SX Live Events
Joe Andrews 59 min

The Impact of AI on the Post-Sales Customer Experience


At SX Live in Austin last week, the following support and success leaders gathered to discuss the impact of AI on the post-sales customer experience. - Tina Branson Dobie, CCO, Calendly - Jenna Koontz, VP Support, Certinia - Eran Ashkenazi, CCO, SentinelOne - Tim Barnes, VP Financial Services Support, Oracle - [Moderator] Joe Andrews, CMO, SupportLogic Throughout the lively discussion, the panelists shared their insights and experiences with implementing AI in their own organizations. They emphasized the importance of taking a pragmatic, customer-centric approach to AI, and using it to enhance the support experience rather than replace humans.



0:00

I first want to welcome everyone to our SX Live event.

0:03

A few of you have been with us before,

0:06

and we're so happy you could join us again.

0:09

My name is Joe Andrews,

0:10

and I'm the Chief Marketing Officer,

0:13

Support Logic, and the Host and Moderator.

0:16

I want to provide some brief context.

0:18

Who's familiar with the Gartner hype cycle?

0:22

Presumably most of you.

0:24

They basically publish a curve

0:26

of where emerging technologies are in terms of hype.

0:31

This is the peak of hype.

0:33

Generative AI is at the very top of that hype cycle.

0:36

Now what happens, as we know,

0:38

from other technology innovations

0:41

is you come through this trough of disillusionment

0:44

where realistic expectations are brought in,

0:49

and then it sort of normalizes

0:51

into a plateau of adoption use.

0:54

So it's not hyperbole to say that we are hyped right now

0:58

with Generative AI.

1:00

In fact, Jenna, who's on our panel,

1:02

she was here at this event a year ago,

1:06

and we touched on Generative AI very briefly

1:10

because it was a brand new topic.

1:12

Chat GPT had just launched.

1:14

It was brand new.

1:15

We were sort of testing it for our personal use,

1:17

but there wasn't a lot of focus for real

1:20

in the world that we all reside.

1:23

That said, right, there is a tremendous opportunity.

1:28

This is data that Boston Consulting Group, BCG,

1:33

shared in a previous event.

1:35

We were actually up in Seattle last week,

1:38

and I thought this was incredibly compelling.

1:40

14 billion hours per year

1:43

that customers spend contacting customer service.

1:46

Now that huge number is really hard to visualize.

1:49

So I did a quick calculation.

1:50

It's like taking every person in the United States.

1:54

If they sat down for a full 40 plus hour work week

1:58

and they were on the phone with customer service,

2:01

that's how much interaction time there would be,

2:04

except it's on a global basis.

2:07

So there's a ton of opportunity

2:10

to improve those interactions and those experiences

2:14

that you all, we all, have with our customers

2:18

when they're engaged with you as vendors.

2:23

Now, the challenge, we talked about the hype cycle,

2:28

is that the perception by C staff,

2:33

some of us here are in the C suite,

2:35

some of us, many of us report to officers of our companies.

2:40

And just six months ago, again, this is BCG data,

2:44

they reported that more than 50% of the C suite

2:49

did not fully understand the implications of JNI,

2:52

no surprise, there's an investigation to be done,

2:55

and they were discouraging its use.

2:57

Now we know about, many years ago was

3:01

in the cloud infrastructure space,

3:03

and the cloud hype cycle was a protracted event,

3:08

and it's still going on.

3:09

There's still mainstream cloud adoption challenges going on.

3:13

We anticipate this is going to be a long transition,

3:17

but we were very pleasantly surprised,

3:19

this is data that we just saw from BCG,

3:22

where they said, in six months,

3:26

the C suite has shifted.

3:28

Two in three, now view JNI as the most disruptive technology

3:33

in the next five years.

3:36

Now again, that sounds hyperbolic,

3:38

but one third of these same executives said

3:42

they are increasing their investments in AI as a result.

3:45

And this survey, you can go see it,

3:47

it's their digital acceleration technology index,

3:50

and it's a survey I think of 2000 executive,

3:55

senior executives at, the largest enterprise companies.

3:59

So there's legitimacy here,

4:01

it's not hyperbolic to say that there's hype with JNI,

4:06

but it's also not hyperbolic to say

4:09

that there are real use cases and value and adoption

4:13

that are taking place today.

4:15

And that's what we're going to drill into.

4:17

From our perspective at SupportLogic,

4:19

AI impacts every function in the company.

4:23

We really started with a focus here

4:25

for support executives, managers, team leads,

4:29

where you extract signals that might be sitting

4:33

in support interactions, unstructured data, right?

4:36

A lot of head nods here.

4:39

Metadata, pull down menu, severity level type of issue.

4:43

Who is the agent assigned?

4:45

Well, if you could actually harness

4:48

the qualitative information, the sentiment

4:51

coming from a customer and be able to push that upstream

4:56

to customer success, sales, product and engineering,

5:00

be more proactive about fixing the issues

5:03

inherent in the products,

5:06

you can drive financial metrics, KPIs for the companies,

5:10

improved retention, lower operating costs,

5:13

who doesn't love that?

5:14

So this is becoming very real,

5:17

and the benefits of AI are cutting across

5:20

every function in the company.

5:22

And we believe at SupportLogic, as do many other companies,

5:26

now I work in marketing for an AI company.

5:28

So I'm prone to say this, right?

5:31

But we believe just like 10 years ago,

5:33

we would say every company will be a cloud company.

5:37

Today, that's not a surprise.

5:40

Back then, people shook their heads a little.

5:43

Now we believe every company will be an AI company.

5:46

Every company will incorporate AI into what they build,

5:51

what they deliver, how they serve their customers,

5:54

which is why all of us are here today.

5:57

So just a couple more slides,

5:59

and then we'll bring up our panel.

6:01

Where do you start is the question?

6:04

So there was other data from BCG,

6:08

we just actually published a blog on this,

6:10

where approximately 40% were sort of just getting started

6:15

or not getting started with AI,

6:17

10% are fast and furiously running towards it scale.

6:22

And then the big mass in the middle, like 50%,

6:27

we're sort of tinkering with it, proof of concept.

6:31

Where are folks like in the room,

6:33

who is actively driving AI solutions today?

6:37

Wow, that's two thirds of the room roughly.

6:41

I mean, Jenna, year ago, right?

6:44

Very different response, scratching the surface.

6:47

Who is sort of tinkering with it?

6:48

Testing it out, okay, a few.

6:51

Who, it's okay to admit it hasn't started yet,

6:54

but has good intentions, okay?

6:56

So there's a mix in the room.

7:00

We're gonna have a mix of perspectives on our panel as well.

7:03

But what BCG said is that there's kind of four areas

7:07

where you can start.

7:09

The area on the right, number four,

7:10

is within the actual support response itself.

7:14

It's more tactical, but it's more immediately

7:17

realizable in terms of value,

7:19

enabling support teams, agents to improve

7:23

the customer experience in the moment,

7:25

do right by the customer, resolve issues,

7:28

better, faster, cheaper, right?

7:30

And then you go over to the left,

7:32

where you have self-help, enabling customers

7:35

to get the resolution that they seek by themselves.

7:38

Self-service, a lot of companies for many years in support,

7:42

because support was very cost-centered focused,

7:44

were really focused on self-service to deflect cases.

7:49

But we know if many of us are in complex technology companies,

7:54

the element of the human will never go away.

7:58

You can push some of the easier to resolve cases

8:02

down to self-service, but you wanna still focus

8:04

on delivering the best service.

8:07

So this is an area of investment.

8:09

Self-healing, being able to preemptively improve

8:14

and resolve the issue for customers

8:15

before they even realize it's an issue,

8:18

wouldn't that be amazing?

8:21

And with very little effort.

8:22

And then finally, the preemptive,

8:24

being able to fix product issues proactively

8:28

before anyone notices them.

8:30

I think some automobiles and maybe some other high priced

8:34

products do that to some degree.

8:37

But you can imagine this is the hardest area, right?

8:40

So most companies are starting here

8:43

with the support response space,

8:44

being able to improve the life in the moment

8:48

for the customer and improve operational efficiency

8:53

while you're at it.

8:54

So I'm not gonna present to this slide.

8:58

It's an I chart, but there are,

9:02

again, this is a BCG slide.

9:04

They have a lot of amazing content on this space.

9:07

And if you look across these four areas,

9:11

there are, I think, 36 or 38 different use cases

9:15

where they are starting to see value

9:17

where companies are actually investing in AI

9:21

for the post-sale customer experience.

9:24

And what do we see?

9:26

This is a support logic slide.

9:28

And we see that, again, this is also an I chart.

9:32

This is on our website.

9:34

But we see three kind of core areas.

9:37

Agent assist or agent productivity.

9:41

We see quality monitoring and coaching,

9:43

being able to auto-QA every single case

9:48

in customer interaction and provide feedback

9:50

and coaching against a program rubric.

9:53

Wouldn't that be fantastic?

9:55

And then core support operations efficiency.

9:58

Being able to understand customer sentiment,

10:02

reduce escalations, intelligently assign the case

10:05

to the right support engineer or agent,

10:08

be able to better more efficiently route cases,

10:12

alert people in the organization,

10:15

bring them in to swarm to resolve customer issues.

10:18

This is kind of a big area of investment.

10:20

So these are kind of the things

10:22

that we're going to start to dive into as a panel.

10:26

And really would love to make this conversational

10:30

and continue this throughout the evening.

10:32

And these are some of the support logic customers

10:35

that are seeing incredible results using AI.

10:39

In the customer experience camp,

10:41

you have significant reduction in escalations,

10:44

which has a very real operational cost,

10:47

but also improves the customer experience.

10:49

You have retention and the CFOs love these metrics.

10:53

Being able to show we were able to preserve these renewals.

10:57

We were able to ensure that customers don't churn.

11:01

And then on the operational efficiency,

11:03

being able to reduce the time to resolution,

11:06

being able to shrink the number of cases,

11:09

get support manager productivity time back in the day

11:13

so that companies can redeploy those valuable resources

11:17

into things that are more proactive

11:19

to improve the customer experience.

11:22

And I see a lot of heads nodding.

11:23

And this is sort of the panacea.

11:26

And this is where the AI is really delivering value.

11:30

And we would love to have more conversations with you all.

11:34

And so I'm going to call everyone up.

11:37

We'll start at the bottom here with Tim.

11:39

You want to take your place Tim Barnes,

11:42

VP Financial Services Support at Oracle.

11:45

Next, we'll call up Jenna Coons,

11:46

who's the VP of Global Technical Support at Sertinia,

11:50

Iran Ashkenazi, the Chief Customer Officer at Sentinel-1.

11:53

And then Tina Dobie,

11:56

who's the Chief Customer Officer at Calendly.

11:59

So we are going to dive into these topics

12:03

and engage the audience, bring your best questions.

12:07

But let's start with a quick introduction

12:10

and an icebreaker.

12:11

So Tina, would love, welcome.

12:12

Thank you for joining.

12:13

I'd love to hear a little bit about your role at Calendly

12:17

and what you're focused on,

12:18

and maybe a fun fact about yourself.

12:20

- Okay.

12:21

Hi everyone.

12:22

Thank you for having me.

12:24

Appreciate it.

12:25

Y'all should get started.

12:26

It's like a family meal.

12:27

- Yeah.

12:28

(audience laughs)

12:29

- So I'm Tina Dobie.

12:31

I am the Chief Customer Officer at Calendly.

12:34

I see a familiar face over there.

12:35

Hello.

12:37

And I've been with the company for about two and a half years.

12:40

I was previously at a local company called WP Engine.

12:45

Chief Customer Officer there for seven plus years.

12:48

So lots of experience working with customers and teams.

12:51

I'm responsible for all of the post sales teams,

12:56

everything from 24/7 customer support

12:58

to customer success for our large enterprise customers.

13:01

We have an implementation on-boarding team.

13:04

I'm an operations team that manages all of our tech stack

13:07

and analytics as well as customer education and training

13:11

and enablement that makes everything happen at scale.

13:14

So that's a little bit about me.

13:17

Has everyone heard of Calendly?

13:18

Does everyone know what Calendly is?

13:20

Yeah, okay.

13:20

Lots of good nods.

13:21

That's great.

13:22

- You use it for your bookie.

13:23

- Good.

13:24

Glad to know that.

13:25

So fun fact about me.

13:27

Hmm.

13:28

We were talking about this

13:29

and wondering what we were gonna say.

13:31

So my husband and I are empty nesters.

13:34

We're currently trying to sell our house

13:36

and when that happens, we are hoping to be homeless

13:40

for about a year and just travel around

13:42

and do Airbnb's in different cities

13:43

for about a month at a time.

13:45

So wish me luck.

13:46

The housing market in Austin is not very good right now.

13:48

- Yeah, all right.

13:49

- All right.

13:50

- All right.

13:51

- Hi everyone.

13:52

My name is Aaron Ashkenazi.

13:53

I'm a chief customer officer at Sentinel-1.

13:56

Probably less familiar than Calendly.

14:00

Sentinel-1 is a cybersecurity company.

14:01

Who knows Sentinel-1, by the way?

14:04

- Yeah, I think we have even at least one or two customers

14:06

in the audience.

14:07

Cybersecurity company that was founded about 11 years ago.

14:12

I've been there for 10 years.

14:13

So I'm in employee number 12.

14:16

I've been running post sales in once, you know,

14:19

some shape or form since I started.

14:21

I didn't have anyone reporting to me the first year.

14:25

There's a team of over 430 people today.

14:27

So everything really ranging from support to success,

14:30

but also managed services, high consulting, training,

14:35

you know, professional services.

14:36

A lot of different functions there.

14:38

Global team obviously.

14:41

The company kind of focuses on, you know,

14:43

B2B, enterprise, cybersecurity,

14:46

input protection, cloud protection.

14:49

You talked about cloud journey.

14:51

We've had a lot of fun during the pandemic grew significantly

14:54

and very fast and went through an IPO process

14:57

about three years ago.

14:59

So we're listed on NICEI if you wanna check it out.

15:02

And a fun fact, I learned to ski when I was 40

15:08

and I'm not really good at it.

15:11

So, yeah.

15:12

- Hello everyone, I'm Jenna Kuntz.

15:17

I am the VP of our global technical support organization

15:21

at Sertinia, formerly known as Financial Force.

15:25

Curious if anyone has ever heard of that company.

15:29

- Perfect.

15:30

So Sertinia is built on the Salesforce platform.

15:37

So much of what I'll talk about today

15:40

will be sort of in that lens of the ecosystem.

15:43

Is any curious if there's anybody

15:45

that also is a Salesforce customer

15:48

or maybe even your own application is on the platform?

15:51

Okay, perfect, wonderful.

15:53

So I've been with Sertinia about three years now.

15:57

I've always been part of a support organization.

16:01

Kind of made my way through Oracle

16:04

and other Motorola back in the day here in Austin.

16:07

That was 25 years ago.

16:08

Motorola is not here any longer,

16:09

which gives you some indication

16:11

of how long I've been around.

16:12

So, but as part of my organization structure at Sertinia,

16:17

we have sort of life cycle of community

16:22

as part of my remit, install deployment like TINA.

16:27

Frontline support, we're not tiered,

16:31

so we're swarming in our company.

16:33

But if you were to look at it as a tier,

16:35

tier one, tier two, tier three.

16:37

And then we have sort of our team

16:41

that kind of parachutes in when there's challenges,

16:43

as well as support the customer success organization

16:45

with things like adoption.

16:47

That's our technical success organization

16:50

and a dedicated escalation team.

16:52

So that is, that's my part at Sertinia.

16:57

Fun fact, gosh, I am struggling with that,

17:00

but I guess I would say I'm originally from Boston.

17:04

I'm a huge Patriots fan.

17:06

I don't know if that's fun, but it's definitely a fact.

17:08

So there you go.

17:09

(audience laughs)

17:11

- And maybe controversial.

17:12

- Yeah.

17:13

- By the way, that's mine too.

17:14

Originally Boston.

17:15

- Hi, I'm Tim Barnes.

17:18

I'm Vice President of Financial Services Support at Oracle.

17:23

I've been there for 20, I don't know how many years.

17:27

I didn't plan on it when I went there.

17:29

I came up through the Austin tech scene

17:33

back in the kind of the turn of 2000.

17:36

And we're responsible for risk and finance products,

17:41

crime and compliance, revenue insurance products,

17:43

things like that.

17:44

We have a global team in probably more countries

17:48

than I could rattle off right now.

17:51

I came up through development.

17:52

I've only been with support for maybe nine years or so.

17:57

So I feel like sometimes it holds me back,

18:00

but also a different perspective

18:01

from being working with our partners

18:04

over on the development side.

18:07

So fun fact about me.

18:09

And I came prepared for anybody that wants to nerd out.

18:12

I was at one of the first Java one conferences,

18:14

so this is a Java ring from 1998.

18:18

And it helped me buy a cup of coffee.

18:20

And that's all I did.

18:22

But it's stainless steel and it looks cool

18:24

and I still keep it in my drawer.

18:26

So that's my fun fact.

18:27

That's awesome.

18:29

Yeah, thank you everyone for joining.

18:32

So I want to start with the AI topic.

18:34

That's front and center.

18:34

It's funny because last year it was the last topic.

18:37

Now it's the first.

18:38

It's been quite a year.

18:40

We talked a little bit, show of hands here

18:42

indicates that two thirds or so are in the process

18:46

or actively using it.

18:48

So let's talk about where do we see

18:50

the benefits in the post sale customer experience.

18:54

I mean, we'll start with Jenna,

18:56

because we've been on the journey with you

18:59

for a good year or so.

19:02

And we'd love to hear your perspective on that.

19:04

And then everyone else, like where do you see use cases?

19:07

Where do you see the biggest opportunities?

19:09

Where are your organizations actively investing?

19:13

Yeah, great question.

19:15

So I think obviously within a support organization

19:17

it's all around scale efficiency.

19:19

And those are the obvious answers.

19:23

And I think as we've been quite thoughtful on our approach

19:26

to AI, we actually coined it the pragmatic AI approach

19:32

internally, both with our own product and our usage of AI,

19:37

but also as we're looking at tools and evaluating tools.

19:40

So I think there's a multitude of business cases.

19:44

Joe showed a really interesting slide,

19:46

I think, on some areas.

19:48

But for me, the focus for me is really around,

19:53

how do I build an experience and elevate

19:56

that experience for our customers

19:58

and capitalize on AI doing it?

20:01

So I think there's loads of opportunity here.

20:06

So I run a global organization.

20:08

And I also have an offshore team.

20:10

And sometimes there are language barriers.

20:11

And sometimes there are challenges

20:13

just bringing that continuity in an experience

20:16

that you want as a support leader.

20:18

So I think as we're evaluating tools and assessing,

20:22

we want that human touch.

20:24

We want to lead with empathy.

20:25

So the irony of using AI to bolster

20:28

some of those situations is kind of interesting to me.

20:32

But also, yeah, that's definitely a focus area for us

20:37

this year in terms of investments

20:39

that we're looking to make.

20:41

Also for our Knowledge Center Services Knowledge articles,

20:46

a brilliant area to also look at

20:48

in terms of the efficiency and productivity and consistency

20:51

you can get with AI building your knowledge articles

20:54

and your Knowledge Center Services and Knowledge Base.

20:57

Same with community, there's opportunities there

21:00

as customers enter through your stream

21:02

and the experience that you're offering right

21:04

from the get go with self-serve versus assisted.

21:07

So I think, yeah, I could keep going,

21:10

but I'll pass the mic.

21:12

>> Thanks, John.

21:13

How about 10?

21:15

>> So yeah, almost end to end, the entire process,

21:18

there's opportunities here from the time.

21:20

And I know you talked about some of the kind of self-heal

21:24

before the issues even happen, but issues will happen.

21:27

That's why jobs.

21:29

So when they do happen from the time the tickets logged,

21:33

how we're interacting with the customer

21:34

to gather as much information as possible,

21:37

make sure that data is as structured as possible

21:40

to get to the agent, to where they can then look at that

21:43

and quickly make a decision, or even cross-reference.

21:46

You talked about Knowledge Base,

21:48

cross-reference other areas that might be similar.

21:50

I don't think we've gotten to the point

21:52

to where we're so confident in the solution

21:56

that we would say, in some rare cases where we would say,

22:00

"Here's the solution."

22:02

And matter of fact, I'm so confident

22:03

close the service requests, right?

22:05

Especially with some of our customers

22:08

that are doing things like regulatory compliance

22:10

or crime and fraud.

22:13

We wanna be sure, and that requires human interaction.

22:17

And that sort of human interaction that you said,

22:20

it's never gonna go away.

22:21

We don't expect it to.

22:23

To take every step along the process

22:25

and make the agent's life easier.

22:27

And then when they are interacting with the customer

22:29

to make sure it's as high value as possible.

22:33

That's what we're really looking for.

22:36

We know that they've been sold value, product value.

22:39

And now you move to the services side

22:41

and you want them to see the value of the service as well.

22:44

So I think every point along,

22:46

we've identified some sort of opportunity.

22:50

And then it just becomes, what do you do first?

22:53

How are we gonna build it?

22:56

How are we gonna test it?

22:57

Do we trust it?

22:57

And trying to collaborate with all the different lines

23:00

of business to make sure you're doing it the right way.

23:02

>> Well said.

23:03

Thanks, Tim.

23:04

Tina, how about you?

23:05

>> Sure.

23:06

So I think about these AI technologies

23:09

in kind of two separate ways.

23:12

And one is really thinking about the internal team's usage

23:16

of the tools in order to drive kind of an agent assist

23:21

experience such that there's technology right there

23:26

that can help them with a customer in a live chat

23:29

in the moment.

23:30

And our goal is to basically serve our customers

23:34

via live chat for as greater than 80%

23:37

of all of our interactions.

23:39

So having an agent assist tool right there in the moment

23:42

is really, really key for us and important.

23:44

And I can talk more about it later,

23:46

but that's kind of one area.

23:48

And then the other area that we have been experimenting with

23:52

and have not yet nailed what we're looking for

23:56

is in the customer's chatbot experience.

23:59

So if you think about I'm a customer,

24:01

I come to the website or I'm in the application

24:04

and I'm looking for help,

24:05

the first place you're gonna go to

24:07

because you're authenticated and we know who you are

24:09

is you're gonna go to a chatbot.

24:11

And that chatbot is going to help you with your questions

24:16

and provide you and route you to a level of service

24:18

that we deem as the right path based on who you are.

24:22

And so we have been testing generative AI

24:26

through that chatbot extensively.

24:29

And I think it's also okay to say that we haven't found

24:32

we haven't found the solution that we're looking for yet.

24:35

Our declarative answers,

24:36

which are the answers we write for the chatbot,

24:38

have about a 92% deflection rate at this point

24:40

or a success rate.

24:42

And the generative AI solutions that we've been testing

24:45

to have not been able to cross that bar yet.

24:47

So I also think you have to kind of think about it

24:51

in terms of what kind of solution are you offering

24:53

if you're a regular e-commerce,

24:55

e-commerce product or solution,

24:57

that's a lot easier to train a chatbot on

24:59

than it is to train on a more technical type of solution.

25:04

So I kind of think of those two areas.

25:07

There's obviously tons of other,

25:09

other areas like translations and all kinds of things.

25:12

But I think of it in terms of how we're helping our customers

25:15

and what I want to use generative AI for

25:17

to be that front door of support, not quite there yet.

25:20

And then also the agent assist for internal usage,

25:24

which we're having really good luck with.

25:26

>> Good to hear.

25:28

>> Or on speaking of more technical products.

25:31

>> Yeah, actually, I like what Tina said about

25:33

like looking at it internally versus externally

25:35

because that's actually what we're doing as well.

25:38

We have a very complicated product.

25:40

It's like B2B operating systems.

25:43

Some of it is like kernel based, malware,

25:46

all sorts of things that are happening over there.

25:49

And it's very, very technical, you know,

25:52

in terms of like the support infrastructure,

25:54

very different than like kind of a pure SaaS product

25:57

because there's a hybrid deployment that includes an agent

26:00

in most cases.

26:01

And so, you know, we're, honestly,

26:05

everywhere I look, I see opportunity.

26:07

And so if you go back to that hype kind of chart

26:10

that you showed, I think there's so much hype

26:12

and we're like trying to figure out really

26:14

where do we want to put our eggs in which basket

26:16

because there's opportunities here all over the place.

26:18

We are experimenting with,

26:21

we started with an initial AI based tool

26:23

which actually is kind of like connected

26:26

to our intranet to some degree.

26:27

And that gives you the ability to ask simple questions

26:30

and get things like from different knowledge base,

26:32

internal knowledge base and/or Slack channels

26:35

and other, you know,

26:36

powerport presentations and things like that.

26:38

And then we're like,

26:38

maybe we can do more with that.

26:40

And we're starting to look into other areas

26:43

primarily because we still don't trust

26:46

the hallucination factor, right, of the AI

26:48

and the fact of the accuracy that you were talking about.

26:50

So we're starting internally for it.

26:51

Like let's see what we can do for the agent itself

26:55

to help them be a little bit more productive.

26:57

The problem there is that it's hard to really understand

27:00

how much more productive are you actually making that agent.

27:04

But then really the next step is looking at

27:06

the AI chatbot, externally chatbot.

27:09

We actually did not have any chat as a medium so far

27:13

just because of that.

27:14

Like we just couldn't really get,

27:16

I would say good outcomes from like the standard,

27:20

you know, templated responses.

27:22

But that might be an opportunity for us.

27:23

And then we're looking at so many other areas

27:25

like from, you know, churn prediction and mitigation

27:29

to like understanding where there's up

27:31

so potential with customers and so on.

27:32

So it's really like all over the place.

27:35

And now I just, I have a relatively small op-steam

27:39

that needs to work with IT and I need to like prioritize

27:41

and pick and choose, okay, what am I investing in right now?

27:45

Let's start with what I feel as the most ROI

27:48

and then maybe a call out for the vendors.

27:51

You know, when vendors talk to me, I say,

27:53

you know, they tell me, do you have like a budget for an AI tool?

27:55

And I'll be like, no, I don't actually.

27:58

But if you can prove to me, because I have a team

28:01

of hundreds of people, you can prove to me

28:03

that you can save me, you know, the effort of 10 people

28:06

while improving some customer experience,

28:08

then yeah, it's easy for me.

28:09

I mean, I'll write the check tomorrow.

28:10

I change my AOP and I put, you know, from, you know,

28:13

from the FTEs I put it in the Opics.

28:14

So.

28:15

- Good tip for everyone for that.

28:19

So let's stay with you, Aaron.

28:20

Let's flip the question over and talk about

28:24

what's getting in the way.

28:25

We've talked a lot about the opportunities

28:27

for AI.

28:29

There's a lot of them.

28:30

But there is still some hype

28:31

and there is still some resistance internally.

28:33

So what are the challenges you're facing

28:36

that are preventing you from moving from us?

28:38

- Yeah, well, it's, you know, one thing

28:41

that one of my VP sent me was like,

28:44

remember there was like, there's an app for that, you know,

28:46

like the commercials of Apple.

28:47

So apparently there's a website that says,

28:49

there's an AI tool for that.

28:51

So you write like what you're looking to do.

28:53

And apparently there's like, you know,

28:54

a couple of like a thousand plus AI tools already there.

28:57

I think that one of the problems, like, where do I go?

29:00

Like, you know, which tool do I actually invest in?

29:02

A lot of the tools are actually promising a lot more

29:05

than what they can actually deliver.

29:07

And I think that to me, that's like a first problem.

29:10

Second thing is that we're already inside a year,

29:14

which probably we did not plan for an AI project.

29:18

Yet we realized probably in the last couple of months

29:20

or I don't know, three to six months that we needed.

29:23

And now we need to figure out how we're gonna do it.

29:25

That's why I talked about kind of the conversion

29:27

between ROI, like instead of investing more in FTEs,

29:31

actually getting that from a tool.

29:32

I do not have any issue actually

29:34

from like executive leadership.

29:36

In fact, it's my CEO that's like pushing the entire company

29:39

GoAI.

29:40

And I have to say one thing about it is that

29:41

we are an AI company to some degree in cyber security space.

29:46

You know, it was called machine learning.

29:48

We didn't know, you know, that, you know,

29:50

mature into an AI models eventually.

29:52

And also we integrated our own, you know,

29:55

we call it an AI analyst or a co-pilot.

29:57

We call it purple.

29:58

It's basically a capability that's gen AI within our console

30:02

that allows an analyst to do their job

30:04

without really knowing that much.

30:06

Like they could do a threat hunting, for example,

30:08

without knowing how to write a query, you know.

30:10

And so we're actually going to use some of that technology

30:14

into the use case of support,

30:16

because in my mind, one of the things that we need to do

30:18

is to bring things as close as possible to the customer.

30:21

Like if we can, if the chatbot resides within,

30:24

you know, the first thing that the customer sees,

30:26

maybe even before the community, maybe before the agent,

30:30

that's a win, right?

30:31

So, but there's a lot of hurdles.

30:34

So I'm like, we're still in experimenting, right?

30:37

Katja, who wants to add to that?

30:39

Like I guess, one of the challenges I think

30:43

is the complexity of the space.

30:45

I don't think, you know, we fully understand it.

30:49

I don't think maybe if we're talking to our peers

30:54

that they fully understand it.

30:55

And so that hype curve that you showed,

30:59

it kind of works to our disadvantage,

31:02

because people say, "Well, wait until it's gone past the hype."

31:05

But some of it is in many things, it is the name, right?

31:10

AI itself makes people think of very complex.

31:16

But if you take away the AI and you say,

31:18

"I want to automate things.

31:20

"I want to industrialize things."

31:22

Things we've all been doing for a long time.

31:25

And now I want to do it with Gen AI,

31:27

or I want to tack on some additional things.

31:29

It becomes much more palatable than just saying,

31:32

"There's this huge thing that only data scientists

31:36

"and PhDs understand."

31:38

And I think that would get us past some of those hurdles.

31:40

In full disclosure, in the first few years

31:42

of the support logic business,

31:45

we didn't really tout the AI.

31:47

It was just the technology we used to solve problems.

31:49

And then really in the last year with this tailwind,

31:53

we said, "We kind of have to."

31:55

If we don't, we're crazy.

31:58

So we sort of jumped on that bandwagon as well,

32:02

and to sum it to hype cycle, but so on and so forth.

32:06

Tina, has it gotten easier or harder?

32:10

You saw the trend data from the last six months

32:13

to get your peers in senior leadership on board?

32:16

- No, it's not hard.

32:18

- No.

32:19

It's not hard.

32:20

I think we have a lot of people at the company

32:23

who are kind of chomping at the bit

32:24

and really excited about new technologies.

32:28

We're an innovative company.

32:29

We track folks who really care about innovation,

32:32

and it's kind of, you don't want to call it

32:35

the shiny new toy, but it is to a certain extent,

32:38

and we've been out evaluating just a number of technologies

32:41

and have done proof of concepts

32:42

with a number of different companies.

32:44

And what I would say is,

32:46

like to go back to kind of the barrier question,

32:49

don't expect smashing success out of the gate, right?

32:54

I would say, like in our case,

32:56

we're in a proof of concept right now

32:58

with a specific technology,

33:00

and we've been able to prove in just a few short weeks

33:04

that we can get in a minute,

33:05

we can save a minute and a half on every, on average

33:09

for every live chat that we do,

33:12

and our average live chats are about 16 minutes,

33:14

so you can do the math, and the next thing you know,

33:17

that adds up when you're doing 30 to 50,000 support requests

33:20

in a month, right?

33:22

So have realistic expectations, start small,

33:26

get a small win, and then it multiplies it as a,

33:29

like a multiplier effect,

33:30

and that's just for one particular sliver of age and assist,

33:34

and so if you think there's way more you can do with it,

33:38

and have that mindset, then it's not hard

33:40

to get people excited about it,

33:43

and you know, there are a lot of companies out there right now

33:46

who are happy to do proof of concepts

33:48

because the technology is new, and they wanna prove it,

33:51

and they want customer testimonials in order to do it,

33:55

so you know, as leaders of post-sale support organizations,

34:00

use that to your advantage.

34:02

- That's a good tip.

34:03

Jenna, with the journey you've had over the past year plus,

34:07

what are some of the barriers you're still facing?

34:10

- I'd say one of the areas that you really wanna consider

34:13

is organizational change management with AI,

34:18

and really any technology, but I recently had to,

34:23

was asked to speak to a company town hall,

34:27

and the topic was of course AI,

34:29

because that is the topic du jour,

34:31

and I didn't consider the impact to my organization,

34:36

and maybe even like a support analyst,

34:39

listening to me as a leader speak about how we were looking

34:43

at it and how we were approaching it,

34:45

and a lot of the survey feedback that we got on that

34:49

all hands was, "Is my job going away?"

34:52

And so we really have to reframe those things,

34:55

and just talk about it with intention,

35:00

and what it really means for you and your organization.

35:03

For us it is not about, it's not a head count reduction exercise,

35:08

it is truly about creating opportunities and pockets

35:11

for our people to do more high value work,

35:13

and again elevating that support experience for our customers,

35:17

so that we can free them up from the mundane,

35:20

and create those efficiencies and those pockets of opportunity.

35:24

So I actually ended up having an all hands,

35:29

just try to walk the team through this to say,

35:31

"Hey, look, we are pursuing these,

35:35

like every other company,

35:36

and oh by the way, our own company is employing AI,

35:39

and it brought it to market a couple of years ago,

35:42

so it's something that you really want to consider.

35:47

For those early in the journey,

35:49

just think about how your teams may respond to this,

35:52

because people will have different feelings about it along the way.

35:56

There's the hype and the noise,

35:57

but then there's other people that are fearful.

36:00

>> Well said, shifting gears a bit,

36:02

you mentioned a phrase that is near and dear to my heart,

36:05

support experience,

36:07

and the fact that we have two support leaders,

36:11

and then two chief customer officers here on the panel

36:14

is not a coincidence,

36:15

we want to get a slightly different viewpoints,

36:18

and then talk about the collaborative opportunities,

36:20

but first Jenna, let's keep it with you.

36:22

What does support experience mean to you and your organization?

36:26

>> So I wake up thinking about how our customers

36:34

and how they're treated, the technology solutions

36:38

that we're bringing to them, how quick we were,

36:40

the quality, both from a technical standpoint

36:43

and from an engagement standpoint,

36:45

what that looks like, holistically,

36:47

and the entire journey.

36:49

You really have to kind of map that journey

36:52

so that, by way of example, we were on a call

36:55

about the whole leadership team today together

36:57

to talk through our case submission form,

37:00

something that simple, but really it's the first experience

37:03

your customers are seeing,

37:05

and they're making impressions of you right then and there.

37:08

So we're looking at it, we're walking down it,

37:10

and the team had a lot of really interesting observations,

37:13

and quite frankly, a few of them have never even seen

37:15

the case submission form, so I was just shocked,

37:17

and I thought, this is the kind of stuff,

37:19

those details matter, what does the survey say

37:22

when it goes out, in that whole entire journey,

37:25

you really need to really just look at it.

37:28

And then, I think you said end-to-end,

37:29

and I agree with that.

37:31

>> Tim, what are your thoughts on support experience?

37:34

>> Yeah, I mean, I think it comes back to providing

37:38

absolutely the best interactions that you can.

37:42

We have some really high performers on the team,

37:46

and they really want to provide the best service possible,

37:50

and so we do want to take those more mundane tasks

37:54

and give them the time to provide the best service.

37:57

We want our customers to feel like we're their partners,

38:00

and they have goals.

38:04

Every one of our customers has a different set of goals

38:08

that they want to provide to their customers,

38:10

they can only do it.

38:11

We feel with our products,

38:13

and so we want to help them to achieve that.

38:15

And the other thing I think is that

38:18

anyone that's in the services area,

38:21

you really want your customers to become your advocates,

38:26

to become your champions.

38:30

And I heard someone say that

38:33

certainly if someone's going to be a reference for you,

38:36

maybe sometimes you have to twist their arm a little bit

38:40

in order to be a reference for you,

38:42

but they're your champion.

38:44

They're going to talk about you when you're not in front of them.

38:47

And those are the kind of customers that we want to build,

38:50

and we're the ones that are interacting with them

38:52

on a very frequent basis after they bought the product.

38:55

So that's what we have to strive for.

38:57

>> Good point, I saw a stat that said that in B2B,

39:01

especially complex, B2B,

39:04

more than half of customer interactions

39:07

occur through support.

39:08

So that your function really is front and center

39:11

in many cases and responsible for delivering

39:14

an amazing experience.

39:16

Let's hear from the chief customer officers in the room.

39:21

We've had a decade, we were talking earlier

39:24

about your path, customer success,

39:27

and then professional services,

39:30

but there's been this convergence and collaboration

39:32

between customer success support

39:36

in really delivering a more holistic, unified customer experience,

39:40

post-sale customer experience.

39:42

From your perspective, what's working well,

39:44

and maybe what are some opportunities for,

39:47

what are opportunities for improving the collaboration

39:50

between support and success?

39:52

Tina, do you want to start?

39:53

>> Yeah, so when I think about this,

39:56

I think about, you know,

39:57

I'm responsible for all the post-sales teams,

39:59

and so rather than thinking about how support supports

40:03

their customer and customer success supports the customer

40:06

and community, which is also under my remit,

40:09

you know, the community supports the customer

40:11

or in the help center, I don't think about that way.

40:13

I think of it in terms of we have a customer

40:16

and we need to think about their experience

40:18

and where they get help from in the moment doesn't matter.

40:22

They just want their problem or their question solved.

40:25

And so if you put the customer view first,

40:27

then you're going to make decisions

40:29

about how the teams support that customer

40:31

as opposed to the teams that are coming first.

40:34

And so by that I would just mean we architect, you know,

40:37

journeys for the customer that we try to make it

40:40

as effortless as possible.

40:41

So an example of this is we want to make sure

40:44

that our teams that are frontline

40:46

have the information they need

40:48

that will be helpful in the moment.

40:49

So for example, if I'm a support agent,

40:53

I know because of the integrations

40:54

that we've been able to create through our tech stack,

40:56

I know if that customer is in there a first 14 days,

40:59

which for us is really important

41:01

because that means they're on a trial

41:02

and we're trying to convert them to a paid plan, right?

41:06

If there are a customer that's been with us for two years,

41:09

we have, you know, segmentation that we put in place

41:11

based on product usage data that tells us

41:14

and they put them into a profile of this customer's

41:17

at risk for these reasons and, you know,

41:19

we have analytics behind it.

41:20

And so the agent knows exactly what playbook they need

41:23

to run and what questions they need to ask

41:25

based on those profiles.

41:26

And so we're putting data into the agent's hands.

41:29

You know, conversely in the community, for example,

41:34

we know when somebody goes to the community

41:36

and then we know when they go to support

41:38

and then the CSM, the customer success manager,

41:41

can see and gain side, for example,

41:43

that they had a support interaction yesterday

41:46

about this topic and they went to the community

41:48

earlier today about that topic.

41:51

And so we get kind of this 360 degree view.

41:53

And so again, it's about the customer first

41:55

and our protecting that journey

41:57

and making sure that all of the data is integrated

41:59

amongst all of the different technology solutions

42:02

that we have so that we can see what's going on

42:04

and it's not just, I'm in the community

42:06

and I only talked, you know,

42:07

I only just done what's happening

42:08

with my customers in the community

42:09

and I only do support and so I only know what they're up to.

42:12

We just try to make it more on the channel.

42:15

- Mm-hmm, that's fantastic.

42:17

Or with your perspective?

42:18

- Yeah, I mean, obviously, I mean,

42:21

echoing a lot of what Tina said,

42:22

but also I think that there's the realities that,

42:27

you know, the groups actually a lot of times

42:29

operate within silos.

42:31

They all have the best intentions,

42:33

but we know also that the road to hell

42:35

is paved with that, right?

42:36

So we have to make sure that, you know,

42:38

and I think this is where technology comes to play,

42:41

you know, that technology helps basically

42:43

connect those groups together.

42:44

And I think that, you know, a lot of times

42:46

you see situations and when you think about

42:48

a customer journey, there's actually this pyramid

42:52

that I constructed and I like to show it

42:55

almost in every like sales kick off

42:56

and customer event that we do,

42:59

which is a take on the muscle pyramid of needs.

43:02

So the muscle pyramid of needs, you know,

43:03

you start with the basics and you go, you know,

43:05

all the way up to like, I don't know,

43:07

feeling accomplished, et cetera.

43:09

And I took that and I put it in a CS mindset

43:11

and I said, well, it's start with,

43:13

keep the business running or like,

43:15

I like to say, don't break my shit.

43:17

So like, if I install a piece of software,

43:19

I wanna make sure that it doesn't really ruin anything else.

43:21

That's the first thing.

43:22

The second thing is that it needs to work,

43:24

like do what it's supposed to do.

43:25

Like, I don't know, set meetings

43:26

or protect my endpoint or whatever, right?

43:29

And then you go into like the Assistant and advice piece,

43:32

which is like, all the support pieces,

43:34

the success, the professional services,

43:36

and eventually you get to that end of the pyramid

43:39

to meet the delight piece.

43:41

That's where the customers becomes your fans.

43:43

That's where they start talking about you.

43:45

That's where they come to panels, right?

43:47

And like, you know, talk about their experience, I guess.

43:50

Jenna had a good experience with you guys, right?

43:52

So I think that like, I'm trying to convey that,

43:55

not just by the way to my teams,

43:57

I'm trying to convey that to the rest of the organization.

44:00

I go with that to sales.

44:01

That's like a pitch I talk to the Salesforce as well.

44:04

Because to me, customer experience is, of course,

44:06

it's everything we do within my org,

44:08

but it's also like the way that

44:10

the quoting process looks like,

44:12

or it's the way that the sellers work with the customer,

44:15

how they interact in case of a escalation, for example.

44:18

Instead of like being an amplifier,

44:20

they work together with the team and we're one team.

44:22

We're not like sales here and this there

44:25

and whatever and product and engineering.

44:27

So, you know, I'm trying to bring all of that together.

44:30

I'm not saying that it's all successful.

44:32

I'm saying that that's the vision

44:33

and that's where we're aiming in terms of a target.

44:36

- Great, thank you for sharing that.

44:38

I wanna go back to Tina for one question.

44:40

You mentioned it earlier.

44:42

Being a product-led growth company,

44:44

you talked a little bit about the playbooks

44:47

and for helping get customers successful early,

44:50

especially I think you mentioned a 14-day trial period.

44:54

What are some of the other challenges

44:56

that you're facing in terms of that,

44:58

delivering that amazing experience early on

45:02

being a PLG company?

45:04

- Yeah, that's a good question.

45:05

So, PLG basically means product-led growth

45:08

and it's where the philosophy is that customers

45:11

should be able to self-serve basically everything.

45:14

And I think there's this healthy tension

45:17

between we as a support organization,

45:22

we quantifiably know that if we're able to help a customer

45:25

in the first 14 days, actually first 30 days,

45:29

that we have a conversion rate that is significantly higher

45:33

so they convert to a paid plan that's significantly higher

45:36

than if they don't talk to support at all.

45:38

And the challenge there is that we have

45:40

like 1% of customers in their first 30 days

45:44

actually talk to support, right?

45:46

So, it's a healthy tension between

45:48

you wanna be a self-surface product

45:51

and at the same time, we know if we can talk to them,

45:54

we can get them to convert, right?

45:56

And so, I think that's probably the healthiest tension

45:59

is it's like, do you wanna have a call to action

46:02

in your application to contact support

46:04

when you're in your first 30 days or 14 days?

46:07

And then do you wanna give other resources after that?

46:11

So, it's like this healthy,

46:12

I think that's probably the hardest thing to--

46:14

- So, you're literally paying off what Tim was saying earlier

46:17

where the support experience is actually benefiting

46:20

the customer's experience and the conversion

46:22

and retention of that customer.

46:24

- We have, yeah.

46:25

We can quantify the impact.

46:27

- That's fantastic.

46:28

One more question for you all

46:30

and then I wanna open it up for some audience questions.

46:33

AI and all this conversation around new technologies

46:37

brings the opportunity for skill development.

46:41

You all have a diverse set of teams within your functions.

46:46

How are you all thinking about training for this new era?

46:50

One of the questions we were talking about earlier

46:53

is do you hire more generalists or specialists?

46:57

Is the role of all being to just become

47:00

essentially prompters for the AI?

47:03

Or are we thinking about it too simplistically?

47:05

How do you think about skill development for your teams?

47:09

Tim, do you wanna start?

47:11

- Sure.

47:12

I think the team that we have right now

47:15

that's focused on automation,

47:17

which will expand out to be more AI or J&AI

47:22

or some of these other tools,

47:24

those are gonna be our specialists.

47:26

And, but I do think that anyone that's using these tools,

47:31

even if it's an internal chat bot

47:33

that's just a natural language model,

47:37

which is giving suggestions,

47:39

either in solutions or helping you look up things,

47:44

unified search and things like that or topic clustering.

47:48

There's a variety of things you can do internally,

47:50

but we can't just say, here's the tool, use it.

47:53

They need to have a base understanding.

47:55

I think when Cloud, you mentioned Cloud having this explosion

47:58

years ago, and you start to then look more

48:03

when you're hiring people or when you're training people,

48:05

you say I'd like to have somebody

48:07

that at least has that base understanding.

48:09

You hear it on the sales side all the time,

48:11

selling on-prem and selling Cloud very different.

48:14

You have to have a base understanding

48:16

before you can move forward.

48:18

And I think support's gonna be no different.

48:20

We have to have everyone trained at a generous level

48:23

that's gonna be using these tools,

48:24

but the people that are building them

48:27

and trying to convince others to invest more in them,

48:32

they have to be our specialists.

48:33

It's application and special depth.

48:34

Jenna?

48:37

>> Yeah, same.

48:38

I mean, great, great response.

48:40

I definitely think it's a hybrid model.

48:42

And I definitely think that as you consider tools

48:47

that you're introducing, how are you,

48:50

we stumbled a little bit when we implemented

48:52

support logic in terms of our adoption of the tool

48:56

and our preparedness for teaching the tool.

48:59

So with support logic sort of predictive,

49:02

analysis of, hey, this looks like it's likely to escalate.

49:06

You're offered an opportunity to say,

49:08

yes, that's true or no, it's not.

49:11

And we were a little slow on the uptake,

49:12

but I think we've got it, we're cooking on gas now

49:15

because it is almost like the accuracy of it is scary good.

49:20

So over time, we were able to kind of build

49:23

that generalist community that's coming in

49:25

and telling it, yes, this is real, this is not.

49:28

And our touch points on that are lessening.

49:30

But to your point, when I look to the future

49:33

with things like co-pilot, yeah,

49:35

you absolutely need specialists

49:38

for that type of advanced technology.

49:41

- Just tangential follow up to that.

49:44

What would be helpful to make you get

49:46

over that learning curve faster?

49:48

- I think it, gosh, that's a great question.

49:53

I think we owned our part in that.

49:56

I think it was more our readiness and our ability

49:59

to really drive it a little quicker.

50:03

But it was, once we began to,

50:07

it came together pretty quickly, relatively speaking,

50:09

it was just a bit of a stumble at the onset.

50:12

- I actually think that, just going back to the example

50:18

of cloud and on-prem, I think that once this technology

50:22

actually really ripens, then it actually should be

50:27

that complicated because when I connect to data

50:29

service, I don't really know if it's on-prem or cloud.

50:32

I mean, there's the saying that cloud

50:34

is just someone else's computer, right?

50:36

So I think that if you think about it that way,

50:39

then eventually it will be very simple and intuitive,

50:42

but in that interim, which could take a year or three,

50:45

or I don't know, whatever, but we will need to have,

50:48

obviously, the data scientists, AI expert types

50:51

that are going to do a lot of the training to the models

50:54

and kind of fine-tuning, and then you need to make sure

50:57

that the general population is actually,

50:59

the adoption there is in such a way that they understand

51:01

that they have a part in it.

51:02

Like, for example, just in a supervised model

51:05

when you say yes or nay, right?

51:07

Like, is it working for me or not?

51:09

Just that thing alone goes back to the data scientists

51:11

and go back to the AI team that changes it,

51:14

but down the road, it would get to a point where like,

51:16

it kind of just works.

51:18

It will take us some time, I don't know how much time,

51:19

but like, that's my view.

51:21

- Fantastic.

51:24

- Yeah, I think in the here and now,

51:25

the way that we've implemented kind of our agent assist,

51:29

that we're doing right now, I think of it kind of even

51:31

as a journey within agent assist.

51:33

So for example, what we're doing right now is you have,

51:38

and this goes to kind of more of the specialist,

51:40

I'm sorry, the generalist topic,

51:43

but the first kind of step in agent assist

51:47

is having the technology do the transitions

51:53

between the different topics a customer is asking about.

51:56

So it does a salutation, does a greeting,

51:59

and then based on how the sentiment comes back

52:01

from the customer, it prompts for,

52:03

let me understand your question in a kind of,

52:06

so it guides the agent not in the answer,

52:09

but in what to say next,

52:11

from like a greeting and sentiment type of thing.

52:14

And that's kind of the first evolution, right?

52:17

The second evolution is really thinking about

52:20

what's called like custom intense.

52:23

So why did a customer reach out to you

52:25

for support in the first place?

52:27

And so we've been able to map all the reasons

52:30

why customers reach out to us with all of our helps

52:32

and our content and a tool we have called Guru,

52:35

which is an internal kind of knowledge base,

52:37

map those together, and then from that,

52:40

the technology is surfacing up answers

52:44

for the agent to then say, yes, this is the answer,

52:48

but you can't, as an agent, you have to know the product,

52:50

you have to have been trained on the product

52:52

because then you become QA and you become,

52:56

you know, you free up your time to be able to work

52:58

on cooler, bigger type projects

53:01

and spend more time training and up level in your skills.

53:04

And so I think that, you know, yes,

53:07

the evolution will be more towards specialist in the future,

53:09

but at this point, the technology is all about

53:12

helping a generalist be just more efficient

53:15

in what they're doing and have better quality

53:17

in the process.

53:19

So generalist and then specialist evolving to that.

53:23

Why don't we open it up?

53:24

Any questions from the audience?

53:26

The question was about any concerns

53:29

in exposing case data to the LLMs

53:33

and sharing some of that customer information.

53:36

So I don't know if anyone wants to think.

53:38

- No direct experience right now, yes, certainly concerned

53:41

and have done some research around large language models

53:47

that can tell lies for the query.

53:50

And so certainly if you were to ask a general statement

53:53

to a large language model, it may give you something.

53:55

But if you say within this context,

53:58

and so there are concepts within AI like vectors

54:02

that siloed off, right?

54:04

Private business data, customer data, things like that,

54:07

which you may have use cases for.

54:09

We haven't gotten that far down that road yet,

54:13

but I think the data science appears to address it.

54:18

But I think it's still early in the days, at least for us.

54:22

- For us, we did examine that concern.

54:27

And it was really a matter of retention, storage,

54:31

and some of those areas that for us

54:35

alleviated the concern.

54:38

The data was not leaving our

54:41

system of records, so.

54:45

- That was really it for us,

54:48

on our factor moving forward.

54:50

- We're just being very deliberate

54:54

about where the technology is pulling information from.

54:59

So for example, we know that we'll get better output

55:02

if we structure the input.

55:04

So all of our help center articles

55:06

have the same kind of construct, taxonomy, et cetera.

55:10

And we're not at this point putting case data

55:13

through those channels, but it's really more about

55:17

the internal help knowledge and customer facing help knowledge

55:20

and you structure it appropriately so that, you know.

55:23

Yeah.

55:28

- I'll add to that, and I'll say,

55:30

because we're talking about JNAI,

55:32

and I'd like to ask how many here know

55:35

that their company have actually a JNAI policy

55:40

related to security and privacy.

55:43

And this is something that, like for example,

55:45

we had to construct and share with the entirety

55:48

of the company.

55:49

I mean, we're a cybersecurity company

55:51

and we're a public company, so that gives us

55:52

a lot of scrutiny.

55:54

That also means that you have to do a couple of things,

55:57

both the tagging, the vectoring,

55:59

kind of leaving some data out and minimizing data, et cetera.

56:03

But also investing sometimes in, basically, in-house,

56:06

LLM enterprise models out of the big kind of AI companies,

56:11

which actually cost a lot of money.

56:13

So some of these things, if you wanna do them yourself,

56:15

especially when I integrate them in your product,

56:17

you have to make a huge investment.

56:19

And other times, you just have to have a lot of scrutiny,

56:22

both on the in and the out, I guess.

56:24

And it starts also from having an awareness,

56:27

because guess what?

56:28

Anyone in the company can go and do a trial

56:31

or do something and start pushing data.

56:33

And all of a sudden, maybe you got some of your stuff

56:35

out there, so that's why it's important to also start

56:37

with internal education and policy

56:40

from a compliance and privacy standpoint.

56:42

>> Very good point.

56:43

>> Thanks for sharing that.

56:45

>> Any other questions?

56:46

>> You need an article for that,

56:47

which is where we're moving towards.

56:50

And as I mentioned at the top of this time together,

56:54

we're using Salesforce Einstein

56:57

for a lot of our next generation AI capability.

57:01

So in this case, we're keying off of some of that.

57:06

We also use a blend of COVEO as part of our tech stack as well,

57:11

that also helps, you mentioned agent insights

57:17

and some other areas.

57:18

So the combination of these tools are helping us

57:21

kind of build out consistency in our article creation.

57:27

What we're not doing well right now is,

57:30

and I'm hoping that we get there with this next level

57:34

of what we're introducing with Salesforce,

57:36

is reducing article,

57:41

like I don't know how you're managing that today,

57:42

but over time, things become legacy real fast, right?

57:46

You're moving new product versions in and out,

57:48

and the articles don't always keep tempo with that.

57:52

So that's an area that we're struggling with today as well.

57:56

We're trying to be careful in some cases

57:59

of not automatically creating new knowledge,

58:03

because there is that maintenance aspect of it.

58:06

So if it can be as structured as possible,

58:10

either in tagging along the way to where you can present

58:14

to the agent a draft of an knowledge article,

58:17

but then also train in the kind of the diagnostic methodology

58:22

of you going through all these steps, does it make sense?

58:25

Is there another one like it?

58:26

Do we have to add yet another thing to maintain?

58:29

So that's really important.

58:31

- We'll take one more audience question.

58:36

- So I'm hopeful that those are all gonna converge,

58:39

and it's gonna be embraced across not just support,

58:43

but all the different lines of business,

58:44

so I'm excited about that.

58:46

- Fantastic, well with that,

58:48

I thank you all for joining us.

58:51

Thank you all as well, and let's continue.

58:53

I think we have an hour plus, more food, cocktails,

58:57

yep, thumbs up in the back.

58:59

So let's continue the conversation.

59:02

- Thanks again.

59:03

- Thank you.

59:04

(audience applauding)

X

This is a test comment.

X

This is a longer test comment to see how this looks if the person decides to ramble a bit. So they're rambling and rambling and then they even lorem ipsum.


author