00:00:00
The labor market, like many
other sectors, is dealing with
00:00:03
the influence coming from
artificial intelligence. And one
00:00:07
component of the labor scenario
that is changing is how people
00:00:11
go about promoting themselves to
get a job, things like the old
00:00:16
fashioned cover letter. Judd
Kessler is Professor of Business,
00:00:19
Economics and Public Policy here
at the Wharton School. He's also
00:00:22
author of the book <i>Lucky By
Design</i>. He wrote recently about
00:00:27
the cover letter going away, and
he joins us right now. Judd,
00:00:32
this is an interesting topic,
and I guess, like many things in
00:00:36
life, even the cover letter
changes over a period of time.
00:00:41
It's been around for hundreds of
years. So it's— you know, dates
00:00:45
back to the 15, 1600s, that era.
But you know, technology
00:00:51
changes. And what is a useful
signal of your interest in a
00:00:58
firm and your fit for a firm, it
changes as technology changes.
00:01:02
Right. And obviously, in this day and
age of artificial intelligence,
00:01:07
I would imagine it becomes a lot
easier to write a very good
00:01:12
cover letter because you have
that extra assistance.
00:01:16
Yeah, and this is the problem. I
mean, it's a little
00:01:18
counterintuitive. We made this
thing very easy to write, and
00:01:21
now it's not useful anymore. But
that's exactly how signals work
00:01:26
in the labor market. So you
know, to paint the picture,
00:01:29
right, this is what we're trying
to achieve with a cover letter.
00:01:32
We're trying to both show the
employer that we are a good
00:01:40
catch, that we're high quality,
that we're producing high level
00:01:43
work, and we're also trying to
signal to them that we are
00:01:46
actually interested in this job.
And one of the reasons that
00:01:50
second one is so important is
it's become very easy to apply
00:01:55
to many jobs at once. So a job
application that you send to a
00:02:01
firm, you might be sending out
1000 applications like that. So
00:02:04
the firm has to decide, is this
person worth investing time and
00:02:08
energy in evaluating? If we
recruit them, are they actually
00:02:13
going to want to be here? Are
they going to be looking over
00:02:15
their shoulder for their next
opportunity? And a good, well
00:02:19
thought out detailed cover
letter used to provide that
00:02:22
assurance, because it was hard
to write, and you couldn't write a
00:02:27
thoughtful, detailed, tailored
cover letter to every employer.
00:02:31
You only sent those to the ones
you really, really cared about.
00:02:34
and that's just not true
anymore.
00:02:36
So let's circle back here for a
second. You talk about the
00:02:39
component of signals. And I
guess when you're going for a
00:02:41
job, the cover letter is one of
the signals that shows that
00:02:47
you're extremely interested
about this position. You think
00:02:51
you would be a great candidate
for this position. That's kind
00:02:53
of been the historic norm of
what a cover letter was supposed
00:02:57
to do, correct?
- Exactly.
00:02:58
And so that's—
and the reason it was a
00:03:01
good signal, the reason that
when a firm received a
00:03:04
thoughtful, detailed cover
letter, they could be confident
00:03:08
that you really were a good fit
for the firm, believed the firm
00:03:12
was a good fit for you, was that
it was hard to write a cover
00:03:16
letter like that. If it took
hours to research the firm,
00:03:19
figure out exactly what they
were doing, what they needed,
00:03:22
and how you fit in, then, you
know— if it took hours to do
00:03:25
that, the firm could be
confident that you weren't
00:03:28
writing letters like that to
every place you applied. You
00:03:32
were— you were selecting the
places that were actually good
00:03:36
fits for you. But if ChatGPT
or Claude Code or whatever your
00:03:40
preferred AI— generative AI
model is can write hundreds of
00:03:45
those letters in a matter of
seconds, then the firm that
00:03:50
receives that no longer can be
confident that you actually
00:03:55
want to work there. They— you
know, they might believe that
00:03:59
you're sending letters like that
to all the places you apply, and
00:04:02
you very well might be.
00:04:04
So you actually suggest that the
component of the job
00:04:09
recommendation may become even
more important as we move
00:04:12
forward here.
- Yeah.
00:04:13
I mean, the question is, what is
going to be a useful signal for
00:04:17
the labor market to identify
folks who are really good fits?
00:04:20
And I should say, you know, I'm
not just— when I wrote my book, I
00:04:23
was kind of speculating that
this would happen, but there's
00:04:26
data coming out that is showing
exactly this. That when
00:04:31
platforms, labor market
platforms, introduce AI tools to
00:04:35
help people write their cover
letters, the cover letters get
00:04:39
more tailored. They get more
detailed, and, you know, better
00:04:43
fit matches— look like there
is a better fit between the
00:04:47
applicant and the job posting.
And those cover letters are
00:04:52
better. They're
successful, but the firms that
00:04:55
are looking for candidates stop
paying as much attention to
00:04:59
them. They kind of recognize,
"Oh, I'm getting a lot
00:05:01
of these," and the cover letter
becomes less important. What
00:05:04
becomes more important? Well,
the signals that are not as easy
00:05:08
for a candidate to generate with
AI. Things like your prior work
00:05:13
experience and recommendations
from prior employers or other
00:05:18
people who know you well, that
might be written by AI. There's
00:05:22
nothing wrong with having the
letter of recommendation be
00:05:25
written by AI, because the
actual signal is not that the
00:05:29
letter is good. The actual
signal is that you got somebody
00:05:32
to write the letter on your
behalf, that somebody out there
00:05:35
in the real world believes you
are a good fit for this firm.
00:05:38
So should the— for those that are
out there looking for jobs, do
00:05:43
they bother with the cover
letter at this point, or do you
00:05:47
make it just about those
recommendations from other
00:05:50
companies, other employers?
00:05:51
So I think we're in a transition
period where you actually do
00:05:55
still need the cover letter,
unless a firm stops asking for
00:06:00
it, which is where I think we're
going. Eventually the firm will
00:06:03
realize this is pointless. We're
having AI write these cover
00:06:06
letters. And, you know, many
firms now are using AI to read
00:06:09
the cover letters. And so the
whole— the whole process is a
00:06:13
little silly. But right now, the
cover letter is kind of the
00:06:18
prerequisite. It's no longer the
differentiator, because if you
00:06:23
decide not to send a cover
letter, or you send a bad cover
00:06:27
letter, then the firm has to
think like, "Well, why didn't
00:06:32
they send— why didn't they just
put it through one of these
00:06:35
generative AI models? It's so
easy to write a good cover
00:06:37
letter, they must really not
care, if they're not bothering to
00:06:41
even do that, even do the kind
of lowest-cost thing of
00:06:44
having AI write a good cover
letter for you. So yeah, I think
00:06:47
we're still in a time when we
need to send a good quality
00:06:51
cover letter, but I think you
can be confident that the firm's
00:06:56
not really looking at them. And later
on, you know, this won't be necessary.
00:07:01
You also talk in your article
about the importance of
00:07:06
networking and how that kind of
plays into— obviously, people
00:07:11
have done networking for a long
period of time, but how does
00:07:14
that get impacted by this shift
that's going on?
00:07:17
Yeah. The issue is when AI
comes and makes a lot of the
00:07:23
signals that we used to send
less costly, easier to send,
00:07:29
then the signals that will take
over are going to be ones that
00:07:34
are happening in the real world.
It's kind of this great irony of
00:07:39
AI, is that we're making things
online just so easy, anything
00:07:43
that a computer can aid you with
has become very, very easy to
00:07:48
do. And so the signals that will
be effective are ones that
00:07:52
cannot be replicated by AI. That
are in the real world. And so
00:07:56
when I find somebody at a firm
and offer to take them to coffee
00:08:01
and spend half an hour with
them, chatting with them, that
00:08:05
is a signal that I'm interested
in the firm. That is a
00:08:07
relationship that I am building
that cannot be replicated with
00:08:11
AI. And as a result, those types
of relationships are going to
00:08:15
become more important. And if
what is happening is that the
00:08:21
cover letters or the referrals—
oh, sorry, the letters of
00:08:25
recommendation or the referrals
are going to replace the cover
00:08:27
letters, well, then, having more
interpersonal connections,
00:08:31
having more relationships,
having more folks who are
00:08:34
willing to go to bat for you, is
going to be all the more powerful.
00:08:38
Is this just another— another
change for the HR department
00:08:41
that's going through an
incredible swirl of change right now?
00:08:45
I think they're gonna settle on
this. I think, you know, when you
00:08:48
are flooded by hundreds or 1000s
of applicants for each position
00:08:53
that you have open, and they're
getting better and better, and
00:08:56
you're having trouble sorting
through them, right, you're
00:09:00
going to need to figure out,
what is the alternative way that
00:09:02
I'm going to cut through this
noise? And I think that HR firms
00:09:06
are getting there, and you see
it in the data from the papers
00:09:09
that are coming out looking at
this. But yeah. It's— it's a
00:09:13
period of change. I do not envy
those folks. But it's—
00:09:16
it's going to get better once
they figure out, what are
00:09:18
the signals that actually are
informative in this—
00:09:21
in this market today.
00:09:22
Judd, always great to chat.
Thanks again for your time, sir.
00:09:25
Thanks so much for having me.
00:09:27
Thank you. Judd Kessler,
00:09:27
Professor of Business, Economics
and Public Policy here at the
00:09:30
Wharton School.
