Evolved Radio

Every MSP I talk to is working out how much of the help desk to hand over to AI. 

My guest has spent two decades training service teams, and he has the data on what happens when a customer needs a person and can't reach one. 
Jeff Toister's courses have been taken by more than five million people on LinkedIn Learning. His new book, Human Service: The Skills AI Can't Replace, is built on research from over 370,000 phone calls, and he calls it a pro-AI book. 

We get into which tickets your techs should never hand to a bot, and why making a human easy to reach pushes more clients toward self-service.

This episode is brought to you by Opsleader Pro. A place for MSP owners and managers to get the systems and tools they need to build a stable and growing MSP. Part group coaching, part peer group, everything you need to run a successful MSP.

What is Evolved Radio?

Evolved Radio Podcast: Interviews with technology experts, industry thought leaders, business leaders and other interesting minds. Exploring the evolution of business and technology.

Todd Kane: Every MSP I talk to
is working out how much of the

help desk to hand over to AI.

My guest has spent two decades training
service teams, and he has the data

on what happens when a customer
needs a person and can't reach one.

Jeff Toister courses have been
taken by more than five million

people on LinkedIn Learning.

His new book, Human Service: The
Skills AI Can't Replace, is built

on research from over 370,000 phone
calls, and he calls it a pro-AI book.

We get into which tickets your
techs should hand, uh, handle,

and what goes to a bot, and why
making a human easy to reach pushes

more clients towards self-service.

Jeff, welcome to the
Evolved Radio Podcast.

Jeff Toister: Hey Todd, how are you?

Todd Kane: I'm very well.

Great to see you

Jeff Toister: Good to see you as well.

Which, it just strikes me as odd people
are listening and they're like, "I, I

can't see either of you." But you should.

You should be watching us

Todd Kane: Yeah, yeah.

Jeff Toister: your mind's eye

Todd Kane: you can check out the, the,
the YouTube feed if you like, if you

wanna see our smiley faces as well.

Uh, but, uh, we… otherwise
we are in people's ears coming

to them about AI and service.

You, uh, you and I connected years
and years ago, uh, and we had sort of

this common theme of, uh, you doing
a lot of the work in more enterprise

space helpdesk to, and service, and
that's the space that I actually

came from before I kind of found,
uh, the majority of my career in MSP.

But I think a lot of this is incredibly
relevant because I've always felt

that the work that I do in both the
operations that I did working inside

of IT organizations and MSPs was really
lending from what I learnt in enterprise

and trying to apply it in an SMB space.

And I think, uh, the, the conversation
today is, is sort of a retreading of

exactly that, is what are we seeing at--
we can learn at an enterprise level that

should be applied at SMB and the MSP
space about the human and AI interface?

This is, as I noted in the open, this is
something that's coming up a lot, right?

Like I, I, I was initially a little
concerned at how slow the a- the adoption

of AI was in MSPs, but it's starting
to come a little more fast and furious.

But there's a lot of sort of
hand-wringing and consternation about

what is the right approach here.

And I've seen the tale of
two MSPs here, for example.

One organization, you know,
we will not use any type of

AI in user-facing activities.

And then we had another MSP that I
saw, you know, e- they went to the,

the full extent where, like, they
didn't even pick up the phone anymore.

The AI was answering the phone
for them, and I was like,

"How does that work for you?"

And they're like, "Clients love it."
I was like, "Really?" Like, I was

actually really surprised by that.

So I find, like, the ground is
very shaky on this about, like,

what people could and should do.

So I'm, I'm excited to get your
perspective on this because you've

got a lot of data, a lot of history
that, uh, I think informs this.

So maybe, uh, all of that to say,
we can start with kind of your

thesis on this of like the interface
between human service, and AI?

Or, uh, maybe to start, what is h-
how do you define human service as

the sort of the title of the book?

Jeff Toister: So human service,
that's a good… I think that

is a good place to start.

It, it's, it's any service
in any environment that

involves uniquely human skills.

So a few examples of those would be
connection, that ability to just build

rapport with someone that you're serving
so they like you and they trust you a

little bit better, or understanding.

A- and understanding goes beyond just,
let's say, finding the root of the issue.

You know, "My computer's not working," but
it's understanding the intent behind it.

"My computer's not working, and
I have a big client presentation,

and I'm worried about not landing
the deal." And then advocacy.

Advocacy means I'm working on your
behalf to try to get a result.

A- and so those are things that humans do.

Now, I think there is a very important
distinction between human service and

people who are providing service, but
they're not using human qualities.

So these are people who
are very transactional.

They're, they're almost
trying to impersonate a robot.

I mean, they're, they're trying to be
AI without the capability, and that's

what I found writing the book is that
that's kind of the worst place to be.

I, I don't want an
uncaring, inefficient human.

So it's a better choice to say,
"Can I automate this and automate it

effectively?" Or in these specific
situations, will a human add extra value?

Todd Kane: I think this is huge, right?

Because, uh, I think you get to the
heart of the problem that we see in a

lot of tech industry is highly logical
people, you know, not terribly emotive.

Em- empathy is something you sort of
have to, like, teach them and train them

to some extent, and that- that's not--
that's a broad generalization, but I

think it's also generally kind of true.

Um, and, you know, this gets to sort
of the fundamental disconnect, I think,

at a lot of service desks with around
IT is, uh, you know, I'm here to solve

a problem, not to be your best friend,
and I think there is obviously a

middle ground in that, especially when
you're doing, you, know, phone support.

Uh, you're interacting with
someone, you're talking to them.

There, there is obviously an
emotional and human connection

because they're frustrated.

Something doesn't work, and
most people are mystified by

technology, and they kinda hate it.

And the, the technical, uh, sort of wizard
on the other end of this is just like,

uh, "Yeah, yeah, I know what this is.

Hang on, let me get this
fixed for you right?

And they, like, they're just not
interfacing with the humanity on the

other end of the line here or sending
them an email and not recognizing how they

come across, like what the, the implied
intonation in how you write something.

Or the one that drives me the
most crazy is like, "Hey, did

you call that client back?"

They're like, "No, I don't have an
answer for them yet." Well, like,

call them back and let them know that.

Like, that's a part of this
communication that is necessary

to service them as a human, right?

Jeff Toister: I think you, you hit
upon a big challenge and, and it, it's

made worse in a way because we're in a
world that thinks in terms of tickets.

So we, we've taken the human almost out
of the equation because of the process.

"Hey, can you help me with this?" "Yeah,
I don't see a ticket, so maybe I can't."

But you also raise a good issue about that
disconnect between, you know, you know how

to fix the issue so it doesn't bother you.

And, and, and s- and you might
even be almost amazed that I

don't know how to fix the issue.

Like, it's so easy in your
mind, so tapping into empathy

is, is extra hard there.

One, one of the exercises I've
often asked teams to do is think

about, you know, why is this person
really frustrated or anxious?

And, and if they can get to that,
then ask yourself, h- have you

ever f- had that same feeling?

Like, you might not have been anxious
because your computer crashed right before

the big presentation, but maybe your
car broke down on the way to work, and

you knew you couldn't be late that day,
and you didn't know how to fix your car.

You can fix a computer but not the car,
and now you're helpless and you're waiting

for that mechanic, and you hope that
they're cool and they show up on time, and

they don't make you feel like an idiot.

Uh, I mean, it's a similar enough feeling
for then you can ask, "Well, what would've

made me feel better in that moment?" Or,
"What did make me feel better in that

moment?" And if you understand that, then
you can probably tap into empathy and say,

"Okay, I can do the same thing for this
person I'm serving." But there's a lot of

intention behind all of that, and it is
really easy just to be dismissive of it

and say, "Look, it's another ID10T issue.

Let me just get on it and away we go."

Todd Kane: Yeah.

Well, I guess, and that kind of
gets to, you know, um, what you're,

you're noting here is sort of the
human in the loop, uh, in this.

Because quite frankly, like
an AI can solve this stuff.

Like we're getting to this place where,
you know, it can recognize what the

issue is, it deploys a script, and
whoop-de-doo, like problem is resolved.

So if that's where we're headed,
where is the human in the loop?

What value does the human interaction, uh,
provide, uh, when s- a lot of this will

start becoming sort of first contact or
first call resolution being done by a bot?

Jeff Toister: There's a few
places where humans add, I

think, some unexpected value.

Um, and, and we'll go maybe
obvious to not so obvious.

One of the obvious places is, is where AI
reaches the limits of its capabilities,

and that's where we need a human.

So those are more so edge cases or,
um, you know, we still rely on humans

sometimes to document properly what's
actually going on and what they need.

And, and, and sometimes we need a
human just to kind of suss that out.

So that's the kind of more obvious.

Um, there's a second where the human's
just not confident in the output.

A- and there could be a lot of reasons for
that, but, you know, typically the higher

the stress levels, the more urgency, the
more perceived complexity of the issue,

the less tolerant we are of entrusting the
issue to automation or AI, especially if

it doesn't feel highly competent or easy.

But then the third area is, I think,
some of these unexpected places,

let's say, where relationships
are really, really important.

And it's more than just
troubleshooting an issue.

It's more about providing confidence
that, hey, we're on it, and

we're going to help with that.

And in those cases, that's where a human
can really add a lot of extra value.

So it's, it is a little bit of a
moving target, and A- AI is able

to do more and more each day.

Uh, but I think it's a mistake to
just we're gonna put everything

on AI or we're never gonna touch
AI a- and, and it's all humans.

I mean, AI, like anything else, is
a tool, and right tool, right job,

you probably get the job done well.

Todd Kane: So what about like,
uh, like first contact, right?

Like, uh, 'cause I, I noted, you know,
the story that I heard about this

MSP that just went pure AI answer,
uh, of phone calls, and I was shocked

that the clients didn't, like, revolt
around this, uh, because that's what

I tend to see is like, you know, uh,
we're of the generation, I suppose.

Uh, I find a lot of this is generational
actually, because, like, the older

generation than us, like the sort of
people in their 50s and 60s and beyond,

they'll pick up, pick up a phone.

Like, they wanna talk to a human,
and that's very much sort of a

generational, uh, approach for them.

Um, you know, I don't
know how old you are.

You, you look like a young guy.

But in those, in those 40s and
below type thing, um, that,

Jeff Toister: No comment, and
definitely not 40s and below.

Todd Kane: Okay.

Well, you look great.

You look great, Jeff.

Um, that, you know, I'm, I'm there,
like mashing zero, like yelling agent

into the phone when I'm stuck in a phone
tree, but I'm also the first one to go

to the, the web and look for a chatbot
that can actually solve the problem for

me, and then the generation below that
is just like, uh, either I will solve

this myself, or, you know, maybe I can
send a text, or I can find a chatbot,

but I am, I am absolutely not phoning
and talking to someone about this.

So there is a bit of a generational split,
but, what would be your opinion about, you

know, having that first line of interface
with a client being more agentic?

Is that a good or a bad thing,
or is It circumstantial?

Jeff Toister: It gets circumstantial.

Uh, however, I, I think generally
there are some real advantages.

So if we think about first line,
if I have to go to a human, I'm

probably going to have to wait,
and if that wait time is tolerable

because it's not urgent, that's okay.

Um, but also one of the things where…
So AI can often answer this question

immediately, especially simple stuff.

You know, my, my internet is down.

Okay, well, let's start,
let's start a process.

We c- we can walk through a
troubleshooting process without waiting

in a queue or submitting a ticket.

That, that's an advantage.

One of the other challenges e-
especially when we're reacting to

issues is having the right amount of
information to diagnose the problem,

and that is classically a challenge.

But if we can have a conversation
with AI, it could probably guide us

pretty quickly if it's programmed right
to provide the correct information.

So there's definitely
some advantages there.

I think when it exceeds complexity,
uh, or, or there's a level of urgency

where I don't have the tolerance to
go back and forth with AI, especially

if it doesn't initially seem to
understand me, that's where the

human is, is gonna be more needed.

So sometimes it's first contact is okay,
but do we have a clear escalation path?

And, and, and what we're seeing it is
kind of all over the place, but AI is

sometimes way too tenacious in terms of
holding onto the issue versus being able

to recognize, "Oh, this is an edge case.
Let's get you to a human, and we're gonna

send you to the human with context."

Or, "Oh wow, you're super, uh,
this is super urgent. We're gonna

get you to a human." Um, I think
you're probably right about there's

some generational differences.

I think there's also
some power differences.

Your… Is your CFO or CMO going to
tolerate chatting with an AI, AI agent

because they have a need that's urgent
or are they just going to talk to the

person that they know is gonna help him,
help them right now versus I just started

with this company, I don't know anybody
yet, and I'm trying to get my, you know,

my, my workspace set up for day one?

That's probably a different
level as well in terms of what

people are willing to tolerate.

So not a clear answer, but I think it's
good to think through those things.

Todd Kane: Yeah.

Well, let's, let's explore this
a little more then because,

Jeff Toister: Yeah

Todd Kane: examples that I find
frustrating in my interactions

with, uh, with s- uh, help desks and
service desk in general, uh, it's

funny how you become sort of critical
and you see this everywhere in your

life when it's a part of your job.

Like, I never thought much about,
like, phoning, uh, you know, an

airline or, uh, support for a product
back in the day before I was in this

industry, and now I'm keenly aware
of all of my interactions with some

type of help desk or service desk.

Um, and one of the things that I think
bothers me is there are lots of places

where I almost feel like an interaction
with an AI that was sufficiently capable

would be better than the human interaction
that we have to wait for in a lot of

the enterprise environments, Right

Like, you phone, uh, you get someone
who, uh, uh, you know, is potentially

offshored, uh, they're doing their
best, But the part that bothers me

about those interactions is that
they're incredibly scripted, right?

And it's to the extent of, like,
the feigned empathy bothers me,

of like, I tell them what my issue
is, and they're like, "Oh, I'm so

sorry to hear that blah, blah, blah.

That must be really frustrating for you.

I'll see if I can help you with the
blah, blah, blah, blah, blah," right?

And it's like you can tell they're
just reading from a script, and if

you try to, like, break them out of
the script and interact with you in

some way, like, it, it just, it's,
it doesn't derail the conversation.

It just seems to, like,
go down the script.

You can't break out of it, right?

So, like, I may as well
be talking to a robot

Jeff Toister: Right

Todd Kane: The human
interaction is not additive.

In fact, it's actually a detractor from
my experience in that interaction, right?

Jeff Toister: But I think that
that's what I would consider

to be, uh, not human service.

That, that's, it's transactional service.

It's, it… That person is impersonating
a robot by sticking to the script.

They're not bringing their humanity,
so they don't add any value.

And I, I think you're right.

If, if I put that person up against a
capable, um, AI, um, AI is probably going

to be just as charming or not as as…

But, but it's probably going to, to
be a little bit faster at working

through the issue, I hope, right?

If it's programmed correctly.

I, I think we have to look at places
where humans add the extra value.

So, uh, let, let me give
you maybe two examples.

One is I was about to start a webinar with
a client, and I, I desperately needed this

webinar software to work, and it was not.

And I was able to, to… I, I didn't
see a clear option to troubleshoot it.

I didn't have time to go through
a whole script and, you know,

Clippy, the AI bot, to tell me how
to reset the software and all that.

I just needed a person
who could just handle it.

And I fortunately got a person
right away, and they solved it.

But then they did something
that AI just can't do yet.

They understood the real source
of my anxiety was that I'm

starting this client webinar.

I'm worried about letting my client down.

And I'll never forget, she
said, "I think we've solved it.

But I know your webinar's about to
start, so I'm gonna stay on the line

with you, and if anything goes wrong,
I will be right here to help fix it.

And if it goes smoothly, which I think
it will, then I'll just quietly get

off the line and you can facilitate
your webinar." That's a, that's a

different level of understanding where
that person is adding value, and I

was ready to cancel that service, and
I stuck with it for two more years,

despite it being an inferior product.

And then Zoom came along and, you
know, some other really good products

and, you know, enough is enough.

But they bought me two years
more of service from that person.

Now, the flip side of this, the
other day I was trying to solve a

problem that involved both payroll and
accounting, and it was after hours,

so I, I didn't wanna wait for a human.

It was, this was my time to get it done.

And I first started with my payroll
provider's AI bot, and they had a

hole in their documentation, which
means the AI bot couldn't solve it.

And then I go to ChatGPT, which is
able to kind of source from multiple

places, and I solved it on my time,
and it was a very helpful thing,

without having to wait for a person.

So I, I think, you know, there's
a, a case to be made for having

both, and both have their place.

Todd Kane: I think the, the, the
first example I think is what

that reminds me of is that in, in
problems is almost some of your best

options for creating loyalty, right?

Like, I tell this story all the time.

We're in an enterprise, uh,
consulting environment that I was in.

Um, we made a huge error and deployed
like s- over 60 servers into the

field for this very large client
with the wrong image on them, right?

And like a huge i- issue because we paid,
you know, $400, $500 shipping to get

these, all of these servers out and all
the work that had been done to build them.

Um, and everyone's like, "Oh my
God," like, "This is such a problem.

We're gonna-- Uh, we have to pull
them back." Like, "The price of the

project is gonna have to go up."

And the CEO is like, "Nope, this is our
error. We own this. We are not charging

them any more, and we're gonna communicate
that to them." Two things about that.

Like, um, one, owning it.

Two, the fact that you own it and
that you're not charging them more.

A lot of people will s- try to hide
these issues and, and sort of s- like

circumvent to, to make themselves look
good when in a lot of cases, like how

you handle an issue is far more i-
impactful than the issue itself, And

they became an incredibly loyal client of
ours because of what we were able to do.

Same thing for you, right?

Like that person is probably incentivized
by their call handle time and the number

of, uh, uh, tickets that they're able
to close through the course of the

day, but they took that extra five to
10 minutes to hang on the line with

you to make sure that you were okay.

They didn't have to do that.

But that simple action of rather
than saying like, "Looks like

you're good, Jeff. Let us know
if you need any, any help in the

future," and then click, right?

Uh, they h- they held on, and that, that
created that sense of loyalty for you.

So I think that's massive.

Jeff Toister: And you're bringing up
a, an even broader issue, which is, you

know, uh, taking ownership of mistakes.

That's, uh, that's something that
does, it, it, it creates a s- a

different set of trust and builds
a, a relationship with your clients.

So, you know, I don't know that
a machine knows how to do it,

but, uh, people certainly can

Todd Kane: Yeah.

And maybe we're sort of getting to
that, that state of the world as well

is like when an AI does something
wrong, like how do they, how do

they take responsibility for that?

Where does the responsibility lie, right?

Like this gets into some weird
spaces on, uh, you know, errors

and omission, uh, insurance.

Like our AI did something, uh, and it
went rogue and did something it shouldn't

have in your client environment.

Like that's something to consider as well.

Like what's the scope of control
that you give these things?

But also, you know, it can't really be
held responsible for the good and the

bad things that, that, that it, it should
be doing or shouldn't be doing as well.

yeah, this

Jeff Toister: We've seen
a case already about this.

Um, I, I, I'm, I'm not gonna, I'm
gonna get the details wrong, but I

think there was a case, or maybe it's
going through the courts now about are

we responsible for what our AI does

Todd Kane: Mm-hmm.

Yeah.

No, I wouldn't be surprised.

I think that's, that's
super, super relevant, right?

Because e-especially as
they become more agentic.

Jeff Toister: Ja

Todd Kane: and we've seen
sort of the previous cases.

I've personally not run into these
issues, but like when Clawbot and

stuff s- first showed up on the scene,
and it would just sort of rip through

and start deleting stuff for users.

Like, "Hey, I'm cleaning up your
inbox. I'm deleting the 4,000

emails that you've archived."
It's like, "No! Stop." All right?

So I've never had an issue like this,
but you know, you could see that

being an issue where, um, I don't
think it was an AI necessarily, but

I've seen those i- the issues where
automation deleted all of the backups,

Jeff Toister: Ja.

Todd Kane: in order to make room
for, for, for the new backups.

And it's like, well, no, like
we have a regulatory compliance

requirement for keeping these backups.

Like, you can't just go and delete them.

Um, so there, you know, this
is maybe gets into those spaces

of like, where is AI useful?

Where is, you know, the, the, the
human in the loop needs to be required?

So we've kind of touched on like
empathy being one that, you know,

robots right now are not terribly good
at, and it's an important component

for especially the service desk.

I believe a lot of sort of the IT support
in the future will be back to the future

of more consulting, a lot more human
interaction, because everything else on

the back end is largely commoditized.

Like I don't think having an
AI being able to resolve your

client's issues for you actually
differentiates you in a lot of ways.

It may make you more efficient, but I
don't think the clients particularly

care about how the issues get resolved
as long as the experience that you

create for them is, is positive, right?

So I think that to this place of like,
what are the skills that humans are still

highly adapted for that are necessary
in, you know, the IT future that, you

know, for now, maybe the next, at least
the next year or next three years,

uh, AIs are not particularly good at
taking on some of those tasks or being

capable of doing those human activities.

What would those look like?

Jeff Toister: I think one of the biggest
things is, you know, go back to advocacy,

that when something falls out of a
normal process or breaks, do you have

a human who's available to run with it?

So as an example, you hire a new
employee and you're counting on

your MSP to make sure that that
employee is ready on day one.

Whether or not they're ready is
critical for that day one to be a great

experience, "Hey, I'm starting to learn
how to do this job and understand the

culture," or, "My day one is spent
tracking down that computer they sent

me and, uh, trying to get the right
credentials and log in." And so there's,

uh, two very different scenarios.

I think AI can help monitor and
run processes like that, but what

happens if something goes wrong?

Who's going to find that solution
that is not in your normal process?

I mean, if that laptop you shipped
out got stuck in a, a shipping hub

somewhere because of a snowstorm,
what are you gonna do about it?

That-- And I think those are,
those things, humans are the ones

that can find creative solutions.

They can work around systems.

They can work through
bottlenecks and limitations.

They can think laterally at, at, at a
very different level and understand when

a rule is a rule and more of a suggestion.

A- and having a human available for
those more complicated scenarios or

when something goes wrong is, I think,
a good way to use your human talent.

It's okay to mo- to automate, you
know, a normal request, for example,

that, but you still need somebody
to monitor that and fulfill it.

So that's-- I think that's where
you, you run into challenges is where

companies either go all in on AI and
forget that you still have to train

and monitor it and have a fail-safe.

But also, if we're just having humans do
all of that, that's inefficient as well.

So there is a balance to be had, but those
edge cases are, to me, one of the biggest

opportunities to have humans involved.

Todd Kane: Yeah.

You, you mentioned kind of the
advocacy, I would say, versus process.

Uh, I think this is a, a big piece
that, that I'm fascinated by because,

um, I often say that, um, critical
thinking skills is one of the

hardest things to actually teach to
people, in my opinion, because so

much of it is experiential, right?

Like you can, you can give them sort
of, uh, value-driven judgments and, uh,

give them as much process as you can,
but either they get stuck in following

the letter of the law and like we talked
about of like, "I'm supposed to follow

this script. If I go off script, I'm
gonna get in trouble, so I'm just gonna

follow the script," and, you know, it
devalues the relationship that I have.

Versus, you know, "Okay, I'm gonna go
off script here outside the scope of

what I'm responsible for and do something
amazing for this client that may help

them, make them a memorable experience
that they really value." But, you know,

I also can't bury four hours on every
ticket that comes into me, and I think

the judgment between these is really
difficult for people, and how you actually

train that I find is quite tricky.

Do you have any thoughts on sort of
like how you educate or give people the

tools to figure out where should I draw
outside the lines and look for something

creative and really bury some time on this
versus I just need to follow the process,

and if the process is wrong, then I'll
tell my boss about it after this call?

Jeff Toister: It's the system
and the culture that usually

creates one or the other.

A- and so I'll give you
a couple of examples.

For the s- typical MSP, if you're
listening to this, a- ask yourself,

have we collectively defined what
a great experience looks like?

And the answer is no
for most organizations.

A few have, and that's a huge first step.

But if we haven't collectively defined it,

yeah, Yeah

Todd Kane: that they've
defined it is it's, "I want the

customer to feel great," or, "I

Jeff Toister: Right

Todd Kane: customer outcomes.

I want them to, you know, give us a
good, uh, CSAT or NPS score." Right

Like, that's, that's
not really a definition.

That's just sort of like, how do
we know that this is working or

not, Right It, doesn't give you
the guidelines of what to do.

Yeah.

Go ahead.

Jeff Toister: it, it, it's, it,
it has to be something like,

what did you do for your client?

And, and how did you… If you're
on a good day when your client

is happy, what are you doing?

And you can test that.

Ask a group of employees at
random, and if they all give you

the same answer, you're on target.

But if they don't, what people probably
default to is process and procedure.

And if you look closely at the process and
procedure, we're trained to follow them.

So critical thinking is almost
pushed out by process or by procedure

because, uh, you know, my boss is
gonna notice if I didn't follow the

process, and maybe I'm gonna get in
trouble if I don't follow the process.

So, uh, now, I don't even, I don't
even worry about going outside of that.

But I, I will tell you that
I was lucky to learn a lot of

lessons from, from customers and
clients, and years and years ago.

I had a client, um, I, I, I had given this
client all kinds of excuses why I couldn't

get something done, and it was all
because of process and procedure, and my

client said something I'll never forget.

He said, "Isn't that what they pay you
for?" And it just, it stuck with me.

It like, wow, I'm adding no value here,
and I need to find a way to add value.

And I worked for a big bureaucratic
company, but I've, I worked hard

to find a way to get things done.

And then I started visiting all
these other departments that I

thought were holding me up and
realized they had challenges, too.

And so instead of just saying, "You're
a jerk and you need to work harder for

me," I realized that wouldn't land.

I started asking them, "Well, what would
make it easier for you to get this through

faster?" And they would say, "Well,
you know, this is how you do it, and I

don't know why no one knows this." So I
start kind of documenting and building

relationships, and eventually I'm able
to get more done, get more done quickly.

But that took a lot of effort,
and none of it was encouraged.

In fact, my boss actively
discouraged that kind of work.

So I think as leaders, we have to
define what great looks like, make

sure everybody knows it, and then
we need to make sure that all of our

processes and procedures are aligned
with that definition of greatness

Todd Kane: I love that example because
it's something that I see so commonly in

companies writ large, MSPs in particular,
certain st- friction points between like

sales or account management and tech, uh,
and let's say the owner and the service

delivery teams or the administrative
group or, you know, the service team

and the professional services team.

Like, there's always sort of
this conflict between these

groups, and I agree with you.

Like, I would say 90% of those issues are
strictly communication and process, right?

Like, we have not talked about what this
process looks like on both sides of the

fence, and we keep on just pushing work
over to the other group and then, like,

getting upset when it doesn't get done in
a timely fashion without any consideration

as to like, could this be done better?

Or if I just gave you these five
points of information before I

shuttled this work over to you, all
of a sudden it's so much easier.

It's just like, "Ah, why don't these
guys know what they're doing? How

come they can't get this done," right?

Uh, it's just so incredibly common.

So I think that's, that's a
huge takeaway is just, you

know, go over and investigate.

Like, sit with them and understand what
their workflow looks like, and then I

often tell people, like, um, if you're,
if you're having trouble interfacing with

another department, build a form that,
like, says, "Hey, hey, here's all the

information I need you to submit to me to
get this done, And I guarantee you the,

the effort will be worth it because I'll
probably get this done like three X faster

if you just give me this information."

You get their buy-in on that, and then
all of a sudden everyone's winning, right?

Jeff Toister: And you solved a problem
for them a- and you're right, now

they're saying, "Okay, well now we're
on the same team. Let's, let's go."

Todd Kane: Yeah.

Yeah.

It's, uh, it's, it's funny how so
many of these things are… I often

describe, like, so much in business
is just back to basics, right?

Like, we just do the simple things, and
then magic just starts to happen, right?

It's

Jeff Toister: It really does.

But I, I think even when you think
about basics, like, you know,

we could go the other extreme.

Like I, I worked with a team where,
you know, they, they said, "Dude,

take as much time as you need to
solve the problem." And that in

a way became part of the culture.

That the problem with that is that
if I take as much time as I need,

you s- you'd see people in like tier
three that would just say, "Well, I'm

just gonna focus on this one ticket
today," and they would ignore all

this other work that would pile up.

Or I, I saw this issue where everybody
was doing their own thing and there was

a kind of a new issue that was taking
people on average 30 minutes to solve.

But, but one person, Stephanie…
Stephanie figured out how to solve

the same issue in five minutes.

And Stephanie was frustrated because
she didn't have a mechanism to

share her five-minute solution
with the rest of the team.

So if, if you think about the
inefficiency of the rest of the

teams, like, "Take as much time as
you want," it was killing their queue.

And Stephanie's like, "Um, uh, how
do I share this?" And, and so even

having this culture of do whatever
you want is… That's not it either.

It's how do we constantly
serve our clients better?

That's the culture we're trying to get to

Todd Kane: Yeah.

Uh, there's so many rabbit holes here.

This is great.

Like we could, we could jam on
this stuff all day 'cause what that

reminds me of is, like I encourage
people, uh, to give positive feedback.

I, I have this structure around
this of like five to one ratio.

You wanna give five points of
positive feedback to every one point

of negative feedback because quite
naturally we are, uh, adaptably

evolved to have a negativity bias.

It's so much easier to nitpick and find
things that are wrong in any business

environment, and people just become
disheartened with this when, your example,

if you had taken Stephanie and held her
up to the rest of the group and be like,

"Stephanie, you're absolutely killing it.

Like here are all the KPIs
that are 40% lower than all

the other peers in your team.

Can you tell us what you're doing
differently to be able to do this?"

And then she does like a little dog
and pony show, does some cross-training

with the team and every- all of a
sudden everyone is benefiting from what

she figured out and, and you maximize
that across the team rather than just

like, "You guys are doing terrible.

I wish you would do better.

This needs to improve."
Like, I'm trying to

be constructive and tell you that
you should do better, but you're

actually demoralizing the team
rather than using the positivity of

a single person and promoting that
in some po- positive fashion, right?

Jeff Toister: Oh,
that's, that's very true.

Uh, or, uh, y- you just kinda
let it, let chaos slide, right?

Todd Kane: Yeah.

Jeff Toister: And then that's when we say,

Todd Kane: I wish they were better, yeah

Jeff Toister: now we need AI because
AI will be way more efficient."

Todd Kane: That's right.

Yeah, yeah.

It solves all problems.

Jeff Toister: That's right

Todd Kane: the oth-the other one
I, I wanted to get your thoughts on

is, uh, the reachability paradox.

I find this really fascinating.

I wanna hear a little more on this of
like, um, because, uh, the example of

where I see this show up a lot is people
will ask me like, "Hey, you know, what's

the best portal that I can use, uh, so
that my clients can enter their tickets

and manage their tickets and see their
tickets?" Uh, and I'm like, Well, you

know, the portal and portals in the
industry are okay at best, like none of

them are great, but, you know, here are
your options." Before you spend a lot of

time on this, is this something that a
client is specifically asking you for, or

is this something that you want them to
use and you think it would be really cool?

Because these two things
are very, very different.

Most portals go largely unused.

A little bit in enterprise, I think
there's some sort of general uptake

of this and a, and an assumption
that these things will be utilized.

In SMB, if you tell people to go
check a portal, I, I think you're,

like you're, you're, you're kind
of pissing up the wrong tree.

Like the-- people are just not gonna be
interested in this in general, unless

you have maybe an IT director who
wants to actually monitor these things.

Um, you, you know, we're talking a bit
about, you know, using some portals

or maybe an AI entry point for these
things, but, you know, if we're directing

them away from people, what is, what
is the impact on whether or not people

actually feel that there is options to
use this and, you know, you're triaging

on, on bots versus humans, um, and you
have an idea, uh, the, uh, of the more

that you amplify these things, the
less that they have an adoption, which

seems sort of counterintuitive here.

So let you take it on that.

Jeff Toister: let's start with one thing.

Is, in this example, is the portal,
if people were using it, does

it save their time or make their
experience better in any way?

And I'm, I'm asking you
just generally, does it

Todd Kane: Uh, I think it

would give them better insight, right?

So like one of the things I really
advocated for, we had some portals back in

the day at the MSP that I ran, uh, and one
of the things that I really liked about

it is it showed like this race car track,
and it would s- show like this is when

the cr- the ticket was created, here are
all the time entry points of when it was

updated, and here is when it was closed
or when it's predicted to be closed.

And I think that that, that timeframe
of visualizing the life cycle of

a ticket can be very beneficial.

Problem is I think you're telling
people like to self-serve, like,

"Go find the information yourself,"
and they're like, "Really?

Like you're giving me homework?

Can't you just tell me
like what's going on here?"

And this is why like people
will do like the, the ticket

m- the, the ticket messenger.

Like they'll call the service desk
and be like, "Hey, I just opened a

ticket. Do you guys have that ticket?"

Jeff Toister: Right

Todd Kane: "Yep, we got the ticket," Right

'Cause they just don't trust that
anything's gonna happen unless

they contact a person and are
able to chase someone down on it.

Jeff Toister: Well, I think
you're, you're bringing up, I

think the core issues, right?

There's trust, and then we're also asking
people to change how they do things.

And it's counterintuitive, but if, if
you have a system or a portal or anything

else that actually can be beneficial
to the people who use it, even if it's

in certain situations, one of the ways
to improve the adoption of self-service

is by having a human readily available.

And so at first it's a little more
labor-intensive, so maybe I'm talking

to that person who checks on the ticket
and says, "Hey, um, I wanna show you

something because if y- you have another
ticket and you wanna check on it, it's

actually a very easy, faster way for you
to do it. Do you have a minute?" Right?

And so that takes a little extra time,
but if the person says, "Oh, wow, this

is super easy," and then you send them
a, you know, a message with a link

to it, a couple of steps, that person
is more likely to try that again.

There's also been multiple studies, I've
been part of some of myself, that just

knowing a human is available if you need
it makes someone more likely to try it on

their own and be happier with the outcome.

And I think we have to design
that, you know, for humans.

Often these, these things are designed
saying, "It's gonna save us more time

if you don't contact us and you just go
to the portal." But we're not thinking

about the person at the other end of that.

A- and I'll give you two
ex- two quick examples.

One, a CIO was getting a lot of flack
because his, his service desk was

way behind on things like, you know,
they're setting up a new office in

Europe and, you know, they're not
ready in time, that kind of thing.

And he goes to investigate a- and the
issue is that executives in the home

office would bypass the ticketing system
completely and just walk in and find an

employee and say, "I need help with this."
And what are you, what are you gonna do?

You're gonna tell the CFO no?

I'm trying to set up this office in,
the network for this office in Europe.

Get in line, buddy.

Todd Kane: shoulder tapping is difficult.

It's even more difficult
when it's an executive

Jeff Toister: Exactly.

So that's what was causing all the
problems, and it, it's, it's really

not the employee's fault in this case.

So… But the CIO just wanted to
give them some training or something.

I mean, that, that process was broken.

Another company, same issue.

They said, "We're just gonna
recognize this is what happens."

So each member of the executive
team was assigned a person.

That's your person.

You need help, you go to them because,
hey, that investor presentation's

coming up, and you probably do need
someone to just make sure it happens.

And then we're, we're just going to
acknowledge that, and by creating

those, uh, relationships, having a
go-to person, it allowed the, the work

for the entire team to be smoothed
out so that the shoulder tapping

was a lot more predictable now.

I just have to check on maybe one
person or two executives, and that's

it, and the rest of it I'm going to
be working the queue like normal.

And, and so just, it was acknowledging
this is how our humans are going to work,

so let's design a system that's efficient
and works the way our humans work.

It, it just ended up being
a way better solution

Todd Kane: Yeah, I think, uh, that hits
on sort of the point that I rail against

this is, uh, it has to be a full court
press of behavioral change and education

if you're gonna get people to adopt this
stuff and change the way that they work.

I like your suggestion of, um, you
know, you take the ticket and then

at the end you're like, "Hey, by the
way, there's this cool tool that you

could use in the future," and just
re-educate them on how to use that.

You give them support.

It's not just sorta, "We sent an
email last month about this portal,

And no one's using it." Like,
no, of course they're not, right?

Because, you know, you can
assume 70% of the people didn't

even read the email, right?

So, like, that's, not the extent of
communication, and I think making it

more human of like, uh, uh, as you
said, like, "This… Here's how you

can do this in the future because
you may have multiple tickets open.

You don't have to wait to get ahold of us.

You can just open this link, and then
you can check on the tickets yourself.

Like, isn't that cool," right?
And they're like, "Oh, yeah, cool.

Okay, I'll use that in the future,"
versus just quiet being frustrated by

the fact that people are not using the
tool that you deployed but didn't really

educate them on how to use, right?

Jeff Toister: And that's
that's exactly it.

The framing is this is how it helps you.

And, and if people believe that,
then they're more likely to try it.

A- and if they say-- and if you say, "Hey,
I think this will save you some time.

You'll get an answer quicker." And
you can even add, "But if that doesn't

work for you, I'm always here to help.

It might just be a little delay 'cause
I'm working on some other issues."

Now you've given people the option,
and you've opened the door for them to

say, "Well, maybe I'll try it 'cause I
know I have to wait to get that person.

Oh, hey, it's a good
experience, all right.

This works

Todd Kane: 100%.

right.

So Jeff, uh, this has been great.

Really appreciate all your input.

Uh, maybe we'll close out with,
uh, sort of an actionable here.

If an MSP owner is like, "Cool, I think
I'm gonna start deploying some AI on

my service desk next quarter with XYZ
tool," uh, what is one thing that they

must protect in doing that, in your mind?

Jeff Toister: Well,
that's a great question.

I think you have to
protect the experience.

And what I mean by that is
the goal is not to deploy AI.

The goal is to create a better
experience for your client to do

it efficiently and profitably.

And if AI is a tool that can help
you do that really, really well

and your client is happy with
that, then that's a good move.

And if it's not the right
tool, then don't do it.

So it's right tool for right job, but
think about the experience first and

finding the right tool and the right way
to deliver that experience consistently.

That should be the correct order of
thinking rather than, "Well, let's use AI.

Oh, let's see, how do we make it work?"

Todd Kane: Yeah, I like that, um, thinking
about it more holistically in the client

experience and how that, that tool is
a component of the overall experience

rather than the tool really being focused
on an internal deployment in order

to make things more efficient, right?

I think you kinda mentioned that earlier
as like thinking is not incorrect, it

just, it, it, it, uh, it sort of loses the
forest through the trees essentially of

like what is the ultimate thing that we're
trying to do here is to create a h- high

value experience for our partners so that
they feel value in us as their IT partner

and will continue to use us in the future.

Not like, "Well, we shaved, uh, you
know, 14 calls off of our, our queue

today, and therefore that's a win,"
despite the fact that we don't know

if the clients liked it or not, right?

Jeff Toister: Exactly.

I, I think the real easy way to think
of like customer or client-facing AI

is it, it should meet three conditions.

One is confidence, that your end user is
confident using it, so they probably will.

If they're not confident, they won't.

It's routine, so in other
words, we have enough process

procedure to be able to train AI.

This is how to do it correctly.

This is not how to do it correctly.

And it's predictable.

Uh, we have to be able to get
the same result every single time

we run the same procedure, and
if we're not getting predictable

results, it's not working for us.

So the, it, it's confident,
routine, and predictable.

That spells CRAP.

We're just gonna automate the crap
that our clients don't want to have

to wait for a human for, and frankly,
our humans don't wanna deal with.

And if you can focus on that
first, I think you'll be okay

Todd Kane: That's awesome.

All right, Jeff, this has been great.

Uh, I'll link to your LinkedIn profile,
your book, and, uh, any other fantastic

links to check out your content.

Anything else that you would call
people to or, or ask them to, to check

out if they wanna see more from you?

Jeff Toister: I think the
book is a good place to go.

Uh, it's, it's human service,
so humanservicebook.com.

And also, I, I know a lot of MSPs,
a lot of help desks, et cetera, uh,

subscribe to my, uh, newsletter.

It's my customer service tip of
the week, and you can find that

at toystersolutions.com/tips.

One email once per week, one
very specific tip you can use

to help your clients this week

Todd Kane: Links in
the show notes as well.

Jeff, this has been great.

Thanks for your time

Jeff Toister: Thank you.

I really appreciate it.

It's been fun