#13: Get Better Employee Engagement Metrics with AI
Welcome back to another episode of The Junction.
All right, Chase, today we're
talking about employee engagement.
I'm really excited about this one.
I've always felt very passionate about communications,
keeping people in the loop, and just
understanding the health of an organization.
So let's just jump right into it. Yeah.
People make organizations move without them.
You really can't do a whole lot.
I mean, unless you automate your entire
business, which that'd be really impressive.
But even then, there's still one person involved. Right.
We found that to keep people engaged,
you have to talk to them. Right.
You have to engage with them and see how they're doing.
And that's something that's really a struggle
for maybe nerdy folks like us. Right.
We're just like, oh, I have to
talk to you to figure that out.
What do you mean, for you? Speak for yourself.
Yeah, I mean, I'm the ultra nerd.
So one of the things we thought of, I think this
was a couple of years ago, was how do we kind
of gauge employee sentiment in an automated fashion on a weekly
basis so we could have ongoing trends to figure out, like,
well, Mel is not doing so hot this week.
Or, she's been riding a high, and all
of a sudden she dropped real low.
Something happened, right.
Like, that's a leading indicator for us to go
have a conversation or for her manager or for
her boss to come and say, hey, just wanted
to check in and see how things are going.
Well, as an organization grows, you go from being,
like, shoulder to shoulder in an office, being able
to maybe pick up on more physical cues, like
somebody's body language or just by virtue of being
across the way from them.
You can ask them how they're doing.
But then there's also the so as
you start to scale, that gets harder.
You also have organizations that have never
had an office, and they are completely
remote, so that model doesn't really work.
But then you also have the level of anonymity.
Maybe even if you can attribute the survey response.
If you're like gauging how people are feeling week
to week, there's maybe not the comfort level to
tap your boss on the shoulder or tell your
coworker you're having kind of a crummy week, or
whether it's personal or work related.
There's just that layer of I don't
know, kind of removes the awkwardness. Right?
It's just like, hey, I'm not doing so hot.
And that's what we wanted to uncover.
So every Friday, I think, at noon in Slack, each employee
gets a message, and it asks, how was your week? Right.
And there's kind of three options.
I think it's good, meh, and bad.
And of those three, you pick one, right.
And the thought is that every week, you just
pick one, and there's not, like, an underlying tone
behind it, like, oh, Mel picked bad every week.
But the thought is that that's a way for you
to let us know that, hey, maybe something's up.
Raise your hand.
Yeah, even if you don't.
I've seen it before too. Where? I don't know.
I kind of fall into the I can
always find something good in my week category.
But I started to really like, we look at this
every week, at least as a leadership team, to sort
of gauge, especially if we have we've actually noticed a
trend when we do certain employee engagement type activities.
We've got a bunch of green we
never have, like a red and yellow.
Not never, but rarely. Right.
And you see things that you would come to expect,
like if there is an outage or something that requires
folks to work more irregular hours or longer hours that
may take them outside of the bounds of comfort, and
we see, oh, there's a bunch of people who are
yellow this past week, and we understand why.
So there's definitely some things that you may
not be surprised by, but it's really nice
for the ones that you are. Right?
Well, it's great because this
automation is so simple, right.
What we haven't done is put 50 questions in front
of people and ask for 50 different types of answers.
The goal is just to get an overall engagement.
And part of the process, though, at least what
we talk about is there's people, process and technology.
A lot of folks I've been probably on
LinkedIn too much lately, and a lot of
folks love to focus on one thing, right?
Like, well, let me focus on the technology.
Let's talk about Chat GBT, or let me focus on
why it's really important to get this process understood.
Nobody really talks about the people, and if
they do, they're only talking about the people.
It's all of these things combined that help
you understand that employee engagement from kind of
that leading indicator path that we're going on.
So this tool, I think, has been in
place for years, or this process, and it's
helped us uncover things for managers where the
manager maybe isn't fully wrapped in right.
Or read in on what's going
on with that individual person.
And it just gives us that nudge like you were talking
about, like Mel just put in, like, maybe you should go
have a conversation with her and see what's going on.
What we're not trying to do, though, is
figure out, like, well, why was she red?
Well, let's figure that out.
You're not automating that piece of it.
There's a balance.
The technology has enabled us to pay attention.
And now, if we're not already and then the
human element of engaging with that person, whether it's
over slack or picking up the phone or taking
them to coffee to say, hey, I'm noticing a
trend, or I saw that acknowledging that they flagged
an irregular response and now opening up a dialogue.
And hopefully you can get to a path
of resolution or alleviate the stressor, particularly if
it is work or task related. Right.
Well, in the beginning and still today, we want to
be an organization that we want to work for. Right.
Like, I don't want to come up on Sunday.
I'm like, oh, I got to go back into work.
But this, I feel like, is probably
a common thing across other organizations.
You threw an article from Fortune in here that the title
is too many CEOs don't know what their workers need.
Employee engagement surveys can make
that problem even worse.
And I think as I read this article, what kind
of stuck out to me was that most folks just
want you to be doing good, and they don't want
to deal with the fact that you're not doing like,
I don't care if Mel's Red forget it.
Just make her happy. Right.
So we can keep trucking down
the road, and that's the problem.
And then it's like, well, let's just throw
more surveys at them so we can try
to figure out what's wrong or worse.
The companies that do one annual survey a year
with the 50 questions that you're talking about, I've
never worked for an organization like that.
I'm a huge fan of collecting feedback, so
I don't want to knock a survey.
And given this circumstance, I mean, I just
went through one the other day that was
more customer facing, and I stepped through I
don't know how many pages towards the end.
I'm a little like, okay, you've collected
all of my feedback as your customer.
I did it I think I do it because
I'm in marketing, and I know how much the
marketer on the other end of that is super
excited to get that insight about their customer.
Did you also connect your bank account and tell
them your DNA sequence and your blood type?
I'm not supposed to do that.
No, but in all seriousness, so these companies collect.
They ask a ton of questions at the end
of the year, and then it usually takes months
to present it to the leadership team or the
board, and then they collect those insights.
The HR team works with
your communications team or something.
Then they roll out a series of communications and
direct touch plans between manager and staff, and it
takes a long time to act on it.
Yeah, it's indicative of the fact that they haven't
built out processes behind all of these things. Right.
It's like, well, let's just ask
them a bunch of questions.
In fact, in the Fortune article, for instance,
it says, a poorly worded question might ask,
do you feel management supports your professional development?
I've seen that question in lots of surveys that I'm
taking in, and it's like, well, if they were really
worried or wondering about that, wouldn't you focus on connecting
with me directly and just asking, well, we went and
did a lot of things together. Right.
Do you feel professionally supported.
It's kind of like you've lost touch, right?
If I have to ask you that, then I don't really
know you or I don't understand how you as a person
are affected by the organization and the work that we do.
And if I don't have that, well, yeah, a survey might
help me out, but I'm already a step behind, right?
Like, if I have to go read the paper to
figure out how Mel feels about it, we've already lost.
I worked for an organization years
back that we partnered with.
Uh, he was a founder at the time.
His name is Stephen Huerta. Shout out to him.
He started a company called Workify, and he
took this principle of we need to collect
feedback in smaller, more like regular pulses.
And so it could be a weekly thing.
Then they would have quarterly pulses, and then you
would do sort of an annual roll up.
But they really emphasized the fact that you
don't have to tackle all of these issues
or things at once in one big survey.
Let's focus on one area of the organization that you
want to tap into or don't understand, or you want
to engage or understand the employee engagement scores there.
And then we can take those little bit of
insights and roll out small iterative changes and then
do the same thing with different topic areas.
So that's one way that I've seen it done very well,
and that was several years ago, and now I don't know.
I'm sure there's other newer
capabilities within that platform itself.
But then also, I just actually did a Google search on
AI tools for employee engagement, and there's this list of seven
on some random blog article that I pulled up.
So these things are out there, and so you don't
have to go build a Slack integration like we've done.
We actually built our own tool. How does it work?
Is it built in salesforce? Break it down for me.
Yeah, I'll break it down for you.
No, man, I'm definitely never going to be cool.
We took Salesforce and another tool called Workato,
which are big partners of and obviously Slack.
And so between those three, there's a chat bot
type functionality inside of Workado where we can pose
questions to all of the active employees. Right.
So we have people come in the door.
We need to add them in.
It'll automatically do that.
It'll ask them the question every Friday.
If they don't respond, it actually asks them again.
It kind of bumps the question.
I think it's on Sunday when they do respond.
We've got a data model inside of Salesforce that then
attaches that to we call it a resource record. Right.
So every time I've ever answered
that question, you have that record.
We go way as long as I've been here.
Yeah, three years.
For every time that I've said good or bad. Yes.
All the way going back.
And I kind of get the sentiment even from
you talking that it's kind of Big Brothery like,
oh, wow, how much data are we tracking?
But let me ask you, there was a thought.
I think this was like right when COVID was coming
around and people were going all remote and they were
taking three and four jobs or maybe quiet quitting.
How much data is too much?
If we've got data going back three years
on your three responses, should we could we
maybe not use the sentiment in your calls? Right.
The transcripts.
Let's start with the data that you have going back
three years of just asking me how my week was.
I actually think that's no on the Big Brother.
I think that's fantastic because can we
get to a place where there's more?
Can you predict maybe there's like, seasonal
I don't think I do this. Okay.
But again, I generally feel like I
can find something good in my day.
I try to answer honestly.
If it's been a little bit of a hairdo of
a week or something, that to just try to flag
it from a maybe I'm drowning and I haven't been
good about communicating with my manager about it.
But I do think that that's really neat because it
takes a lot of data for these models to learn
on, to train on, and it's just one simple indicator.
Now you start getting into listening to calls.
I think that's a really interesting point.
We've talked about that in the context
of sales calls, and we've talked about
that in the context of recruiting.
Now you're talking about it in the way of how is Mel
feeling this week or last week or three times a year?
She hits a wall, maybe.
I don't know.
Seems like an awful lot of data, though.
And what about the employee that never gets on calls?
We do have some team members that are not
all customer facing, or they may have three to
5 hours of more, like internal meetings a week.
Well, we've talked about this before.
Maybe we judge how many emails this person sent.
And I wish I had that article in front of me
because it talks about how some of these bigger orgs are
watching, like, well, how many times did you log in?
How many emails did you send?
How many phone calls did you have?
Even with present sensors, like in the buildings right.
They walked through the door.
I'm sure there are organizations that are going way
too far down this to just collect a bunch
of data and then figure out right.
But I keep going back to in all of our
discussions, it's about the people, the process, the technology.
If you have way too much technology, people
are already kind of lost in the weeds. Right.
You already lost. Right.
And it's going back to this focus
of, well, what is employee engagement?
It's ensuring that the employee wants the
job they get it and that they
enjoy it, that they're actually learning.
And all of these things that we've talked about,
if you're going to automate them or you're going
to add some level of AI, some kind of
sentiment analysis, those are easy things to think about.
It's a lot harder to think about, at least
for me, kind of being the nerdy guy, right?
Like, well, let me sit Mel down and let
me have a one on one conversation and ask
her these really difficult questions, because now I'm already
like, let me just tech it.
Let me just throw some surveys at
her and figure out what happens. Well, okay.
So let's step back for anyone listening, if you're
not already doing maybe something like a weekly pulse
check, something that has worked really well for our
organization, we also look at we are a professional
services organization, so we do track times, so we
look at hours logged.
And so that paired with you
can look at multiple data points.
So if you can see someone was red
and they worked 62 hours last week, then
you have data to lead your conversation.
These things I don't think should
ever just be used in silos.
So we're big on that, connecting the data.
But right now, none of this beyond us receiving
a time submission from the team member and receiving
the I'm feeling this way this week.
The rest of it is a manual.
We look at it as a team via a
salesforce report, and then we act on it.
So if you're not already doing it, maybe consider that and
then take it a step further with, can we combine some
of these things to maybe do, like, a roll up score
that can help us be a bit more proactive in how
we lead our teams and lead our people?
Yeah, I agree.
Mel, actually, I have a question for you.
I'm looking at the data, and this indicates
that you have never been read in the
three years that you've worked here.
Tell me the secrets.
How are you never read?
I'm in marketing.
It's really fun.
Well, I'm looking at mine and your data.
I've never had a red.
You've never had a red.
I'm looking at mine, and I've had in the three
years that you've worked here, you haven't had a Red.
And this goes back to February 2021. That's accurate.
I mean, I started late 2020.
I've had 1234 Red weeks up until last week.
I think it's got to be pretty like scorched earth.
Something is nuclear.
I'm just making it to work.
I don't think I want to be here that week.
I don't know.
I'm also like a ten on the logic
for the culture index, so you might not
even know that I'm feeling that way covert.
Yeah, I don't have a secret to that.
But again, I do think it points to the fact
that you can look at three years worth of data
on me and say generally, maybe that means maybe you
can look at someone's threshold for certain things.
Right.
You kind of can see even through certain times.
Not trying to pump myself up
here, but maybe there is that.
And there's also probably just a little bit of level
of I know after the first time I put meth,
my manager reached out to me, and I didn't know
that the team looked at this every week at that
time, and I was like, oh, they know.
So if I put meh or bad, someone's going to talk to me.
So, I mean, just keep those things into consideration.
I think it's none of this you shouldn't
roll this out and not tell the team
why that starts to feel big brother.
You have to remind your teams, we're doing
this so that we can continue to gauge
the health and well being of our organization.
It doesn't replace one to one managerial conversations.
It does not replace you feeling like you can come
in and tell me that you're having a crap week.
But it's a tool for us to look at
the organization as a whole, along with other data
sets, timesheets, events, outages, just a variety of things
and be able to see how those things impact
us and then lead our teams well through it.
What's really interesting about this is you can
see the trend for a specific person. Right.
And so I don't know what that individual was going
through in that moment, but I do know that it
kind of dips down and then it dips back up.
I wish we could do, like, a marker the
way that you have in SEMrush, Google Analytics.
You could kind of be like, this event happened here.
Not to say that, to necessarily attribute it to that.
That would be interesting because if you saw, like, a
spike or a flag, you could kind of point on
the record, well, there was this particular change.
Right.
Well, it would be nice to utilize the
data beyond what we're already doing to further
define, hey, in this moment, this thing happened.
And because of that, three of
the product dev folks right.
Were not feeling so hot, be like,
well, we probably shouldn't do that again.
Whatever that thing was.
Whatever that thing was.
Yeah, I think that's really insightful.
All right, well, let's go on to headlines.
I love this part.
So IBM says 40% of the global workforce
will have to learn new skills over the
next three years due to AI implementation.
This headline feels a little
like rinse, rinse and repeat.
Yeah, same stuff, different day.
What makes this one new?
I've been spending a lot of time trying to figure
out how AI is going to change our world.
There was a ton of hype at the beginning of
the year, and now it seems to be trailing down.
But I do believe that what is already in place, like
with chat GPT is going to accelerate your productivity to a
level that prior to any of this, wasn't there.
And so if you don't keep up with what's going on,
even with just like regular old prompting regular old like it's
been around forever, but if you don't keep up with Chat
GBT and how it can expedite some of the things that
you're doing, well, you're going to be behind the person that
is right before you go off of that.
How does one keep up?
That's such a broad term.
I've heard some suggestions
around enrolling in classes.
I know I personally try to register for webinars on
these topics, but what's a good way for someone to
actually gain a skill set or understand basic principles beyond
listening to The Junction and attending a webinar?
And what other suggestions do you have?
I actually just figured this one out recently.
My LinkedIn fee has been terrible forever and I
really haven't ever enjoyed LinkedIn beyond the Rolodex.
But I realized that it's because I've never really kind of
massaged my filter, the people that I follow and the topics
that I'm in and the groups that I'm in.
And I just recently started doing this and it made
me realize that I've been doing LinkedIn all wrong.
But there are, to answer your question, algorithms whack
because you haven't been in there and engaging.
Oh, totally.
I'm like the worst social media guy ever.
But I've started following people that in
relevant topics that are interesting to me. Right.
And now all of a sudden my
feed is a little bit more exciting.
Opening up a whole new world. Oh, totally.
Well, this is like if you want to get involved in
understanding how AI is changing the world, go follow some people
out there that are at the top of that list.
And I'm going to call out a
name only because I recently found him.
His name's Ruben Hasid.
Hopefully I got his name right.
But looking through his story, it was almost a
year ago when he started doing LinkedIn and posting
chat GPT videos and how tos and all these
things that you can do with this stuff.
And while most of it, it's kind of like yeah, I mean
I could figure out how I would use that in my work.
His thought process, right?
Like how he does these things,
how he's connecting the dots.
His video, I think, from today is taking YouTube
names from YouTube, executing some not necessarily code, but
he's just doing something unique to get the titles.
He's putting that in Google Sheets and then he's pasting
all that in the Chat GBT to get the most
viral to generate a viral YouTube video name.
Right.
The response from GBT is pretty good and
I'm not posting tons of YouTube videos and
I'll never have a hundred million people following
me, but not with that kind of attitude.
You're right.
I'm starting today.
But what he does is it's the
combination of those where he takes chat
GBT, Google Sheets, this kind of browser.
I'll call it a hack, but it's not really a hack,
but he's combining all three of those to expedite this one
thought process of if you were trying to name your YouTube
video in a way that was most SEO viral, right?
If you did, like, you would expedite
how fast it would take you.
It would be so much faster because of this.
And that's where I think a lot of people
are kind of missing the boat on this AI
stuff is that it's not just chat GBT.
You have to connect chat GBT to your
data, to these other programs to accomplish this
one thing, and that's really challenging.
So the next headline is, what if generative
AI turned out to be a dud?
Sometimes I wonder that myself.
In playing with chat GPT and anthropics chatbot,
you kind of wonder, Is this going anywhere?
And I think there's probably a sentiment, especially if
you've read this article, like, okay, that was cool.
It didn't really help me, but all right, whatever.
I'm going to go back to my work.
But I think what people are missing is
the idea that it's not just chachi, BT.
It's all of these other things in combination that
are going to help you be more productive, to
be a better person, to solve these problems faster.
Like, you name it.
And we've only got kind of one
piece of the pie right now. When you look at chat
GBT, it's like generative content.
That's kind of what people have
kind of honed in on, right?
HubSpot has chatspot.
We've seen some of this with kind of email generation.
Everybody's kind of focused on, well, let
me write the email for you.
What people aren't focused on is what this
guy, like I was talking about earlier, what
Ruben's doing, and he's connecting multiple platforms to
then do something really unique.
And that's probably the biggest challenge because
we're also focused on just one part
of that people process technology.
We're just focused on one thing, not all of them.
You talked about in an earlier episode that there's
a lot of people using it to do things
faster, but they're not using it in the way
that gets them insights that they wouldn't otherwise have.
So I think that's kind of where
you're going with connecting more applications.
It's like supercharging those insights that you
may not get from just connecting applications
or even doing it manually.
Yeah, I think what I've found is that it's not
going to be, at least in today's model, what we
have access to right now is not something that is
going to do something totally new and unique.
Like, it's not going to go figure out how to
bust the stock market and make you a billionaire overnight.
But what it can do is it can help
you replicate things that you're already doing manually.
Right.
Or maybe things that you haven't tackled because you know, it
would take you a very long time to do it. Right?
Like the transcripts. Right.
We're taking the podcast transcripts, we're dumping them in
there and we're having it generate the topics.
Okay, well, that's not I mean, I don't
think that's necessarily revolutionary, but it's the time
that you've now saved so you don't have
to go back and do that yourself.
Those are the, like, forget the transcript
piece and pulling out the topics.
It's that idea that I can utilize this thing
to do something that I was already doing to
speed up what I was doing before.
One of the things that this article also points out,
they actually pull in a tweet from someone saying, I've
been using Chat GPT for half a year and I
really don't see any valuable use for it.
I'm kind of disillusioned.
My friends feel the same way.
I think I've seen this a little bit in
some of my more personal circles professionally working for
a technology consulting firm that specializes in automation.
There's a lot of urgency around this here internally right
now to understand how can we use AI, how can
we adopt it, how can we roll it out, how
can we deliver on it for our clients?
Whereas I'm seeing a disconnect just
in kind of my personal life.
People aren't even using it.
I'm like, you know, you can go out and just do
this just a little bit faster or a little bit better.
And I don't know if you've seen the
similar thing, like, are we feeling it because
of the industry that we're in?
And there's still a ton of people out there
that don't have a ton of urgency around understanding
how it works and why they should adopt it.
I kind of see a little bit of that in
my personal life, but that's primarily because people don't fully
understand what Chat GBT can even do, right?
There's a lack of education on that front.
And the people that do have kind of figured out, at
least with this one tool, it can do some cool things,
it can do some tips and tricks and help me out.
But it's not like for me as
an individual, something that is completely revolutionary.
I'm going to go start up a new
business and automate the whole thing, right?
It can help you write some text.
And that's what I think most folks kind of feel like.
I read something the other day that reminded
me it was an article around SEO and
know that's this whole don't sleep on Chat
GPT marketers because that's where it's going.
It's basically your next Google search.
It was kind of what was happening with TikTok.
We started exploring, should we be on TikTok?
And we were for some time.
And it was the use case, or the
compelling use case for me was we want.
To be where our audience is.
And so if they're searching on how do I
integrate salesforce and HubSpot, we wanted to at least
be in that conversation on that platform.
I definitely see a generation of people
that will eventually be using Chat GPT
the way that we use Google today.
I would say that I probably use OpenAI more within the
work context, and I never think about it outside of, like,
other than if it comes up in a conversation.
But the other night I was trying to I don't frequently
use recipes, I just kind of prefer to cook by feel.
And I went to Google to know, how do I do this thing?
I could have totally punched it into Chat GBT.
You should do it tonight. I should.
I'm going to go home and do that.
So I don't know, that's where
I'm kind of seeing it going.
Like there's just still this huge gap.
We keep saying, adopt it, learn
it, it's not going away.
But then I'm seeing this disconnect in my more
personal circles, and I'm trying to figure out for
the person out there listening that hasn't played around
with these tools, where should they start and why?
I think that's totally fair, and I think
it's primarily because Chat GBT, not the OpenAI's
Large language model, is not going to exponentially
increase your organization's processes on its own.
You have to connect it with something else, some
sort of data source, some sort of roping it
into some level of automation or integration, and then
you'll start to see that exponential value.
One of the things that people don't talk about a
lot right now, and I think it's mainly because it's
very much a coding thing, is that through the API,
you can give the model access to function, right?
So in the documentation, what it refers to is like, here's
a function that allows you to type in the city.
And in the function, we'll just
return the temperature right now.
So if you pass that in and let OpenAI in the API
call know that it now has access to this function, right?
What's, the city?
And it'll spit back the temperature.
It'll then like as a system message to itself.
Let me query what the temperature is in Dallas.
Well, it's like a million degrees, right?
110 yesterday, 110.
We're not well here.
I don't know when this episode is going
to come out, but we're recording it.
We're melting while we chat.
Summer, it's awful, but it knows that it
has access to this kind of program, right?
And then it pulls that 110 out and now it
responds to the user with real time information, right?
It says, hey, Chase, the temperature in Dallas is 110.
It's these functions that are going to exponentially
increase the value of GPT for your organization.
And part of that disconnect is, well,
now I have to code something, right?
And it's no longer chat.
So I think it's going to take some time for this
to be a real big deal, but it's there now.
So understanding how it works to begin
with without the functions is probably the
thing that you should focus in on.
It's all good insight, per usual.
All right, well, that's all we
have for today's episode employee engagement.
Automation.
Can't do it without the people.
People, process and technology.
That's what we're all about.
So thanks for tuning in, as always.
If you have a question, comment, or if
there's a topic out there that you would
like to see covered on a future show,
please email us at thejunction@ventechnology.com.
In the meantime, keep it automated.