide you

[00:18:00] [SPEAKER_03]: and tutor you just ask it how to use it more effectively or you can reframe your questions.

[00:18:06] [SPEAKER_03]: I mean it's very user friendly in that sense we're not talking about upscaling AI as an

[00:18:14] [SPEAKER_03]: predictive and operational AI machine learning natural language sense in terms of an

[00:18:20] [SPEAKER_03]: extension of your data science or data analysis kind of career trajectory. General Devin

[00:18:27] [SPEAKER_03]: AI is much different and much as a much simpler learning curve and as much more approachable than

[00:18:35] [SPEAKER_03]: a data career trajectory I guess. Right, right I would agree and that's where you probably

[00:18:40] [SPEAKER_04]: heard of prompt engineering really what we talk about in terms of generative AI and even if you

[00:18:45] [SPEAKER_04]: have a natural language interface with the predictive AI, you don't need to know the data science.

[00:18:50] [SPEAKER_04]: You need to know your program and you need to know which questions to ask and then you know

[00:18:55] [SPEAKER_04]: what I would start with you need to know what outcomes you're after, you need to know your program

[00:18:59] [SPEAKER_04]: and you need to know what questions you need to ask of the data or of the machine learning

[00:19:04] [SPEAKER_04]: AI of the data that's more important than spending weeks or years learning data science or

[00:19:11] [SPEAKER_03]: even statistics. Yeah absolutely so as you and I've talked about I mean that I've definitely

[00:19:17] [SPEAKER_03]: had a focus on responsible AI and that's not just a complying with legislation I mean that's being

[00:19:23] [SPEAKER_03]: responsible by design and when we think about upscaling and I think of using AI responsibly as

[00:19:32] [SPEAKER_03]: part of that upscaling I think it's important for people to understand enough about like where

[00:19:38] [SPEAKER_03]: the data is coming from but you do have to think critically about the source of the data because

[00:19:46] [SPEAKER_03]: if you're questioning well first of all the output of these generative AI tools it's not a calculator

[00:19:52] [SPEAKER_03]: right so it's not giving you a definitive this is 100% correct and of answer so you know for

[00:19:58] [SPEAKER_03]: what's familiar with the term hallucination because it doesn't really know what it's

[00:20:04] [SPEAKER_03]: irritating back to you but I want to make sure listeners understand how they're using it and whether

[00:20:11] [SPEAKER_04]: they can bust it into the day. I think that's a really good point and I didn't bring up critical thinking

[00:20:16] [SPEAKER_04]: but I think you're raising it is really important to along with that ability to generate the

[00:20:21] [SPEAKER_04]: white questions you have to know your program and your data like you mentioned

[00:20:26] [SPEAKER_04]: at least where the data is generated what it says what a cannon can't do and they have to apply

[00:20:31] [SPEAKER_04]: that critical thinking that's where the human in the loop I think is another term people have

[00:20:35] [SPEAKER_04]: probably heard a lot where you have to think critically about the results that the AI puts out

[00:20:40] [SPEAKER_04]: or what it's doing to make sure that it is not you know hallucinating as you mentioned when I

[00:20:45] [SPEAKER_04]: learned how large language models work and they're really just predicting the order of words

[00:20:50] [SPEAKER_04]: based on these huge data sets that they're using but they really have no understanding of

[00:20:55] [SPEAKER_04]: the data I think there's a lot of really smart people and a lot of investment going into addressing

[00:20:59] [SPEAKER_04]: these hallucinations we can think about the nature of how they actually work that problem will never

[00:21:04] [SPEAKER_04]: be completely resolved which is why I'm confident that AI is probably going to be more

[00:21:09] [SPEAKER_04]: more augmentative than more augmentation than replacement and if we're talking about you know

[00:21:14] [SPEAKER_04]: responsible AI sure there's going to be jobs that change sure there's going to be some displacement

[00:21:22] [SPEAKER_04]: but again with rescilling and really thinking about what a person does best I think there's going to be

[00:21:29] [SPEAKER_04]: a lot of opportunity for people to kind of grow and upscill and do new things rather than be completely

[00:21:35] [SPEAKER_03]: replaced. Yeah no absolutely agree I mean at least I hope that's the way people are

[00:21:40] [SPEAKER_03]: thinking about it because there's a lot of other work that could be done if only we freed up time

[00:21:45] [SPEAKER_03]: to take on those higher value activities. Let me just ask about in terms of the the client

[00:21:53] [SPEAKER_03]: relationships and obviously you spend a lot of time with clients and prospects what what are some

[00:21:58] [SPEAKER_03]: of the you know high level you know challenges and concerns that you're hearing from them.

[00:22:04] [SPEAKER_04]: Yeah sure first of all there's plenty of interest you know I think every

[00:22:08] [SPEAKER_04]: every sea sweet is has to talk about AI and what it means and so that certainly makes its way to

[00:22:14] [SPEAKER_04]: the people who manage the contingent workforce and the technology associated with the

[00:22:19] [SPEAKER_04]: contingent workforce without a doubt the most most prevalent comment or most prevalent reaction

[00:22:25] [SPEAKER_04]: is a concern about the privacy of data and I think from the enterprise organizations they've made

[00:22:31] [SPEAKER_04]: it clear across to their leadership to their managers and down the line we have to protect our data

[00:22:36] [SPEAKER_04]: we can't just feed our data into open source large language models and expect that they're going

[00:22:43] [SPEAKER_04]: back with the right answers and expect that they're going to protect our data so that's usually

[00:22:49] [SPEAKER_04]: the first question when we're talking with customers about gender to bias okay how do we protect

[00:22:53] [SPEAKER_04]: our data and then to your point how do we trust what's coming back and so what they want is

[00:22:59] [SPEAKER_04]: where are those use cases that can actually improve our lives that can improve our program.

[00:23:04] [SPEAKER_04]: Get me beyond kind of the hype and the you know the robot apocalypse type stuff and you know this

[00:23:09] [SPEAKER_04]: futurism stuff to give me some real examples of how this is going to you know make my life better

[00:23:15] [SPEAKER_04]: and improve the experience for for our consultants and for our hiring vendors. Yeah now that makes

[00:23:21] [SPEAKER_03]: sense I guess if we sort of shift that to you know broad goals in terms of like workforce ecosystems

[00:23:29] [SPEAKER_03]: right you know as people think about the agility of some of these systems and the ability to

[00:23:37] [SPEAKER_03]: happen to different talent pools and use technology to make that an efficient

[00:23:44] [SPEAKER_03]: process is a tie into the data concern in terms of we don't understand it now or we don't have

[00:23:51] [SPEAKER_03]: our arms around it now it's going to that problem is going to be exacerbated when we start

[00:23:57] [SPEAKER_01]: connecting more systems and data. Hi I'm Stephen Rothberg and I'm Jeanette Leads and together

[00:24:03] [SPEAKER_01]: we're the co-hosts of the high volume hiring podcast. Are you involved in hiring dozens or even

[00:24:08] [SPEAKER_00]: hundreds of employees a year? If so, you know that the typical sourcing tools, tactics and strategies

[00:24:14] [SPEAKER_01]: they just don't scale. Yeah our biweekly podcast features news tips, case studies and interviews

[00:24:21] [SPEAKER_01]: with the world's leading experts about the good, the bad and the ugly when it comes to high volume

[00:24:27] [SPEAKER_03]: make sure to subscribe today. Blowing sort of multi-directionally across the ecosystem

[00:24:33] [SPEAKER_04]: or is that a sort of separate challenge? No I think that's a challenge but I think it's also an

[00:24:38] [SPEAKER_04]: opportunity depending on how you look at it so as I mentioned earlier because of our architecture

[00:24:43] [SPEAKER_04]: because of our connectivity and data tools we see integrations as an opportunity yes for for a lot

[00:24:49] [SPEAKER_04]: of people right now it's very complex and it's even it's becoming even more complex but we believe

[00:24:55] [SPEAKER_04]: very strongly in this ecosystem approach because what that'll do is I'll create something that fits

[00:25:00] [SPEAKER_04]: each organizations, each customers, unique needs and preferences in terms of that only the

[00:25:07] [SPEAKER_04]: core solution that system of record like the VMS but also the core infrastructure solutions like

[00:25:13] [SPEAKER_04]: the ERP like the invoice you know like HRIS or HCM and also all of the different point solutions

[00:25:19] [SPEAKER_04]: maybe fit for a specific purpose for given organizations you know unique needs at that time. So we

[00:25:25] [SPEAKER_04]: believe very much in the ecosystem approach and bringing the best of the ecosystem what fits best

[00:25:30] [SPEAKER_04]: for a customer into their tech stack and into their contingent workforce,

[00:25:35] [SPEAKER_04]: tech stacks that it runs well for them. That does bring a burgeoning complexity with regards to

[00:25:41] [SPEAKER_04]: data like you mentioned and so obviously the data has to be secure that's where again the

[00:25:46] [SPEAKER_04]: investments that Salesforce is making is really important for our customers and making sure that

[00:25:51] [SPEAKER_04]: connections are secure making sure that we're only bringing the data that's needed for

[00:25:57] [SPEAKER_04]: example every time we do an integration each field that the customer selects is immediately mapped

[00:26:01] [SPEAKER_04]: through the API for reporting it has to be reportable and so you can see what's happening.

[00:26:06] [SPEAKER_04]: And then with the data management, the machine learning and ultimately the AI tools they're really

[00:26:12] [SPEAKER_04]: good and really important in making sense of that copious amount and complex data from many

[00:26:17] [SPEAKER_04]: different sources. So yes it's complex and particularly complex if you're thinking about

[00:26:23] [SPEAKER_04]: the way we looked at big data and the way we looked at integrations, maybe 10 years ago

[00:26:28] [SPEAKER_04]: where you need a platform you need code based integrations and your workflows are a little bit

[00:26:32] [SPEAKER_04]: more rigid. Now we have connectors and we have tools to bring these all together in a secure way

[00:26:39] [SPEAKER_04]: and in a quick way so that the AIs can start working on making sense of all of that complex data

[00:26:47] [SPEAKER_04]: with of course the direction from the user or the program office, the MSP. Yep absolutely

[00:26:53] [SPEAKER_03]: are you seeing a consolidation of responsibility in this space such that there's a single

[00:27:03] [SPEAKER_03]: person that's overseeing this workforce ecosystem and its evolution or you still seeing

[00:27:10] [SPEAKER_03]: organizations are kind of figuring that out or what do you what your view there? Again I'm seeing

[00:27:15] [SPEAKER_04]: some change management needs so without a doubt we're seeing I'm seeing convergence between

[00:27:21] [SPEAKER_04]: HR procurement, IT and finance on a total workforce approach. So yes those silos have been in place

[00:27:29] [SPEAKER_04]: for a long time and for many years look at the HR people would say oh that's a non-employed

[00:27:33] [SPEAKER_04]: don't talk to me about them even thinking about them is going to create a component or in this

[00:27:36] [SPEAKER_04]: classification issue a little bit of exaggeration there but that's generally the approach that

[00:27:44] [SPEAKER_04]: are pros and cons and she procurement officers are recognizing that you can't operate in silos anymore

[00:27:50] [SPEAKER_04]: things are two uncertain there are two complex that silos in organizations generally are going to

[00:27:57] [SPEAKER_04]: limit what you're able to do and particularly in terms of scale and particularly in terms of

[00:28:00] [SPEAKER_04]: opinit optimization of resources and so we're seeing a definite convergence between HR and procurement

[00:28:08] [SPEAKER_04]: sometimes there are HR people being sent to procurement to you know to help them socialize around

[00:28:15] [SPEAKER_04]: how HR looks at the workforce many times I'm seeing contingent workforce programs leadership

[00:28:21] [SPEAKER_04]: being pulled under the head of TA which is which is actually I think becoming more common as well

[00:28:26] [SPEAKER_04]: as under under the CHRO to start to work on you know desilowing the workforce and being

[00:28:34] [SPEAKER_04]: being much more agile and taking that kind of ecosystem approach of we have all these sources for

[00:28:40] [SPEAKER_04]: workers it would be foolish for us to make sure to not make sure we're tapping into them we're

[00:28:44] [SPEAKER_04]: optimizing the span and the resources and the the output from these different resources and the

[00:28:50] [SPEAKER_04]: last thing I mentioned is the talent marketplace that you mentioned are internal mobility

[00:28:54] [SPEAKER_04]: that's like a bridge in my opinion between the contingent workforce and the FTE workforce because

[00:29:00] [SPEAKER_04]: times those opportunities Michael in parallel to internal opportunities maybe they're prioritized

[00:29:04] [SPEAKER_04]: internally at first if there's a robust marketplace no no takers or no likely candidates and it

[00:29:10] [SPEAKER_04]: can get pushed out to the extent of the workforce and that should be automated through the HR I

[00:29:15] [SPEAKER_03]: asked or through the VMS or some combination of both yeah I'm glad you brought that up I spoke to

[00:29:29] [SPEAKER_03]: available project or new role right if you're immediately sort of going and fishing externally

[00:29:38] [SPEAKER_03]: without looking inward there's there's gonna be I'm not saying there's gonna be a mutant

[00:29:44] [SPEAKER_03]: egg but people want to know that listen I'm here I mean I'm trying to be engaged I'm trying to learn

[00:29:50] [SPEAKER_03]: I'm trying to understand the business I'm trying to be a good you know team player I'm trying

[00:29:56] [SPEAKER_03]: to be loyal to this organization and now you're this exciting opportunity comes up and you're

[00:30:01] [SPEAKER_03]: gonna immediately flip it through you know posted in a lecture I could you know a free that's you

[00:30:09] [SPEAKER_03]: know platform or whatever I mean give them the opportunity and give them right of first refusal

[00:30:15] [SPEAKER_03]: and you know if no one we don't have any takers or we don't see a good skills match face on what

[00:30:21] [SPEAKER_03]: marketplaces showing me then then yeah go and look externally it's pretty clear I've heard this

[00:30:28] [SPEAKER_04]: over and over again nothing makes a valued FTE want to go somewhere else then seeing you know someone

[00:30:35] [SPEAKER_04]: from the outside coming in and taking an opportunity that they might maybe have wanted and possibly

[00:30:39] [SPEAKER_04]: making more money than they are and so that's a really important point in terms of retention

[00:30:45] [SPEAKER_04]: in terms of development in terms of having a really more robust build borrow buy and now

[00:30:50] [SPEAKER_04]: a approach which is another conversation one of the things we're seeing inverted though and we've

[00:30:55] [SPEAKER_04]: developed this tool it's a decision wizard so rather than having hiring managers say I need an FTE

[00:31:01] [SPEAKER_04]: I want to go to the market or I need a consultant or a contractor or SOW and all the nuances that

[00:31:08] [SPEAKER_04]: are associated with those you start with a decision wizard and says what's your need and you can

[00:31:12] [SPEAKER_04]: start putting the need in I need a I need a team to do this build and the decision wizard can take

[00:31:18] [SPEAKER_04]: them through the process of what they're building say specific skills and capabilities and

[00:31:24] [SPEAKER_04]: budget they might have and then based on the work flows can push those needs into the appropriate

[00:31:30] [SPEAKER_04]: labor category or the priorities of the organization so maybe someone says I I have a project build

[00:31:37] [SPEAKER_04]: that's going to take six months you might think I'm going to go to my my favorite consulting firm

[00:31:42] [SPEAKER_04]: to do this or I'm going to do some staff augmentation because I don't have FTE but I have budget

[00:31:46] [SPEAKER_04]: or I'm now I'm being told I need to go to the internal talent marketplace rather than you know

[00:31:53] [SPEAKER_04]: having the hiring manager or the unit leader the resource manager have to figure all this stuff out

[00:31:58] [SPEAKER_04]: and know the nuances you start with the need and it takes them to the appropriate place you still

[00:32:03] [SPEAKER_04]: want to leave a little bit of it's a decision wizard so you want to see leave a little bit of agency

[00:32:07] [SPEAKER_04]: particularly for business unit leaders either based on budget or based on availability or based on

[00:32:12] [SPEAKER_04]: education or whatever criteria the organization wants but we're taking away the need to understand

[00:32:17] [SPEAKER_04]: the nuances of all of these categories and your billing those into the work flows through the decision

[00:32:22] [SPEAKER_04]: wizard so that's something that we're seeing increasingly we built and deployed a decision wizard

[00:32:26] [SPEAKER_04]: for for those purposes and I think that's going to be something that continues to grow in terms

[00:32:32] [SPEAKER_04]: of use and is really helpful in navigating these silos that we've been talking about

[00:32:37] [SPEAKER_03]: for breaking them down actually I think that is exactly the direction that organizations need

[00:32:44] [SPEAKER_03]: to go to be able to have that level of adaptability and optimization simultaneously

[00:32:50] [SPEAKER_03]: and look out across all those silos I'm also of course thinking about putting my response

[00:32:56] [SPEAKER_03]: to the eye had on and that's going to be tough to audit some of that terms of the hiring decisions

[00:33:02] [SPEAKER_03]: I mean it was hard enough with different decision points within the recruiting funnel but now

[00:33:07] [SPEAKER_03]: if the job was ultimately filled through an alternative path and an alternative talent pool

[00:33:16] [SPEAKER_03]: you've got to now aggregate that so I guess just stepping back more generally on a personal level

[00:33:23] [SPEAKER_03]: you know there are AI general AI tools that you think are with that you use in your day-to-day either

[00:33:30] [SPEAKER_03]: your work or even in your personal life or their specific use cases that you like there's a

[00:33:37] [SPEAKER_04]: couple of them so I'm a big grammarly user I think grammarly is kind of like an editor

[00:33:43] [SPEAKER_04]: steroids it makes it really easy to use when chat GPT first came out I enjoyed experimenting with it

[00:33:50] [SPEAKER_04]: and I still I still experiment with it and play with it there are some graphic design tools

[00:33:56] [SPEAKER_04]: that are out there even my son for example likes to graphic design so we got them we got a

[00:34:01] [SPEAKER_04]: max as to some gender-to-day eye graphic tool which are kind of fun to play with so there's a lot

[00:34:04] [SPEAKER_04]: out there what I'm most excited about Bob are the it may have heard of small language models and so

[00:34:11] [SPEAKER_04]: so we have these large GPT's that are kind of large language models and kind of general purpose

[00:34:19] [SPEAKER_04]: tools small language models are much more focused and they're much more they're trained on

[00:34:23] [SPEAKER_04]: various specific data sets for a very specific purpose and so for me those again allow you to

[00:34:31] [SPEAKER_04]: hone in on those particular use cases that are most you know most relevant and useful for

[00:34:37] [SPEAKER_04]: your role your organization your unit and then you can specifically train it with the content

[00:34:43] [SPEAKER_04]: that you want to have that you want to have in it so one of the things we're using a tool at

[00:34:48] [SPEAKER_04]: FlexTrack it's a GPT but again we're using it to retraining it on the things we need for our

[00:34:55] [SPEAKER_04]: customers rather than you know what's generally out there certainly we're accessing what's

[00:35:00] [SPEAKER_04]: generally out there but the training is more focused on our content our use cases our needs

[00:35:05] [SPEAKER_04]: and what our customers are asking about so and and that's really accelerated a lot of what we've

[00:35:10] [SPEAKER_04]: been able to do for me particularly in terms of generating content I still take responsibility

[00:35:16] [SPEAKER_04]: for the content that I push out but it takes me probably you fifth of the time to get a first

[00:35:21] [SPEAKER_04]: draft on using those questions then it would have sitting down and writing something or

[00:35:26] [SPEAKER_04]: are coming up with the graphic so and what it does is it's very relevant I don't have to insert

[00:35:30] [SPEAKER_04]: you know our differentiators after certain our nuances into things we just have trained the model

[00:35:36] [SPEAKER_04]: on things that we we use over and over again and then we can adapt them specifically for a purpose so

[00:35:41] [SPEAKER_04]: I'm excited about small language models and the applications that they can create yeah absolutely

[00:35:47] [SPEAKER_03]: so depth this last question I ask all my guests what do you think the keys are

[00:35:55] [SPEAKER_04]: for elevating your AIQ so you have to get involved you have to engage it's almost like a new tool

[00:36:02] [SPEAKER_04]: think of think of those of us who are round in the 90s when the the internet was you know in its

[00:36:07] [SPEAKER_04]: early stages and proliferating you could be an early adopter you could adopt the tools to help

[00:36:14] [SPEAKER_04]: yourself accelerate or you could be someone who's forced to do it but eventually there's going

[00:36:18] [SPEAKER_04]: to be no choice and the sooner you engage and start to learn and start to understand how the tools work

[00:36:25] [SPEAKER_04]: what the applications are and what what skills as a person you need to bring to the work equation

[00:36:31] [SPEAKER_04]: to get the most out of both yourself and these tools I think that's important so make the decision

[00:36:36] [SPEAKER_04]: to engage there are so many resources out there that are designed to help people understand

[00:36:42] [SPEAKER_04]: AI both predictive and generative you know I could name any of the learning platforms there's

[00:36:48] [SPEAKER_04]: Khan Academy there's LinkedIn Learning sales force offers their their trailways or program which is

[00:36:54] [SPEAKER_04]: you know free training on essentially anything sales force including AI Google offers something similar

[00:37:00] [SPEAKER_04]: in terms of their AI and I think less important than picking the right specific vendor

[00:37:05] [SPEAKER_04]: is a decision to engage and starting to learn how these models work and how you can apply it

[00:37:11] [SPEAKER_04]: but there might be nuances like we were talking about how they apply in different contexts

[00:37:15] [SPEAKER_04]: but more importantly is start to learning journey and it's not necessarily like a learning

[00:37:19] [SPEAKER_04]: program you have to go to school for it or community college or sign up for an expensive program

[00:37:24] [SPEAKER_04]: you can and maybe there are things you can do but there are lots of inexpensive resources that will

[00:37:29] [SPEAKER_04]: take you on a learning path but at least develop your basic literacy and develop your understanding

[00:37:34] [SPEAKER_04]: and experiment with how to use these tools and that's the other part is you know I like to

[00:37:39] [SPEAKER_04]: talk about navigating between fear and hype when it comes to AI and the way to do that is to

[00:37:45] [SPEAKER_04]: one start to educate yourself but then in a safe way start to experiment with these tools

[00:37:50] [SPEAKER_04]: and understand you know build your own ability to apply them in specific situations and grow

[00:37:55] [SPEAKER_03]: and scale that ability as you learn. Fantastic. Beth thank you so much for all your insight this has

[00:38:01] [SPEAKER_03]: thanks Bob it's been a pleasure some great questions next one. Thank you everyone for listening and

[00:38:07] [SPEAKER_03]: we'll see you next time thanks again Jeff bye everyone