Working Smarter: Presented by Calabrio

Dave Hoekstra: Welcome to Working
Smarter, presented by Calabrio, where

we discuss context centered industry
trends and best practices, as well

as sharing success stories and pain
points with some of the most innovative

professionals in the industry.

We're glad you're here joining us to learn
and grow together in order to provide

world class customer service to each
and every one of our collective clients.

My name is Dave Hoekstra, Product
Evangelist here at Calabrio,

and I have two guests today.

I'm joined from AWS.

I'm very excited to have Mike Gillespie.

Now Mike is a Principal Solutions
Architect at AWS joining us and

with us again for the second time
on our podcast is Shalima Bala.

She is the Global Lead for
Partner Development at AWS.

Now I assume pretty much
the whole world is of AWS.

pretty aware of what
AWS is and what we do.

And so our goal today, what we really
wanted to spend some time talking about

is the super exciting world of AI.

Has it, have we, have you heard of AI?

Do we even know what
this is at this point?

Absolutely.

We all have, right?

We've all talked about it quite
a bit, and it's starting to

really influence what we do.

And AI in the context center has.

Let's call it let's say it started to
mature a little bit we're still very

early in the stage, but we're starting
to see a little bit of collection

of that going and You know what?

We want to focus a little bit on the
reporting side today but I really want

to start with and either one of you can
answer the question here, but You know

when we talk about AI in the context
center how has it started to work its way

into the functions of the context center?

And, what can context center users expect
to see out of AI today and maybe tomorrow?

Mike Gillespie: Yeah I'll take this one.

So first I'd like to jump into a
little bit of why, like why is AI

all of a sudden such a big hot topic?

Why can it now do those things Do
before, and it's really a confluence of

a number of different industry trends.

And first followed the cost of
compute and storage has gone way down.

So to do the calculations needed to do
the AI used to be prohibitively expensive.

And now the cost has gotten
into it being very economical.

And another aspect is there's a lot less
friction and how to track that data.

So instead of having to take your
call logs and write how long the

call was get the sentiment of each
and every call get feedback from the

customer, those are all automated now.

So that data just comes
along with the ride.

So combining those two things,
you have a much richer set of data

that's easily collected along with a
lower cost for compute and storage.

Now we can actually do things that
are predictive in nature that help

the efficiency of the contact center.

So to do those things are,
like, how do we reduce errors?

How do we shorten our contacts?

When you, as an end customer, when you
call the contact center, your goal isn't

to have as long a call as possible.

It's really, how do I get what
I need as quickly as possible?

So to be able to collect data and
bring that data using AI right in

front of the agent, Or through a chat
bot to get to that resolution faster.

So those are all the factors
that are driving like why

the AI is such a big topic.

Now And what that end result
does is now those data.

The business is now more data driven.

You have more metrics to track against,
and that's where the reporting comes in.

So now we have a richer set of data.

We have a I integrated in to
help collect that data and

make predictions on that data.

Now we can make reporting that gives
very keen and clear metrics on how

we can improve the call center.

It was saying that if you can't
track it, you can't improve it.

Now we can track it and now we can
improve it and develop technologies

around making improving the experience
for both the users or the end

customers, the contact center agents
and the leaders of the contact center.

Dave Hoekstra: Let me unpack that
a little bit, Mike, because are

you telling me that I do not have
to write SQL queries anymore?

Is that what I'm hearing?

Mike Gillespie: Yeah, from a generative
AI standpoint, that's one of the

most powerful components is it makes
that data much more accessible.

So instead of having to know the
data structure or the SQL query

language or how to use the reporting
system, you can just ask questions.

What was the average call time over
the last seven days and it will in the

back end knows the data knows how to
query that data will generate the query

and report the answer to you and you
can do things like make a pie chart

that shows the call disposition over
the last five days, things like that.

So it just makes it much more
accessible and reduces that ramp up

time that learning curve to be to
make that data accessible to you.

Dave Hoekstra: So what
I'm hearing is that.

This is available now, because I know
a lot of times when we talk about

AI we really blur the line between
today and five years from now, right?

A lot of times it gets really murky when
we talk about those kinds of things.

These are things that we can do
today where I could literally take a

data set and just type in a question
and have it give me an answer.

Mike Gillespie: Absolutely.

Yeah, you're absolutely right.

A lot of times when you read
these articles in the newspaper,

it's very spectacular.

Speculative in nature, that's Oh, we
might be able to do this in the future.

This is a case where
you can do that today.

So for example, an AWS RBI solution,
a quick site allows you to just

ask questions about your data.

So you can ask those questions
that I talked about earlier, or

you can build charts and graphs
just using natural language.

To do and that's called Amazon
queue within QuickSight.

Dave Hoekstra: That's pretty awesome.

I, when I started working in contact
centers back in 1999 or let's,

I got into management in 1999.

Let's put it that way.

I literally had to build a
report in Microsoft access like

an entire reporting database.

And it took me weeks to even
get little things going and.

What's really funny is how revolutionary
it was to everybody that I could, you

could click a button and get the last
week's worth of call volume reports.

So what I'm hearing is
we've come a long way.

Have you come across any new
connections in data that potentially

that would get people really excited?

I know it's pretty basic to
say, how many calls did I get?

But have we seen new ways
of data that can connect?

Mike Gillespie: And that's where,
this technology is really impressive

when it's so conversational.

So you'll talk about, show me
the data for the last seven

days, and that's just a report.

But really what it comes down to is
when you can start telling it to ask

questions of the data, like things that
you may not have seen, you can ask it,

how is this data correlated to another
piece of data or, Tell me any anomalies

that are occurring in this data.

What are outliers that we're seeing?

And it knows the context
of the conversation.

So instead of having to run the
report again and put it into Excel,

et cetera, you can run that, those
types of analyses right in line with

your conversation with the data.

That's one area where.

It becomes much more accessible to
non technical analysts than it would

be someone that is an expert in,
Microsoft Access or whatever tool.

Dave Hoekstra: Wait, people
still export data to Excel?

That still happens?

Every day.

Every day.

Every day.

I still maintain, it is my own
personal feeling, that Microsoft

Excel is the greatest soft, piece of
software that has ever been invented.

And we are.

We are hoping to get away from it, but
everybody's infrastructure is so built

on that, that it's an adventure, right?

Excel

Mike Gillespie: is probably the most used
and abused piece of software ever written.

Dave Hoekstra: Oh, that's
such a true statement.

And we all have a special place in our
heart for it, but we all have, there's a

special place somewhere else for it too.

Sometimes.

That, I, it excites me especially
to think about how daunting a new

reporting data set can be for people
and not just for people who know data

already, even for people that are new.

Do you think AI, this kind of AI process
is really going to help us onboard

new people that much more quickly?

Mike Gillespie: Absolutely.

That's one of the key.

Benefits of this conversational
technology is it puts the train the

learning curve down significantly,
so you don't have to be an expert.

You don't have to have a Ph.

D.

in data science to be able
to answer these questions.

You can just be very conversational and be
able to drill into that data just based on

the things that you see in front of you.

And the.

The trends that it identifies and
that you identify in the data.

Dave Hoekstra: That's awesome.

And I'm looking very much forward to
seeing that continue to expand in, not

just in the contact center, but in, in
all the software that we use the number

of times that data is presented to us,
but it's still just so up and throw

up for lack of a better word, right?

It's still just dumped in front of us.

And then, I'm a little worried because
I've made my entire career out of

being able to interpret that data.

So I'm wondering maybe, and you're
gonna, hopefully it's going to

turn us into better analysts as
opposed to no analysts whatsoever.

And that's a little excited.

Now you mentioned.

That, and this question could be for
both of you however you'd to tackle it.

You mentioned that AI used to
be prohibitively expensive.

And it's gotten much, much more
accessible and cheaper, but I've

still noticed it ain't free, right?

We're still using it as a methodology to
either recoup some costs or find a way

to get our hand when we want to get our
hands on that next level of technology.

It's usually not just turned on.

We usually have to go out and find it.

So are there things that.

These contact centers can do or
myself that can help the kind of

these next level functions pay
for themselves in a lot of ways.

Shalima Bala: Let me take that.

Let's face it, any technology,
not just gen AI or AI in general,

the implementation of those
technologies can be really expensive.

However, when implemented strategically,
the cost savings and efficiencies that you

gain from these technologies, especially
artificial intelligence can outweigh

its initial investment, essentially
making technology pay for itself.

For example, AI can automate tasks.

And identify inefficiencies pretty
quickly and make better decisions

leading to significant cost reductions.

Let's take an example from contact center.

Jenny, I can automatically craft a
call summary at the conclusion of the

interaction, dramatically reducing
the after call work for an agent.

So by removing the time agents spend
manually and summarizing a call,

organizations can save millions of
dollars and if done right, it can

also help with the revenue generation.

So by enhancing customer experiencing
personalizing marketing campaigns

and developing new products.

A.

I can generate additional
revenue streams for a company.

Let's take a basic example.

Virtual assistance, right?

So if we implement it correctly, it can
reduce the inbound call volume for a

contact center and also enhance the C set.

So it can automate the
conversations with customers

across digital and voice channels.

E.

I.

At its core can analyze large
amount of data sets to provide

insights and predictions.

So if we power our agents with the
AI, it can listen to the conversation

real time and help agent to offer new
products to the customers and enhance.

The the value that you get out
of a particular interaction.

In turn elevating the performance
and also delivering significant ROI.

Now you must be thinking that all of
this sounds great, but how quickly can.

Customers expect these returns, right?

And the timelines can be very
depending upon the complexity of the

implementation and specific use cases.

But what I have seen that many businesses
start seeing these positive otherwise.

Probably in six to 12 months.

Okay.

Done correctly.

And so are

Dave Hoekstra: you talking about with
the, the voice assistants or the agent

assist pieces, is that specific to just
that, or is it any kind of any timeline

where AI has been brought in to assist

Shalima Bala: on an average?

Dave Hoekstra: Okay.

Shalima Bala: On an average.

Yeah.

Because mining the data, implementing
the technology and learning that

especially Jenny, I learning that
data, it takes six to 12 months.

And also when you're implementing gen AI
particularly, it's not about just taking

the technology and just replace it.

What do you have currently, right?

You have to clearly define what
objectives you're trying to solve.

What is the outcome you're
trying to achieve and also most

importantly, quality of data.

And you have to have the right
talent in your company to monitor

and optimize it continuously.

So that is really important as well.

Dave Hoekstra: Yeah.

The, what the challenge that I've seen
to phrase it in a different way is we're

we've got hundreds and hundreds of years.

Of teaching human beings
how to do this process.

We've learned over many generations
to say, okay, if you want to teach

someone how to do something, you sit
them next to each other and you haven't

one person watched the other person
and learn, and eventually they get

good enough to go off on their own.

And AI doesn't work at all like that.

AI is much more you have
that person sit down and map

everything out before you even.

You're even allowed to have it look
and that, that's the difference is we

have to learn as a race, as a species,
how to teach AI to do things for us.

And that is that's an, a bit
of an adventure because we, our

brains aren't wired that way.

That's what I've noticed for a lot of
people is that I don't think like that.

Learning how to think like that, Calabrio,
we just launched a new process that

uses AI to to answer quality questions.

Did the agent did the agent
use the greeting correctly?

Something like that.

And as we're trying to teach our
customers how this works, they're

so used to just having a human
being listen and go, good job.

Without ever having to
quantify what that means.

They just, they're just
good job or bad job.

And now we're trying to
teach people how to quantify.

Have you ever seen that video where
it's somebody says, teach me how to

make a peanut butter and jelly sandwich.

And then they say, okay, you take
two pieces of bread and they're

like, I don't have any bread.

And they're like Oh, okay.

You have to go to the store.

But how do I go to the store?

Okay.

You get in a car.

What's a car.

And it's trying to teach people.

This is how computers think.

You can't just say to the computer, take
two slices of bread and stick it together.

You have to really, you have
to define every minute step.

And that's what it really
feels like with AI.

So I think bringing this to light for
a lot of people is really important.

You cannot just turn on AI
and have it magically work.

There's a fair amount of prep
work that goes along with it.

We're seeing that across everything.

Mike Gillespie: Yeah.

One other aspect of it, which is
unique is in traditional applications.

If you have the same inputs,
you'll have the same output.

It's very predictable where
AI is a little different.

There's randomness.

So you can ask the same question of an AI.

And exact same question and
get two different responses.

And they could be very different.

So building your systems to be resilient
to that variability in those responses.

Now, the responses may be more
accurate than their comparable human.

There's a benefit there, but you do
have to build that recognition that

you have to check your work and you
have to check the eyes work as well.

Dave Hoekstra: Nobody wants to do that.

Mike.

Nobody wants to check their work.

No you're right.

And it's it's funny because we were
starting to see that with some of the

large language models that are coming
out where, you start to see things pop up

where someone says like, How many ounces
are in a gallon and you might, honestly,

sometimes you might get different answers.

And then we're starting to see things
being taken as fact where it's really

because bad source data, because the large
language models are mining incorrect data.

But that's the thing.

AI doesn't know what's wrong.

Mike Gillespie: Yeah.

Dave Hoekstra: And so I don't think we
should be that worried about Skynet.

At least in the near future, right?

That's the takeaway from this.

Last question, either one
of you can handle this.

Let's say that I am a contact center,
or even just an organization, doesn't

have to have a contact center.

And I'm looking into these tools
and I know that more than likely in

the next 12 to 18 months, I'm going
to purchase some sort of solution.

What should I do today?

To make sure that I'm ready for
when the implementation calendar

starts on that type of solution.

So it could be voice assist, it
could be interaction summaries, it

could be whatever the case may be.

What should I do today?

Mike Gillespie: I'll take a first
pass and then Sheldon can jump in.

But I think having a good understanding
of what those metrics are that really

drive success for your business.

Because what AI will do is amplify
what you're trying to, those

metrics you're trying to manage.

And give you tools to help improve.

Whatever dimensions make
the most sense for you.

Without having a good understanding
of how that, that business works,

you're not going to be able to use
that amplification as these new

technologies, roll into the mainstream.

The other part is, even outside of,
these packages or these solutions

I would say, Use the publicly
available AI, chatbots get used to

how they, how to interact with them.

There isn't a skill, we call
it prompt engineering, but it's

really how to talk to a computer.

Yes, it's natural language and you're
using English or whatever language.

But there is a certain style
that you communicate with the

computer and being fluent in that.

And you don't have to
be a computer scientist.

You don't have to be an expert, but
just getting comfortable with having

those conversations will go a long
way to helping you understand how

they react, how they think, and then
how to utilize them in your business.

That,

Dave Hoekstra: that so reminds me
of when I was trying to teach my

kids how to search the internet
and they would sit down and type in

a long string of, or they'd ask a
question what's the best video game?

And it's and it hurt my head
because, there's just Oh, that's

not how these things think.

And you have to be more specific,
but it, and then you look back and

you realize, how would they have
known that this is the way, right?

It's my job to make sure that
they know how to do this.

So teaching them how to correctly
query and correctly prepare for these

kind of things is such a huge thing.

I think that's a great piece of advice.

Just because something can do it
better, faster, stronger, doesn't

necessarily mean that the old way
is going to completely go away.

Because we're, there's the
world finds that niche, right?

It finds that area.

And I agree with you 100%.

It will actually in the long run, create
more jobs because we're going to have

time to create new and amazing things
and new and amazing technologies that

supplement what's available out there.

So I agree with you wholeheartedly.

And I think anybody who's scared
of AI I don't think you should be.

The benefits of AI are going
to far outweigh any kind

of negative effects of it.

It might affect certain industries
or certain areas a little bit

more than others, but Mike,
anything you want to contribute?

Mike Gillespie: No, I think
that pretty much hit it.

The awesome, the thing
I'll add is that, the.

In the long run, the companies
that are in organizations that are

most successful utilizing these
technologies will be the ones that

are most successful in the long run.

So embrace it learn it, play
with it immerse yourself in it,

and it will be time well spent.

Dave Hoekstra: But I can do this
all in an on prem environment Mike?

Mike Gillespie: No comment.

Dave Hoekstra: No comment.

Oh, this has been fantastic.

You guys.

I learned so much.

It's always great to talk to
some people who really understand

this and get a chance to immerse
themselves in it day to day.

I'm not necessarily one of those people.

So thank you guys so much for
spending some time with us and

really educating on what's going on.

It's been great to talk to you both.

For the rest of the audience, as always.

Thank you guys so much for
spending time with us here on the

collaborative working smarter podcast.

It's great to spend some time
with you and looking forward

to having many more episodes.

So thank you guys so much.

Charlie, my Mike really appreciate you
jumping in and giving us your thoughts

here on AI reporting context center
and all that fun stuff that we have

to deal with on a day to day basis.

So thank you guys so much.

And to our listeners, have
a great rest of your day.

Have a great grit.

Rest your weekend.

We'll talk to you again on the
working smarter podcast from Calabrio.

Thanks everybody.