HR Strategy 2025: Workforce Planning in an AI-Powered Era

Let's jump into this.
First off, I'm really excited to share
that this is actually a collaborative session
between two separate communities within this space.
That's something I think at Achieve Engagement
we're really excited to do.
So I'm excited to welcome Dave DeAngelo, director
of Community, uh, at Built by People.
They also lead an incredible podcasts, uh,
called Built by People Podcast.
So we'll share the link for that in the chat in a second.
But Dave, really excited to develop this new collaboration
between our networks and bring some of these insights
and thought leadership to, to both of our communities.
So it's great to see you
and appreciate you being here with us.
Zach, thanks for the invite. And Zach.
Zach and I actually met through another community,
which was transform.
And, uh, uh, I really admired the work that ZEC is driving,
and it was just so clear to me from the beginning
that we had to find a way, uh, to come together and,
and partner on something like this.
And I can't think of a hotter topic
to do this over than workforce planning and,
and AI powered era.
Uh, so thanks for this
and, uh, I'm excited to get us started.
And, uh, just, just for those in the audience, uh,
as Zach mentioned, um, I am the host of the Built
by People podcast, where we elevate the stories
and insights of the world's top HR leaders.
Uh, we put out about 50 episodes a month across Spotify,
apple, and Amazon.
Uh, Zach will share a link to that.
I hope you're, uh, listening into that
and continue to participate in great, uh, networking
and learning opportunities like the one
that Zeck helped put together today.
Exactly. It is, definitely check it
out after this session.
Right? And there's tons of amazing episodes as Dave shared.
I mean, incredible leaders taking this,
the podcast stage with them on that front.
But that being said, I'm excited
to dig into this discussion with you, Dave.
I mean, this is a very much a, a relevant conversation
for all of us as we try to move our organizations,
our HR departments,
and even our people strategies into the future.
So, that being said,
should we start this off, Dave, and jump into it?
Let's do it. Yeah.
And, uh, we're gonna be doing like two fireside chats here.
It's a, it's kind of a unique approach
that we're taking today,
and I'm excited to welcome to the stage Josh Newman, uh,
who I met through Cal at Transform
because of a, a LinkedIn post that went out
and Cal was just like, you gotta talk to this guy, Josh.
'cause if there's one guy who knows workforce planning
and how to kind of integrate that with ai, it's Josh.
Josh was also just featured on, um, the Josh person podcast.
So, uh, I think that's gonna be shared with everyone after
or toward the end of this call.
Uh, so Josh, thanks so much for joining us today.
I wanna dive into the,
the seven questions I'm gonna be asking you in the 20
minutes we have, uh, together.
And the first question that I wanted to ask you is, what are
CEOs expecting from leaders
regarding AI integration in workforce planning?
First of all, thanks for, uh, thanks for having me, Zach
and Dave, and Glad Cal connected us, uh,
after that LinkedIn post.
Just a quick two seconds on, on WPP and,
and my, my role within that.
And, um, it'll help sort
of contextualize the commentary from here on.
Uh, so I work for WPP.
It is the leading advertising
and marketing, uh, holding company.
Uh, we consist of media, creative pr, commerce,
production agencies.
We're about 110,000 people across a hundred countries.
Uh, my role within that is leading people strategy
and employee experience that consists of employee listening,
people marketing and communications, all things future
of work, which really lately the past two years has been
understanding the impacts of AI on the workforce.
Um, so just to step back from the question.
So now, now with the context on WPP
and my role, this all started when, uh, our board
of directors was asking our HR team, uh, my,
my boss, our C-C-H-R-O
and myself, what is the future shape
of our workforce, right?
We have 110,000 individual humans that comprise, you know,
a number of, of roles, uh, quite a varied amount of roles.
And, you know, our instinct to answer that question was
to pull the microscope out
because we can't make any estimates, uh,
using broad brush strokes.
Um, the only way we thought we'd be able to
predict, uh,
or guide us to a new future with new sort of sets
of tools and capabilities within the organization would be
to unpack the work that every individual is doing.
Now, it sounds like a tall task,
but to understand the impacts of ai,
we were fortunate enough to have at our fingertips AI tools
to do that and accelerate that.
Um, creative industries, uh, particularly
advertising industries, do not do the best job
at capturing process.
There is some magic that happens to produce truly creative
game changing business growing work.
Um, so there's a bit of pride that comes with the magic
behind the process, which means there's very little
documentation for a lot of processes.
Uh, some of our media type businesses do have documentation,
but you know, others do not.
So, again, our first thought was let's unpack who we have,
what they're doing, and really understand the
work architecture.
This wasn't exactly the question that our CEO
and the board were asking, but we knew in order
to answer their question around
what is the future shape size of the workforce?
What is the hierarchy gonna look like?
Are we gonna have more or less junior, mid
and senior people and we do today?
Uh, that was the only way to get to it, right?
So I'll, I'll, I'll pause there. I can keep going.
Um, but, uh, the instinct is to reframe
the initial ask, which, you know, with all of the media, uh,
around ai, of course, that question is going to come
to the people team, but it's up to us.
And, and what we did was reframe the question
to something much more granular to begin unpacking that.
Josh, building on that thread,
how do you see AI changing the definition of strategic hr,
especially in the last year?
I think it depends the starting point of
where HR sits in your organization, uh,
there's obviously spectrum of organizations that
leverage their HR functions extremely strategically
to make business decisions.
And then there's HR functions
that are more compliance based.
And I wouldn't say there's any right or wrong.
Uh, I certainly get more jazzed
by the former, uh, than the latter.
And I think what, what AI
and really generative AI is what, what we're talking about.
'cause that is what instilled this new wave of,
of questions, uh, and ideas.
Uh, what it's given us permission to do is be
at the table in a way that, uh,
we might not have been asked to in the past.
Right? The only way to understand how, uh,
the workforce is gonna change is
by understanding the impacts of new technology on workflows,
how work, how those workflows can be delivered against
by teams that are probably structured differently,
which means individual humans need new skills
or different skills, or different skill combinations.
Um, so we're actually getting a lot of really call it,
uh, traditional HR questions pointed to our team,
um, in a way that we haven't in the past.
Right? There, leaders are more interested in talking about
talent, talent management, uh, skills, capabilities, uh, in,
in a way that, that I find hasn't, hasn't been the case, um,
as fervently in the past.
Josh, where is AI creating tangible value
in workforce planning?
And is there anything that you consider
to be a bit overhyped in this space at the moment?
There's a lot that's overhyped.
Um, you know, I think
what's what's overhyped is that, uh,
AI will instantly unlock all sorts
of capacity across all, all sorts of roles.
You know, in, in the research that we've done, uh,
which included shrinking the complexity
of our job architecture from about 55,000 individual unique
job titles to about 600 archetypes, um, is
that there, there is such a spread depending on one,
the types of tasks and work that an individual contributes,
and also the technology that's available.
Just because people have all
of the latest large language models at their fingertips,
one does not mean they'll not, they know how to use it.
And even if they do, it might be adding time
to their process rather than taking time away
because they're using it on higher value work.
I mean, just from my own experience, I know that I go into
avenues of exploration and research that I
otherwise wouldn't have in my natural language conversations
with, you know, LLMs that I use on a daily basis.
So that, you know, that that I think is the overhype,
that it will instantly unlock a type of capacity
that will then, you know, reduce head count
and we can just drive that to the bottom line.
I think that's bs. Uh,
and, uh, you know, I think you gotta get into the details
to figure out where that value is.
So your first question on value, I think value needs
to be looked at holistically.
Um, we're, we're lucky,
and I feel fortunate, again, to be pulled into conversations
with our commercial team, right?
In, in media and advertising,
we're basically professional services.
So the bulk of our revenue comes from time
and materials where we,
we sell our people's time and their expertise.
And so if clients are expecting, uh, rightfully so, they're,
it's a, it's a good question to ask if,
if a widget takes 20% less time to create, uh,
we sh we want to pay you 20% less.
So we're, we're thinking through new commercial models
that are win-wins for both us and our clients.
We are pulled into those conversations
from a talent perspective.
Capacity unlocked time savings, if you will.
Productivity increase is a value
that you can derive from certain types
of technologies if applied, sort of
with pinpoint precision on tasks.
But you're not just gonna give someone a tool
and say, oh, over overhaul your work.
Um, the best, uh, the best examples of this
I've seen are from talking to people who represent sort of
that 2.0 of the job that they're already in.
They have naturally organically evolved how they drive, uh,
delivery and outcomes of the
work they've been doing already.
So it's about asking them, Hey,
how are you using the tools differently?
And then setting that as the minimum standard for all
of their peers in those same type, same archetypes.
Josh, what are some of the new AI driven demands
that you're seeing being placed on people teams?
Uh, especially in mid-market pressures?
I, I would imagine
that most HR teams like ourselves are being asked to,
uh, find efficiencies.
Um, and again, there's some validity there,
but efficiency for the sake
of efficiency will never drive growth efficiency.
In our case, the way we think about it, efficiency
for the sake of effectiveness is the only way
we can balance out the bottom line
and growing the top line, right?
And, and I've used this phrase a few times, capacity unlock.
If we figure out, if we predict
capacity unlock based on certain roles by augmenting
and speeding up certain tasks within those roles,
what we can do is go to business leaders
and say, Hey, we expect this role
to have a 15% capacity unlock or productivity potential.
Um, now it's your decision.
Do you wanna have less people in those roles
delivering the same amount?
If you plan on growing, you should have more
people delivering more.
Do you wanna spend some of that time on upskilling?
Do you wanna, do you want to apply some of that time
to wellbeing initiatives, right?
So there's optionality
that business leaders have when you're thinking about
that efficiency and the same exact thing applies
to, to HR teams.
Josh, are there any ethical concerns
that you see arising when integrating ai, uh,
within workforce planning processes?
Absolutely. Uh, you know, I, I think it's something that,
it's something that we keep, uh, at the forefront.
We work very closely with our legal and compliance team
and our data privacy team.
I know not every organization, um, has functions like that.
I think we can, uh, leverage, you know, you can,
you can leverage tools out there, especially LLM tools
to understand some of the latest, uh, regulations
around data privacy.
Uh, we are confident that
we will never be using AI to make hiring decisions,
promotion decisions and things like that.
Um, you know, take something like performance management.
Uh, the traditional way to do that, you know, back,
it's been around for quite some time.
You would use the nine box.
Um, there are more advanced methods,
but you can use generative AI
and certain AI technologies, specifically purpose
for performance to understand where your areas
of opportunity are opportunity from there.
But you're gonna want to use, you know, that unbiased, uh,
human eye on, on making those decisions.
Josh, how should AI powered people analytics influence
daily HR decision making?
In your mind?
How should AI power people analytics
and data-driven decision making?
I mean, and I think it's kind of one and the same.
If we talk to our people analytics teams, um,
they're at the forefront of, of AI capabilities
and have been for the last 10 years.
Uh, if we're talking specifically about generative ai,
you know, I'll give you one example on
the people listening front.
Um, as we roll out things like engagement surveys, we're
training managers
and leaders who get access to the results on how
to leverage generative AI to interpret the results,
to support with action planning.
You know, and you can very simply build,
whether you have internal tools within your organization
or, you know, free or premium tools like Chachi, BT Plus
or Claude, you can fairly easily build an agent
to put parameters around those recommendations.
Um, so you can maintain the kind of responses and,
and sort of limit the ethical concerns.
Josh, as, as we wrap up this, um, fireside chat
around AI driven workforce trends, I have one more question
for you before I bring Z up on stage.
And that is, uh,
how else do you see AI reshaping some macro level
workforce planning strategies?
I think it, you know, I I,
all of the conversation around AI makes me, uh, compare it
to the most recently probably equally hyped,
uh, technology, which was laughably, uh, the metaverse.
Uh, I am not comparing the two in terms of provenance
or sticking power.
I think, uh, generative AI
and AI is clearly here to stay, uh,
and transform everything.
What I think it does is it further enables HR teams to do
what they have always strip strived, strived
to do, um, which is ensuring
that the right skills are in the right place
to do the right job at the right time.
Uh, and the, the more we understand the architecture
of work, the more we can, uh, support
strategic workforce planning and ensure that those initial,
uh, outputs of HR are held true.
Awesome. Well that wraps up our first fireside chat.
I'm now gonna bring on, uh, Zach who's going
to lead the second fireside chat with Taylor.
And then we'll all come together as a group toward the end,
um, to, to discuss, uh, some other things.
So I'm gonna hand things over and, uh, both me
and Josh will go on mute and take up our video. Now.
Thanks for the rapid fire talk soon.
Great kickoff fellas, everyone attending.
Let's give 'em some appreciation
for warming up this virtual stage
for Taylor And I, uh, that was awesome.
So I'm excited to kind of continue that build off of it,
maybe even repeat
or provide some different
perspectives on some of those things.
So Taylor, welcome VP of talent strategy
and success at touring.
You are all warmed up from a couple sessions today,
so what's going on?
Taylor, good to have you here
and, uh, appreciate you taking some time with us.
Thank you for having me.
And I have my cup of coffee, so that's keeping me going,
but I'm always living the dream,
so looking forward to the chat here.
Love it. Nothing, uh, AI can replace our coffee
for these things at this point. So
Not yet. Not yet.
I tried the AI, or at least the robotic coffee
and s o's airport, not the same.
Okay, well if there's one insight for all of you,
stay away from, from that coffee,
but, uh, yeah, Taylor excited to keep this going.
Obviously I think as, as Josh expanded on,
like it gets really excited when we're able
to really understand our workforce job architecture, the,
the mapping out of our talent and the needs from our talent
and, and do a lot of things that we've always wanted to do.
So yeah, one would love to hear more,
like maybe introduce yourself a little bit of who you are,
the type of work you're doing.
But I would love just to start off with like as people,
leaders or CHROs
and the people attending, like what are some
shifts in our mindset and the way we should view
and embrace AI for workforce planning
and going forward with our strategies?
A big question to begin with
and we'll kick off with an intro here.
So my name's Taylor Bradley, head of Talent Strategy
and Success at Turing.
For those unfamiliar with Turing,
we are an AI infrastructure company.
We have two main services.
So one is a GI advancement where our teams work
with researchers
and collaborate with them at leading AI labs,
if you would think of chat, GPT, Gemini, Claude
to improve their models.
And then on the turn intelligence side, that's
where we leverage that expertise on training those frontier
models and helping those organizations release all those
updates to all of you is we leverage that expertise
to help you identify
and implement practical AI solutions in your own workflows.
And we've had quite a range.
There's a lot that I do internally on the HR team.
One case that I found really interesting that I just learned
that we did is worked with a
pretty prominent clothing manufacturer to replace, uh, their
human models
and onsite photo shoots, all digital and AI now.
And that was mind blowing to me to, to see that demo.
So there's a lot we can unpack there, Zach.
Well, thinking about that example too,
'cause it's like, it's hard for me sometimes even
to grasp at should we be looking at this from a workforce
planning standpoint as a AI model that we invest into
to execute a certain job?
Or is this a human that we need to hire and train?
Sometimes it's a blend of both.
What was maybe even that client that you worked with, like
what are some of the mindset shifts we need to have so
that we can start to even include those type
of questions in our workforce train, like planning, like
how, what are some of the shifts you see there?
Broadly speaking, with all of our clients
and internally, the first thing
that we encourage organizations
to do is understand the roles that they have.
And this comes back to job mapping.
It's a skillset that has been getting a little dusty in hr,
but really understand what task each job is doing.
Because our roles are not one single thing that we do.
If we did that
or if that was our role, that would be very easy
to replace with ai.
But our roles are actually a combination of hundreds
of different tasks that have
to come together in our brain matter to lead
to a certain outcome.
And what we do is on my team, look at those tasks
for a people operations specialist,
for a talent development partner,
and not set out immediately to replace those roles with ai.
I think that's a misguided approach,
but instead go what part, what task
of those individual roles are well suited
to be AI managed either through automation or augmentation,
and then go step by step at chipping away, freeing up
that human to work on additional task.
And when it comes to my internal org, I lead a team of
around 40 human HR professionals,
but also on my org chart, I have two ai,
FTE that report into me.
One is a talent development partner
and the other is a people operations specialist.
I mean, how cool is that? Is that your,
like two full-time AI agents reporting to you?
That just sounds wild to me.
Like, uh, how did you start to uncover those tasks?
Like how do you determine like
what are questions are you asking yourself
or can you encourage even our audience
to start asking themselves about their own roles
or their team's roles to figure out which one
of those they can start to augment with AI
to free up that capacity?
Yeah, there's a nice infographic that we made very simple,
and you, you all can see it on my LinkedIn right now, is
it highlights a role
and then there's little dots that represent each task.
First you have to map all those out.
What are the gray on this infographic?
If you look at it, what are the gray
human-centric tasks, right?
And then having an understanding as what are the strengths
of large language models of ai
and how can you take those individual tasks
and convert them from the gray human-centric dots
to blue AI managed ones on the infographic
and having an appreciation.
This is where we see with some of our clients sometimes of
what AI is currently capable of, uh, versus
what people think it's capable of can be a chasm
that's not even manageable to cross.
So the tangible use cases that we help our clients
to identify are ones where they can build momentum behind
and more importantly, either go to their shareholders
or their executives or most people on this call their CFO
and say, here was a use case that we beta tested,
we designed a proof of concept, it led to this
and improvement and actually get tangible results
associated with it.
We want additional investment to accelerate that
and to scale that use case.
That's where we really do a lot of work at, uh,
with our clients to be able to do that.
And I see some people in the chat,
I'll just drop a link in here too
so they can follow along on that infographic.
Yeah, I was gonna ask for that myself.
Like even thinking about my own role, like where,
where are those dots for me?
And um, I think
that's another important point that you hit on.
I think there's also that, that chasm where
we have this idea of
what AI in large language models can do,
and then there's just additional potential what it can do
that if we're not educated
or aware of that it's hard for us to even
support this next evolution of our workforce
to build out some of those AI agents that report to you.
Right?
That's right. And a common,
it was just at CN BBC's CEO summit in Arizona.
And that always affords me the unique opportunity to connect
with CEOs to some of the world's largest companies
to get an understanding of how they're a applying ai.
And last year when I went to the event,
I saw this big disconnect of
what I ended up breaking down is short term versus long term
getting to orbit versus getting to Mars.
I used a SpaceX analogy, which I realized may not, uh,
resonate with everyone, uh, given the recent climate,
but going with me on this journey is we had so many CEOs
that were fixated on, we need to get to Mars
and we're gonna launch this huge AI program.
It's gonna take multiple years to implement.
That's, you do need
to be thinking about those long-term initiatives,
but if you don't even get to orbit with your workforce
of having them leverage the LLMs that are existing today
or embedding in the workflows that's going
to be dead on arrival, you're just never gonna get
to Mars in this analogy.
And so that's one area that we're actually pretty diligent
with our clients of.
Do they have a realistic expectation of where they are today
and a realistic expectation of where they want to go
to make sure that it's a good fit for us.
Because if they have an unrealistic expectation that
for Star Wars fans,
they're gonna immediately have C3 PO walking
around their office office, that may not be a realistic
outcome in the near future one day though.
Yeah, yeah. It's funny
'cause uh, obviously May 4th was, was just around the corner
and I just watched, uh, star Wars again.
I was like, oh my gosh, is this,
am I gonna have that someday?
Um, so yeah, fingers crossed.
Hopefully you, you help us build that out.
But, uh, I wanna back up to Dorothy's question too,
and there's a few others I want to cover,
but since we're on the topic for your department as well,
uh, could you tell us more about these AI agents that report
to you and give some examples
of those two employees and what they do?
Absolutely. So on the talent development side,
we made a use case in 2024 at this point,
so a little over a year ago where we wanted
to rapidly scale our asynchronous learning modules for our,
uh, employees or for our team members around the globe.
And what you learn is when you're on the frontier
of helping train large language models with open AI
or Gemini, is that the technology rapidly changes,
therefore the processes continuously change.
And a general industry standard is
to create a technical asynchronous interactive learning
module can take anywhere from two to five weeks.
We set out to disrupt that
and started to train a proof of concept.
So before we invested a ton of money, we just went to chat.
GBT used that as our r and d lab
and effectively create it in the aggregate.
What an instructional designer would do
with raw input from a engineering leader,
engineering leader would record a th 30 minute video.
They would put it through the custom GPT
and the output would be anything that we would need
to immediately create an asynchronous learning module
that could be from quizzes to how information is positioned.
And that was all trained.
We helped do the fine tuning on adult learning practices,
best practices in instructional design, English
as a second language best practices,
because many of our team members are not based, uh, we're
or do have English as a second language.
And that output was astonishing.
So we took that two to five weeks
and dropped it down to about three
to six hours from the moment
that an engineering leader would have an idea of
what they wanted to train on.
Three to six hour later we'd have a fully built
out module ready to go.
And that blew me away at that time.
Now we literally have that down to within two hours
that we get the raw input and we generate that output.
And that has not replaced the talent
development members of my team.
I have three and we support an organization of around,
if you know, fluctuates based on the types of work we do,
but anywhere from six to 7,000 people that I'm supporting,
which is three talent development partners.
And that has been the exponential
gain that we've had on that.
I'll take a breath here though,
before we get into the people ops
to see if you have any feedback or questions on that.
I mean, that's incredibly exciting
and I'm sure for everyone listening
who has done curriculum design content development for l
and d and facilitated learning, like how much time
and hours goes into that.
Like I'm sure we would all love to have something
that could take a raw video
and convert that into the deliverables you need
to upskill your workforce on something
that's happening right then, right?
And you think about how fast demands shift in the
environment where we need to upskill people quickly
to be able to deliver to the new demands
and needs of whatever that might be.
Like that's the type
of turnaround you're, you're seeing here.
And and one of the questions that came in was like,
what does upskilling look like as it looks to ai?
So as you think about those three team members
that maybe traditionally were doing that,
that curriculum development and designing the worksheets
and all those things, what was the the,
the change management
and upskilling like for those three individuals now
that you have an AI agent operating in that way?
We first faced the same challenges
that many leaders in this call are facing
as a fear around it.
We're an AI infrastructure company,
yet some folks on the team were
so nervous on am I training my replacement?
So getting them to understand the upside
and showing them more importantly
how they could achieve a significant productivity gains
through A LLM usage.
Custom GPT really started to get them on board.
The biggest shift that I've noticed is the creativity
that's unlocking for these team members.
These team members are generally at the
merging part of their career.
They do not have advanced degrees.
Two of the three as first time they've ever
worked in talent development.
They were internal transfers to the team,
but they're putting out PhD level work because of the LLMs
and the use cases that we've designed.
So now, instead of going through and doing tedious things
and attempting to try to understand the technology,
or I'm sorry, the SME content behind it,
that's all being handled by AI
with a final human in the loop checkoff by our SME, right?
But they get to work on creative things.
And an example of that is one of them said, Hey,
why don't we actually put the senior engineering leader out
of a job of facilitating the session?
And I was like, what do you mean by that?
That's when we started to have one of them
go out on their own and they developed an ai, uh,
lifelike indistinguishably lifelike AI synthetic voice
that would replace the facilitator entirely
and rewrite the script to be high quality voiceover script.
So that way when we need to go change things,
you just edit the script and click publish.
You don't have to go do any reshoots.
If the SME wanted something said slightly different,
no reshoots, no editing,
just change the script, click publish.
And that was because the talent development partner had time
to go think of what else can we do?
What other gray dot can I turn to a blue AI manage one?
And that's also unlocked a tremendous amount
of productivity for us.
And the SMEs love it
because if there is anything that they want to change,
it's very light lift.
It's just a matter of minutes
to go back and change the content.
Ah, it's so exciting and it sounds like you're helping a
few people's even leadership proposals right now,
like Alison and Sharon in the chat with.
That's awesome. And, uh, yeah, I was, I was curious
of maybe lessons learned as you shared, like
as the models change, that also changes how you need
to update and refresh
and train the model differently
for the tasks that you're doing.
And Sue, you asked a great question now.
I was even thinking about like the risks involved to this,
where we hear about these hallucinations that are happening,
uh, or these hallucination problems.
Um, so obviously it sounds like you have
that quality control checkpoint as a part of the process.
Um, but yeah, what, yeah, what's your thoughts on some of
that and what are some of maybe like yeah,
the lessons learned where I know some
of the people in the chat that are like, okay, I want
to get on this journey.
How can I maybe learn from Taylor's mistakes
and issues that he had so I don't repeat those?
Well, plenty need to go around.
And the HR business partners,
when I first started getting them to use LLMs,
a very important guideline that I put in place is it needs
to compliment, not replace your voice and expertise.
And what I encourage them,
and I shared with them some highly publicized use cases
of lawyers, uh, putting forward things to the courts
of made up cases that were hallucinated,
I don't want you to be in that position.
Right? And it's important
that you always remain the human in the loop.
And even if you look at the graphic that I shared earlier,
even in the getting to Mars, there's still a,
there's still a human at the center of it,
even if they're managing teams of LLMs or AI agents.
So we have to have that quality control check.
And the second piece on the talent development side
is recognizing when you
are letting AI write too much.
And some of the feedback
that we got immediately when we implemented, uh, some, uh,
AI assistance on the people operations side is the people
operations team didn't like it.
Not because it was helping them speed up their task, but
because what was being written was clearly
written by a robot.
And so the exciting part for me was, Hey,
that's a linguistics problem.
That's just a matter of training it on linguistics.
So I asked the team share with me some emails
where you feel like you were just the hemming way
of writing a good email back to someone.
What does that look like? Then we can take those examples
from the humans and apply that into the linguistics training
of the LLM itself to say, mimic this.
And once we got the team involved
and they understood that that's just something
that they can find, tune out of the system, they were able
to then embrace it.
It was, oh no, this is exactly writing it
how I would've written it,
or even better than I would've written it.
And that's when you really start
to get the team involved. Yeah.
Gosh, it's so cool and uh, really exciting to hear kind of
how ahead you are with this, and then obviously serving
and supporting other companies.
So that being said, let's keep the conversation going,
but I wanna bring the full group to kind
of continue unpacking this a little bit more.
Uh, let's maybe as a, as an audience,
can we give Taylor some appreciation though for,
for unpacking kind of that journey to, to Mars so far?
Uh, that was so cool. So Josh, Dave, welcome back.
Let me bring you up, uh, how cool
of a discussion this has been so far,
and as we kind of open it up to the group,
I'll just shoot it over to you, Josh, just start, like,
anything pop into your head as we talked about, like
this journey to Mars, but how do we get in orbit first?
Yeah, I, I love that analogy, Taylor.
Um, you know, I I I've heard in some round tables
and some conversations with, with leaders, uh, this tension
or this back and forth between do we invest
and then what, what, what's the value realization look like?
Uh, or do we just kind of wait and see?
You know, I I, I don't think wait
and see is an approach that will win any sort of
game, business or otherwise.
Um, so I think you gotta start, even if you start small,
start very practical figure out, again, I'm going back
to the same few points, but someone told me, say one thing
and just repeat it, uh, to get the message through.
So understand the workflows,
understand the work architecture, pick out the stuff, tasks
that people just don't like doing
and that doesn't add value.
Um, and start figuring out how to, how to accelerate that.
And Taylor had some wonderful examples there.
Zach, this is my, this is my calling now, isn't it?
Uh, jump in. Yeah.
Otherwise, I, I kind of have a follow up
for, for Taylor. Yeah, yeah,
Yeah. Um,
I've Been, I've been thinking about Taylor's, uh,
answers, you know, 'cause Taylor, your company
at its core is very AI driven.
Uh, a lot of companies don't have that luxury of being, uh,
like AI native or, or AI first.
So I, I wanted to ask a question to both Taylor
and Josh on how can we start getting HR leaders
to prioritize being AI first
or AI driven in the context of workforce, uh, planning,
especially with all the other competing demands,
and what are some, some, some tactical
and practical insights you might share to help them take
that first step if they're just getting started?
Yeah, happy to cover that.
The, the first part about Touring's journey
or my part with touring is
we 10 x headcount in the past year
and a half, maybe a little less than a year actually.
And that led to a lot of demands
that just totally obliterated any kind of,
of HR processor system that we had in the organization.
And so what ended up being a joke at first ended up becoming
our mantra I shared with our CEO.
He asked how you would get to scale the team this way.
And I said, I'm very AI our way out of it.
And that ended up being the mantra of the only way
that we were gonna be able to scale is if every team member
in HR started to use AI to get that proficiency.
But not every organization's going
through hypergrowth like that.
So it may feel like, oh, this doesn't connect with us.
And I wanna go back to something that Josh said,
'cause this is the most common complaint I hear from CEOs
when we discuss it with them.
I went out and I got all of our organization chat, GPT
or Gemini licenses, and no one's using it, right?
That is the biggest missed opportunity
of getting folks comfortable with ai.
It's like the least threatening entry into utilizing ai.
And what we encourage team members
to do is encourage your managers to speak to their teams
and start using AI daily in conversations,
just getting people familiar with help.
Help me rewrite this email so it sounds less harsh
or is more constructive.
Have a look at this performance
feedback I'm going to give to someone.
What coaching tips can you give to me as a manager?
May not be always right,
but it's encouraging team members
to really go explore in their own individual workflows
how they can start to leverage AI to be more productive.
And that applies to every one of our organizations.
So those are the two things that come to mind on that front.
I forgot the initial question, uh, Dave,
but Taylor, just to build on that, you know, I, I talk
to people often who are, uh, not in a situation
that you're in with 10 x growth in a year and a half,
but people who are often, you know,
wondering why we don't have budget for additional headcount.
They wanna hire, hire someone, they need some, you know,
junior support and so forth.
And, you know, my my go-to is similar,
the framing is slightly different in the sense that, um,
your average LLM treat,
treat it like the smartest intern you've ever had, right?
You're always going to check an intern's work, you're going
to make sure that they're on message, on task, on
whatever it may be, but they're gonna generate some things
that you're like, holy s**t.
Like, I didn't realize I,
I didn't, I hadn't thought about that.
And it'll take you down a different path, right?
So like writing, writing the email is I think good.
The, the three things, uh, within that intern sort
of umbrella I think about are search, research
and content, right?
Use it for search, uh, play around with it
for search versus Google.
Play around with it for research
because you can ask it to cite references
and sources, uh, most of the models anyway.
And then content, you know, no matter what kind
of content you're building,
and we're all content creators one way or or another.
We're all communicators and storytellers one way
or another use it for, uh,
to ensure you're not starting on a blank page.
And the, the misconception that prompting is hard.
I hate it that somewhere along the time timeline,
someone added engineer prompt, right?
Prompting is super easy.
And, and I gave the example Dave, I think when you
and I were talking, is I'm teaching my
3-year-old daughter how to prompt.
And how we do that is every night she's learned
that if she gives me an animal, an activity
and a location, Chachi PT will create a
custom coloring book page for her that we print out.
And she gets, we've also branched into,
we're creating a little Pixar movie with vo Google's new VO
where she's a character in it.
And I'm a character in it, and we're
making a soundtrack for it.
But the one thing that I I I claim as an hallucination
is I have hair, but every time it generates me is a prompt.
It, it, it's always bald.
So that's, that's the one thing
that we still gotta fine tune that one.
You are next level, Taylor.
I will say, you know, my toughest mornings with my five
and 3-year-old, I turn on the chat function
on chat GBT,
and I'm like, okay, Stella, Julian,
what are you into right now?
Dinosaurs and princesses. Sounds good.
Okay, chat, GBT Stella wants
to hear a story about di princesses and Julian dinosaurs.
Tell it a five minute story, make sure
to ask it five questions in those five minutes
and build on what they share.
And that's, you know, and then I can go get dressed.
I love that. So for some of you listening with kids
and you're looking to open up some capacity in your
mornings, like you already got some use cases here.
And, and I, I feel like you answered this,
and Melissa asked this in the q
and a as well, like, how can humans gain the practice
and experience and some of these more entry level skills
that, you know, to outsource to ai?
And it sounds like in many ways
it's just like getting in the game and getting in the arena
and starting to think about scenarios and,
and it's, it's, it's, as you do that almost your imagination
and your perspective on what's possible expands
through just the usage of it.
And Taylor, what, even as a company,
when you're going into organizations
to enable 'em in this way,
is there certain things you're doing
to like guide their people to become more enabled and,
and part of this transformation?
Yeah. The methodology that we share with
non-technical folks, again, I'm not an engineer, is
the idea methodology is we want every person in your
organization to be able to identify design and experiment
and action and AI use case in their workflows.
And the beauty of it is, I'm partial
to chat GPT though Gemini is certainly growing on me
with a 2.5 pro release is those are your r and d labs.
You don't need to write a single line of code
to create a custom GPT that will enable you
to have the fine tuning.
You want to really test something.
And if it's an enterprise license, which is important
for governance because it's not gonna train on your data, is
that you can use that as a sandbox to safely test ideas.
And that enables you to then go again to your executives
or your CFOs and say, Hey, I did this, Josh, I did this.
Here was my results.
If we invest, it's gonna add more fuel to the fire. Right?
That's a much more powerful positioning than, Hey Josh,
I I wanna do this thing.
We haven't tested it. It looks pretty cool either.
They've heard that before on every SaaS tool
that we've ever requested, right?
So this is a, a great enablement
and it really democratizes your ability
to create some really cool use cases.
And I'll end with my talent development manager.
She messaged me last Wednesday, it was,
and she coded an app that helped
automate a specific notification process when they were
onboarding candidates.
She coded herself, she's never coded in her life,
and she used Codex to help create this tool.
And then I of course said, this is amazing, it's working,
which is awesome, but let's go have engineering
and it review it to make sure that it's right.
And so they looked at her GitHub repo on, uh, our platform
and they were like, this is actually great.
This is perfect. There's
no other things that we would change.
So she did that just by natural language prompting.
So, so getting, um,
executive teams really bought into this.
It's, it takes time, it takes trust, it takes evidence,
you know, building a use case for how this is going
to drive performance results, efficiencies in the business.
I'd love to get both of your perspectives on
what you've done as a p in a people function
to build executive trust, uh, when deploying AI
in the context of workforce planning
or even something as generic as the HR function.
Yeah,
I can jump in first.
Um, I mean, at the end of the day, with the tools
that are at all of our fingertips, we no longer have to go.
And Taylor, you were just broadly making this point.
We no longer have to go
and say, here's what I think's gonna happen.
We can go to our leaders, whether you're asking for budget
or permission or whatever you typically go to your leaders
for and say, here are the results of the test
that I did in my garage, you know, my proverbial garage.
Um, so it, again, it, it accelerates if it gets you
that step ahead.
Uh, so you're going with proof points,
Taylor, you'd add to that?
Yeah, I always joke, yeah, proof not promise.
And especially when it comes to ai, like if we had
that approach, probably going back to Josh
with like metaverse
and some of these other things may have not have expanded in
the way that it did, but certainly there's, there's proof
to be had and I think my executive team,
it's a little unfair comparison
because we're leading the AI charge, right?
So, uh, in fact there's probably some use cases
where I've had to say, let's slow down, let's figure it out.
Let's, let's work through some of these things first
before implementing.
So at times it's raining in rather than getting
buy-in on my execs.
Yeah, I got, I would just jump into, oh, go ahead Josh.
No, go ahead Josh. Yeah, I was gonna jump to one
of the questions in, in the chat.
Uh, it was posted by anonymous.
Um, talk to me like I'm five
sounds like the start of a prompt.
Talk to me like I'm five, what are the basic steps I can
take to begin building a capacity model
or hiring plan aside from just looking at total budget?
Um, I mean, one, one of the things
that we recently were asked to do, you're, you're,
you're talking about the basics of workforce planning.
We were asked to think about the new, you know, the three,
five years from now, what is the shape, structure, uh,
and skillset of the workforce look like
based on the new technology coming in.
So we, we were being basically asked
to do workforce planning 1 0 1,
but with, you know, the tech first component.
Now, I wouldn't propose to do it that way,
but the four steps or the four sort of nesting dolls
that I think about within that,
and maybe this beyond five-year-old, uh,
but you can transcribe
and translate using one of the tools we've mentioned here.
Uh, first is understand the outcomes
and deliverables you are trying
to get out there in the world, whether that's, you know,
physical production or some sort of knowledge management,
whatever, whatever the outcomes
or deliverables you deliver to clients or customers.
From there, you need to figure out the types of teams
that need to deliver it.
What makes up teams.
Teams are made of up of individuals in roles.
Those roles are occupied by humans with skills, right?
So you start with the outcomes you're trying to drive,
and ultimately you get skills.
And if you have enough of that,
and that's like the work architecture I mentioned earlier,
you have enough to start building out, um, the beginnings
of a workforce model or workforce planning model
And a, uh, five-year-old language.
Taylor, you would add on that one
or I feel like yeah, that hit on the nail though.
The, the no, that, that hit the nail on the head.
The thing that I would add is reiterating
what Josh said at the beginning, that sounds
like a good prompt, right?
Popped In. Cha it in.
I actually did that while, while we were talking.
I was trying to find the, the,
I was, I was gonna put that in.
I couldn't find the question to copy it.
Yeah, so I'll, this is a great question that Melissa
re-added in the, the q and A as well.
And I think this has actually come up a few times
and some of my conversations with the network,
and someone sent me a meme the other day of, uh,
I don't know if you've seen the movie Idiocracy
and kind of the future that, you know,
humanity could potentially go to and,
and she kinda shares like employees entering the workforce
for more entry level roles, right?
They're coming in with entry level skill sets
and, um, a lot of times to get to a more advanced skill set,
right, you need to practice
and build those, those more entry level skills like writing,
like to become a better writer, you need to write a lot.
Uh, and I think about even the scenario, Taylor, you added
of, uh, your, your teammate created this app, you know,
using prompts, but you had a more advanced
engineer review it and see if it's a good quality,
but for that person to be an advanced engineer, they had
to go through maybe stages of of development, right?
So I think there's a fear maybe of, of how do we make sure
that the entry level people
that are com entering the workforce, you know,
are continuing developing skill sets,
especially if AI is starting to replace some
of those tasks on the front end.
Like what does that, I guess maybe what does the future,
how does that scenario play out
or what are some thoughts around kind of
that potential reality we're going into?
One of the questions I got recently from a new hire HR
generalist on our team was
what skills can I learn over the next year
that are gonna be most valuable?
And I said, the number one skill you can learn
as an HR professional right now is prompting,
is understanding how it can work.
Because now a new hire isn't inherently
restricted by the years of experience they have compared
to the output that they can generate.
And the example I gave earlier
around our talent development partners, some
of them have never, never been in instruction and design
and they're putting out PhD level work.
Now, there still is likely needs to be a human in the loop
that does have that experience
that can monitor that quality.
That's important to do.
You don't want to find yourself as one of those lawyers
that's putting out stuff that's not true or accurate.
That said there, if I was in the emerging part of my career,
or if I was unemployed
and just getting outta university, aside from looking
for a job, I'd spend all my time on deep learning.ai
or go to open AI's free content
and learning prompting, learning how to do this so
that I can specialize in the field that I want to get into.
And hey, if you're into video games,
you could potentially make one,
like there's nothing holding you back from creating a
portfolio right now
that will make you stand out even in a
sense of unemployment.
And generally speaking, I, I hear this anecdote
that unemployment at the earlier age of, uh,
new hires coming into the workforce, that that part of
that workforce, if you look at Jolts data,
has historically been had higher unemployment than those
that are later on in their careers.
So that not necessarily a new problem, I think it's just a,
going back to an old anecdote, but promptings the key.
Well, we have one minute left here.
Uh, maybe we just go around the horn quick
and I'd love some closing thoughts, especially
as you think about, you know, majority
of people are attending here today are,
are in similar shoes as yours.
Maybe they're trying to get the organization involved
or even their own department.
Yeah. Any closing words of wisdom you would provide for,
for some of these groups trying to get into orbit
as we, as we talk about?
I would just say, um, oh,
did he just freeze? Oh, he's back up,
Back.
Sorry about that. The one thing that I, one piece
of advice I would give any HR professional right now
and really ever, uh, is understand your business.
Understand how your business makes money,
understand the commercial models
and understand the work that the people closest
to the revenue are making.
Because regardless of AI or
otherwise, if you understand the problems they're solving,
you'll be able to figure out how
to apply whatever the latest technology is to go
and support that problem solving.
I, I would add that as HR leaders, we,
we must shift from a headcount driven mindset to one
that's capability centric,
and AI is going to be a core part of our strategy,
but that shouldn't come in a pendulum swing
to everything else that we need to drive for the business.
And exactly what Josh said, if you're tying this back
to the business, if you're tying your use cases back
to the business, you will continue
to be successful HR leader.
It's amazing. Well, everyone joining today,
can we give a final round of applause
and appreciation to Taylor
and Josh for, for bringing their expertise
and their experiences to this?
Yeah, this was amazing.
Uh, huge shout out also to Dave and built by people.
Uh, Dave, this was awesome to collaborate with you on,
on bringing our communities together
and bringing this thought leadership to our, both
of our networks.
Um, so yeah, Dave, this has been a blast
and any parting words on your end?
Thanks. Thanks all for attending and showing up.
And thanks to Taylor and Josh for, uh, taking our,
our calls on, on LinkedIn to participate in this.
This is, this was awesome and we'd
love to do it again sometime.
Absolutely. Thanks for having us. Thank you all.
So for everyone else, we shared their LinkedIns, feel free
to reach out to them.
Maybe you'll get them on, on one of your sessions as well.
Uh, we shared built by people's podcast as well in the chat,
so make sure you check out those episodes.
Check out Josh's
and Josh Burson's episode that they just did together
as well, the SHRM codes and everything in else in there.
But thank you to all of you for joining too.
Taking time outta your day to level up your own skillset,
your expertise, sharpening your craft
to build a better world of work and,
and bring these companies into the future.
I mean, as, as a, a leader of this community, it,
it warms my heart and, and makes it worthwhile what we do.
So thanks for joining today.
So that said, that wraps us up
and, uh, yeah, have a great rest
of your afternoons, everyone. Take
Care.

‍