EP231: Using AI Chatbots To Create Empathetic Customer Service, With Siena Co-founder Lisa Popovici

1
00:00:02,600 --> 00:00:04,420
You're listening to the
Insight Commerce podcast
2
00:00:04,420 --> 00:00:07,140
hosted by experienced
consultants James Gerd and
3
00:00:07,140 --> 00:00:09,900
Paul Rogers. Insight
Commerce is your fast track
4
00:00:09,900 --> 00:00:12,100
to digital knowledge,
featuring interviews with
5
00:00:12,100 --> 00:00:14,160
leading e -commerce
retailers, industry thought
6
00:00:14,160 --> 00:00:16,360
leaders and the C -suite of
key tech companies to
7
00:00:16,360 --> 00:00:18,860
explore the latest digital
strategy, customer experience
8
00:00:18,860 --> 00:00:21,220
and tech. Discover
what the latest trends
9
00:00:21,220 --> 00:00:23,880
really mean and get
practical advice based on
10
00:00:23,880 --> 00:00:25,960
the experience to help fine
-tune your e -com strategy
11
00:00:25,960 --> 00:00:29,160
and tech stack. Follow
us on LinkedIn and reach
12
00:00:29,160 --> 00:00:31,420
out to James and Paul if
you'd like to discuss an e
13
00:00:31,420 --> 00:00:33,540
-commerce project. We
hope you enjoy the latest
14
00:00:33,540 --> 00:00:37,420
episode. Today's
episode is sponsored by
15
00:00:37,420 --> 00:00:39,600
Scale Commerce Engine, who
have productised the scale
16
00:00:39,600 --> 00:00:41,080
and expertise of the
fastest -growing e -com
17
00:00:41,080 --> 00:00:43,120
business in Europe and Swap
Commerce, who help you
18
00:00:43,120 --> 00:00:44,860
improve every pain point in
your opt -to -journey from
19
00:00:44,860 --> 00:00:46,980
shipping, tracking and
package protection to
20
00:00:46,980 --> 00:00:49,760
inventory and returns. So
let me set the scene for
21
00:00:49,760 --> 00:00:51,980
today's conversation. I'll
introduce you to our
22
00:00:51,980 --> 00:00:54,300
amazing guest. We're
talking about using AI chat
23
00:00:54,300 --> 00:00:56,480
balls to create empathetic
customer service. A
24
00:00:56,480 --> 00:00:59,040
really cool topic.
Interesting
25
00:00:59,040 --> 00:01:01,920
chat box is huge amongst e
-commerce businesses at the
26
00:01:01,920 --> 00:01:04,680
moment. We're
talking to Sienna Cofolder,
27
00:01:05,060 --> 00:01:07,700
Lisa, who's a member of
the prestigious Forbes 30
28
00:01:07,700 --> 00:01:09,920
out of 30, which is more
than Paul and I will ever
29
00:01:09,920 --> 00:01:11,780
achieve. So
well done on that, Lisa.
30
00:01:12,000 --> 00:01:12,880
And
welcome to the podcast.
31
00:01:14,780 --> 00:01:15,940
Thank
you. Thank
32
00:01:15,940 --> 00:01:18,420
you, James. Great
to be here. Cool.
33
00:01:18,640 --> 00:01:20,720
So before we ask you lots
of exciting questions, do
34
00:01:20,720 --> 00:01:21,880
you want to just give
people a bit of an
35
00:01:21,880 --> 00:01:25,540
introduction to yourself and
Sienna? Yeah,
36
00:01:25,680 --> 00:01:28,160
of course. I
am one of the co -founders
37
00:01:28,160 --> 00:01:32,800
and CMA of Sienna. And
basically what we were
38
00:01:32,800 --> 00:01:35,860
doing at Sienna, we've
taken on the mission of
39
00:01:35,860 --> 00:01:38,720
humanizing the internet by
pioneering a new category
40
00:01:38,720 --> 00:01:43,780
of empathic AI economists
agents for commerce. And
41
00:01:43,780 --> 00:01:47,080
we've built Sienna as the
first and leading solution
42
00:01:47,080 --> 00:01:50,700
in this space. And
our goal is to transform
43
00:01:50,700 --> 00:01:53,360
our brand's approach,
customer experience and
44
00:01:53,360 --> 00:01:58,640
customer service. Amazing.
I love that position as
45
00:01:58,640 --> 00:02:00,400
well because you describe
it obviously about creating
46
00:02:00,400 --> 00:02:04,060
empathetic AI customer
service rather than just
47
00:02:04,060 --> 00:02:06,300
giving people more
technology and about
48
00:02:06,300 --> 00:02:08,979
humanizing the internet. So
I'd love you to, can you
49
00:02:08,979 --> 00:02:11,460
expand on that? So
can you talk us through
50
00:02:11,460 --> 00:02:13,520
what that really means for
brands and how you see
51
00:02:13,520 --> 00:02:15,960
that as different to just
like standard chat bot
52
00:02:15,960 --> 00:02:19,760
interactions? Yeah,
of course. So
53
00:02:19,760 --> 00:02:24,080
basically what Sienna is is
an AI agent that helps
54
00:02:24,080 --> 00:02:28,980
brands and teams kill their
customer service and it
55
00:02:28,980 --> 00:02:33,160
delivers emotionally
intelligent personalized
56
00:02:33,160 --> 00:02:35,780
customer service at scale.
So
57
00:02:35,780 --> 00:02:40,480
basically what this category
represents is a paradigm
58
00:02:40,480 --> 00:02:43,480
shift in how customer
service has been done so
59
00:02:43,480 --> 00:02:48,080
far where these AI powered
agents can perform roles
60
00:02:48,080 --> 00:02:51,240
traditionally managed by
humans. And
61
00:02:51,240 --> 00:02:54,680
you can think here of very
advanced roles like taking
62
00:02:54,680 --> 00:02:57,160
care of transactions or
managing subscriptions,
63
00:02:57,840 --> 00:03:01,520
taking care of returns. So
basically actually solving
64
00:03:01,520 --> 00:03:05,400
customer's problems, not
just, you know, foraging
65
00:03:05,400 --> 00:03:08,520
tickets and giving agents
actually more work at the
66
00:03:08,520 --> 00:03:11,020
end of the day. As
that's, you know, what
67
00:03:11,020 --> 00:03:13,940
traditional chat bots have
been doing so far. They're
68
00:03:13,940 --> 00:03:17,540
very good at like, you
know, the first year of
69
00:03:17,540 --> 00:03:20,020
interactions, maybe, hey,
how can I help you today?
70
00:03:20,100 --> 00:03:22,440
But
then once a customer places
71
00:03:22,440 --> 00:03:26,900
their question, which
obviously customers questions
72
00:03:26,900 --> 00:03:29,320
are pretty complex, they
might ask two or three
73
00:03:29,320 --> 00:03:33,340
multiple questions that
require you to take some
74
00:03:33,340 --> 00:03:36,460
steps into a different
vendor or, you know, take
75
00:03:36,460 --> 00:03:39,880
care of transaction, the
chat bot will be lost. So
76
00:03:39,880 --> 00:03:42,540
we'll be not be, we'll not
be able to help the
77
00:03:42,540 --> 00:03:45,280
customer. So
that's when, you know, an
78
00:03:45,280 --> 00:03:47,840
agent, a human needs to
step in. And
79
00:03:47,840 --> 00:03:52,280
that's why if probably if
you shopped online, you've
80
00:03:52,280 --> 00:03:54,820
also been very frustrated
when you dealt with the
81
00:03:54,820 --> 00:03:57,380
chat bot was just not able
to understand what you
82
00:03:57,380 --> 00:03:59,680
need. So
you probably type, Oh, talk
83
00:03:59,680 --> 00:04:01,040
to human talk to human.
This
84
00:04:01,040 --> 00:04:03,720
is useless. So
that's what we're trying to
85
00:04:03,720 --> 00:04:06,160
avoid. That's
what we're disrupting. And
86
00:04:06,160 --> 00:04:09,820
these AI agents, they can
offer 24 seven service
87
00:04:09,820 --> 00:04:14,260
significantly reduced
operational costs and really
88
00:04:14,260 --> 00:04:17,959
enhanced customer experiences
with a very consistent
89
00:04:17,959 --> 00:04:23,520
personalized approach across
all channels. Great.
90
00:04:23,820 --> 00:04:26,900
And I guess just we'll
obviously come back to the
91
00:04:26,900 --> 00:04:28,900
CX side and, you know,
some of the mechanics
92
00:04:28,900 --> 00:04:30,740
around what you just talked
about. But
93
00:04:30,740 --> 00:04:33,340
before we go into that,
can you just talk us
94
00:04:33,340 --> 00:04:37,060
through like how you see
brands implementing AI at
95
00:04:37,060 --> 00:04:39,520
this stage and how you
think, yeah, brands and
96
00:04:39,520 --> 00:04:44,500
retailers should be looking
to utilize AI right now?
97
00:04:45,720 --> 00:04:48,020
Yeah,
for sure. So
98
00:04:48,020 --> 00:04:51,980
I think there's a few
criteria that you have to
99
00:04:51,980 --> 00:04:55,160
look at to know that you
are ready and the timing
100
00:04:55,160 --> 00:04:58,960
is right. So
we typically look at the
101
00:04:58,960 --> 00:05:01,400
ticket volume. So
I feel like once you're
102
00:05:01,400 --> 00:05:05,500
around 3000 tickets per
month in your help desk
103
00:05:05,500 --> 00:05:08,840
across all channels, then
you're ready to to adopt
104
00:05:08,840 --> 00:05:11,700
AI. And
also we look at the team.
105
00:05:13,180 --> 00:05:15,720
If
if you have at least a
106
00:05:15,720 --> 00:05:18,740
few in -house members that
are doing the same
107
00:05:18,740 --> 00:05:21,740
repetitive work over and
over again, maybe you're
108
00:05:21,740 --> 00:05:24,620
also outsourcing part of
your customer service. Then
109
00:05:24,620 --> 00:05:28,540
you know that it's a good
timing to implement AI. So,
00:05:30,440
you know, first aspect,
111
00:05:30,680 --> 00:05:33,360
it's timing, right? You
need to be ready. You
112
00:05:33,360 --> 00:05:36,560
need to have at least some
processes in place, some
113
00:05:36,560 --> 00:05:40,240
knowledge in place so that
you can use to actually
114
00:05:40,240 --> 00:05:43,080
train the AI. Because
what's very cool about
115
00:05:43,080 --> 00:05:46,320
tools like the way they're
being trained, it's very
116
00:05:46,320 --> 00:05:49,460
close to how humans are
being trained. You
117
00:05:49,460 --> 00:05:51,900
have to have some sort of
a repository or some
118
00:05:51,900 --> 00:05:55,440
documents, you know, some
handbooks that you that you
119
00:05:55,440 --> 00:05:58,640
give to your new employees
once they join. And
120
00:05:58,647 --> 00:06:00,780
then that's their like
onboarding guide. So
121
00:06:00,780 --> 00:06:04,320
you have to have these
these things into place, at
122
00:06:04,320 --> 00:06:09,240
least like a basic level
of documentation and then
123
00:06:09,240 --> 00:06:10,960
team and then ticket
volume. So
124
00:06:10,960 --> 00:06:13,800
I think these three things
are are what brands should
125
00:06:13,800 --> 00:06:20,560
be looking at. And
I feel like, yeah. So
126
00:06:20,560 --> 00:06:21,520
it's a reason to be that.
I
127
00:06:21,520 --> 00:06:24,000
mean, so I guess it'd be
interesting just outside of
128
00:06:24,000 --> 00:06:28,400
CX as well. So
outside of kind of support
129
00:06:28,400 --> 00:06:31,240
and Sienna, like, where do
you see and I guess,
130
00:06:31,400 --> 00:06:32,740
Andrew is a good example
here. We
131
00:06:32,740 --> 00:06:34,940
both know from current
model, you will hopefully
132
00:06:34,940 --> 00:06:36,280
come on to talk about it.
But
133
00:06:36,280 --> 00:06:40,220
I guess, yeah, outsiders,
those areas like where do
134
00:06:40,220 --> 00:06:43,200
you see brands kind of,
yeah, really investing like
135
00:06:43,200 --> 00:06:46,920
time into adopting either
different solutions or
136
00:06:46,920 --> 00:06:49,600
looking into kind of
building their own kind of
137
00:06:49,600 --> 00:06:53,220
solution. Yeah,
yeah, of course. So
138
00:06:53,220 --> 00:06:55,560
I'm seeing a being adopted
more and more. I
139
00:06:55,560 --> 00:06:58,020
feel like on the marketing
side for sure, content
140
00:06:58,020 --> 00:07:02,500
creation also on the video
side. So
141
00:07:02,500 --> 00:07:05,820
we've all seen like how
impressive the release of
142
00:07:05,820 --> 00:07:09,420
Sora is. So
I feel like now you're
143
00:07:09,420 --> 00:07:12,340
you're going to be able to
scale your marketing efforts
144
00:07:12,340 --> 00:07:16,620
10x easier and faster by
leveraging AI for video and
145
00:07:16,620 --> 00:07:20,000
content creation for social
media. Obviously,
146
00:07:20,200 --> 00:07:24,020
Chagipiti helps out with,
you know, like the actual
147
00:07:24,020 --> 00:07:28,720
maybe strips that you're
using for your for your
148
00:07:28,720 --> 00:07:33,420
content or all of the
social media posting, you
149
00:07:33,420 --> 00:07:35,700
know, interactions with your
partners with your
150
00:07:35,700 --> 00:07:38,280
influencers. I
think that's that. And
151
00:07:38,280 --> 00:07:42,300
then another area that I'm
seeing and I think what
152
00:07:42,300 --> 00:07:46,120
for triple, well, I
actually attended their
153
00:07:46,120 --> 00:07:50,540
event two weeks ago in New
York and they launched a
154
00:07:50,540 --> 00:07:53,240
pretty cool tool. I
think it's called Moby and
155
00:07:53,240 --> 00:07:58,860
you can basically just talk
to Moby and ask any kind
156
00:07:58,860 --> 00:08:00,960
of data that you want to
find out. Probably
157
00:08:00,960 --> 00:08:03,360
Moby is going to be able
to build some sort of
158
00:08:03,360 --> 00:08:05,460
reports based on all of
the data that you have
159
00:08:05,460 --> 00:08:08,620
connected in triple. Well,
so it just helps helps
160
00:08:08,620 --> 00:08:11,380
you, you know, just
identify patterns, feedback,
161
00:08:11,780 --> 00:08:14,680
trends to improve your
business. So
162
00:08:14,680 --> 00:08:17,980
I think that's pretty cool.
Hi,
163
00:08:18,240 --> 00:08:21,140
I'm Thomas, Director Growth
at Scale. The
164
00:08:21,140 --> 00:08:23,640
world's fastest growing e
-commerce platform has landed
165
00:08:23,640 --> 00:08:26,260
in the UK with global
brands like Manchester
166
00:08:26,260 --> 00:08:29,860
United, Dietman, Snipes and
Marco Polo trusting us to
167
00:08:29,860 --> 00:08:32,020
scale their commerce
businesses. Are
168
00:08:32,020 --> 00:08:34,120
you ready for a new e
-commerce experience? Talk
169
00:08:34,120 --> 00:08:36,760
to Scale today because life
is just too short for
170
00:08:36,760 --> 00:08:39,840
painful commerce. Yeah,
that makes sense. Yeah,
171
00:08:39,840 --> 00:08:41,580
I did see that and I
thought it was really cool.
172
00:08:41,580 --> 00:08:42,760
I
was quite annoyed actually
173
00:08:42,760 --> 00:08:44,900
when they released that
because my colleague and I
174
00:08:44,900 --> 00:08:49,860
had a similar idea. I
guess it's different and
175
00:08:49,860 --> 00:08:51,140
their data set is much
bigger. But
176
00:08:51,148 --> 00:08:52,980
yeah, I think we'll see a
lot more of that. And
177
00:08:52,980 --> 00:08:56,180
yeah, I've definitely seen
it within some of the
178
00:08:56,180 --> 00:08:58,920
other kind of off the
shelf reporting solutions as
179
00:08:58,920 --> 00:09:01,400
well where you can kind of
be asked queries about data
180
00:09:01,400 --> 00:09:04,800
and that kind of stuff. So
I guess the next question
181
00:09:04,800 --> 00:09:08,580
is, I guess, how is there
any other kind of, I
182
00:09:08,580 --> 00:09:11,000
guess, actually, a better
way of spreading it? Are
183
00:09:11,000 --> 00:09:13,160
there any other solutions
that you are aware of in
184
00:09:13,160 --> 00:09:16,680
the industry outside of CX
that you're particularly
185
00:09:16,680 --> 00:09:20,320
kind of interested in or
you find particularly
186
00:09:20,320 --> 00:09:26,780
impressive? Yes,
I am very impressed by a
187
00:09:26,780 --> 00:09:30,880
tool that I recently
invested in, which is
188
00:09:30,880 --> 00:09:34,600
called Agent Hub. So
Agent Hub is basically the
189
00:09:34,600 --> 00:09:37,420
new shape here. It
permit the easiest way I
190
00:09:37,420 --> 00:09:41,460
can describe it. So
forget about code, then
191
00:09:41,460 --> 00:09:44,640
connecting a bunch of
things, which is, you do
192
00:09:44,640 --> 00:09:48,580
need to be at least have
a level of expertise to
193
00:09:48,580 --> 00:09:51,760
use Zapier, right? So
with Agent Hub, it just
194
00:09:51,760 --> 00:09:56,340
unlocks an entire new world
of possibility and creating
195
00:09:56,340 --> 00:10:01,660
those flows and connecting
a bunch of tools that
196
00:10:01,660 --> 00:10:03,600
you're doing. So
you can use it for sales,
197
00:10:03,940 --> 00:10:09,640
for legal, for accounting,
for marketing, for, I don't
198
00:10:09,640 --> 00:10:12,220
know, a bunch of things,
and they're using Genitive
199
00:10:12,220 --> 00:10:15,880
AI. And
it was very impressive. So
200
00:10:15,888 --> 00:10:19,260
I would definitely recommend
you looking into Agent Hub.
201
00:10:20,800 --> 00:10:23,340
I
think another area where
202
00:10:23,340 --> 00:10:27,400
I'm seeing Brent, so I was
actually talking to one of
203
00:10:27,400 --> 00:10:30,940
our customers bandit
running, and they are about
204
00:10:30,940 --> 00:10:35,360
to implement a tool that
will help them use AI to
205
00:10:35,360 --> 00:10:41,080
identify buying patterns and
then recommend better
206
00:10:41,080 --> 00:10:42,920
products to their customer.
And
207
00:10:42,920 --> 00:10:48,100
I think this is on the UX
side of things. So
208
00:10:48,108 --> 00:10:50,660
it was, yeah, it's an
interesting solution. So
209
00:10:50,660 --> 00:10:55,640
I see it being applied in
a wide range of
210
00:10:55,640 --> 00:10:59,520
departments. But
yeah, Agent Hub is my
211
00:10:59,520 --> 00:11:02,040
favorite. That's
the tool of the year so
212
00:11:02,040 --> 00:11:06,240
far. Excellent.
I'd like to drill in a
213
00:11:06,240 --> 00:11:08,300
bit on the support side,
because obviously you were,
214
00:11:08,420 --> 00:11:10,020
with the solution you
provide, you work with a
215
00:11:10,020 --> 00:11:12,440
lot of customer services
support teams, so you get
216
00:11:12,440 --> 00:11:16,180
to see people's issues,
challenges, and how AI
217
00:11:16,180 --> 00:11:19,660
benefits. How,
what impact does this have
218
00:11:19,660 --> 00:11:22,280
on support staff? Because
obviously some people will
219
00:11:22,280 --> 00:11:24,440
be out there, worry,
there's the classic media
220
00:11:24,440 --> 00:11:28,480
panic story of AI's taking
all your jobs. I'd
221
00:11:28,480 --> 00:11:30,520
love to see, I'd love to
hear from you, like how
222
00:11:30,520 --> 00:11:33,500
does this empower support
teams? What
223
00:11:33,508 --> 00:11:35,840
stuff does it take away
from them that they
224
00:11:35,840 --> 00:11:38,700
typically don't want to do,
and how is it changing
225
00:11:38,700 --> 00:11:43,740
support staff roles? Yeah,
yeah, great question. So
226
00:11:43,740 --> 00:11:47,660
I feel like if we think a
little bit more high level,
227
00:11:47,880 --> 00:11:51,920
there's two main value
propositions and benefits.
228
00:11:52,440 --> 00:11:53,720
So
the first one is on the
229
00:11:53,720 --> 00:11:55,820
economic, so the impact
that it has on the
230
00:11:55,820 --> 00:11:58,660
economic side, so it can
significantly lower
231
00:11:58,660 --> 00:12:02,940
operational costs, you know,
aligning service margins
232
00:12:02,940 --> 00:12:06,200
with those typically seen
in high margin sales
233
00:12:06,200 --> 00:12:09,340
businesses. So
cost saving, and then it's
234
00:12:09,340 --> 00:12:14,080
on the team and customer
side. So
235
00:12:14,080 --> 00:12:17,160
implementing an AI agent
and bringing it into the
236
00:12:17,160 --> 00:12:19,620
team to start working
alongside the other team
237
00:12:19,620 --> 00:12:22,740
members, I feel like it
inevitably leads to higher
238
00:12:22,740 --> 00:12:26,380
satisfaction scores due to
super fast response time
239
00:12:26,380 --> 00:12:32,560
around the club availability
and reliability. So
240
00:12:32,560 --> 00:12:36,120
taking these value
propositions and benefits
241
00:12:36,120 --> 00:12:41,760
into consideration, I feel
like, teams, they're, the
242
00:12:41,760 --> 00:12:45,240
biggest impact for them has
been just reduced stress
243
00:12:45,240 --> 00:12:51,920
overall, burnout, improved
efficiency, being able to
244
00:12:51,920 --> 00:12:56,120
show to their, you know,
boss and their managers,
245
00:12:56,460 --> 00:12:59,260
okay, we can provide a
consistent service quality.
246
00:13:00,060 --> 00:13:02,520
Their
job satisfaction is enhanced
247
00:13:02,520 --> 00:13:05,160
because they do not need
to spend their time, their
248
00:13:05,160 --> 00:13:09,980
precious time on just copy
pasting the same macros
249
00:13:09,980 --> 00:13:12,700
over and over again, and
just being in the inbox
250
00:13:12,700 --> 00:13:15,620
every single day. That's
just not a job that you,
251
00:13:15,680 --> 00:13:18,000
you know, wake up very
excited about, in my
252
00:13:18,000 --> 00:13:20,520
opinion, you know, it's
like, you don't really,
253
00:13:20,720 --> 00:13:23,960
what's the career path if
you're just into, in your
254
00:13:23,960 --> 00:13:26,820
inbox every single day
doing the same thing. So
255
00:13:26,820 --> 00:13:31,400
they, I feel like what AI
brings right now, it's
256
00:13:31,400 --> 00:13:34,260
unlocking a lot of the
opportunities that they
257
00:13:34,260 --> 00:13:37,800
would otherwise not have.
And
258
00:13:37,800 --> 00:13:44,620
also, it's, it's able to
help them just be better
259
00:13:44,620 --> 00:13:47,300
at their job without having
to, like, you know, put
260
00:13:47,300 --> 00:13:49,060
that much effort, right?
Because
261
00:13:49,060 --> 00:13:52,920
when you think, for
example, let's say, you
262
00:13:52,920 --> 00:13:55,320
know, someone is having a
bad day, maybe something
263
00:13:55,320 --> 00:13:58,420
that's personal that
happened to them, but they
264
00:13:58,420 --> 00:14:00,040
have to show up to work.
So
265
00:14:00,040 --> 00:14:03,700
imagine that agent being,
you know, having to
266
00:14:03,700 --> 00:14:07,560
interact with 100 customers
in that day, and all 100
267
00:14:07,560 --> 00:14:10,940
customers just happen to be
frustrated and angry and,
268
00:14:10,940 --> 00:14:15,320
you know, upset. So
imagine how much energy is
269
00:14:15,320 --> 00:14:17,540
taking from that person
who's also going through
270
00:14:17,540 --> 00:14:21,440
some personal stuff. So
I feel like, you know, we
271
00:14:21,440 --> 00:14:27,640
cannot really compete with
an AI on the, like, mood
272
00:14:27,640 --> 00:14:30,140
level and energy levels,
you know, because we are
273
00:14:30,140 --> 00:14:32,240
humans. So
I feel like with AI helps,
274
00:14:32,800 --> 00:14:35,460
helps team is just deliver
the same level of service
275
00:14:35,460 --> 00:14:39,020
consistently, always very
positive, always very nice,
276
00:14:39,160 --> 00:14:42,760
never gets angry, never
gets frustrated, 24 seven
277
00:14:42,760 --> 00:14:44,720
available, never takes
breaks. So
278
00:14:44,720 --> 00:14:48,260
I feel like it just helps
them take that space that
279
00:14:48,260 --> 00:14:51,740
they need. And
then once they, you know,
280
00:14:51,800 --> 00:14:55,000
delegated some of the
tedious manual repetitive
281
00:14:55,000 --> 00:14:58,580
things to the AI, they can
take a back and say, okay,
282
00:14:58,680 --> 00:15:00,360
like, what should I do?
Where,
283
00:15:00,360 --> 00:15:03,940
where, where would my
skills have the biggest
284
00:15:03,940 --> 00:15:05,700
impact in this company?
Maybe
285
00:15:05,708 --> 00:15:08,400
I can, you know, get more
involved on the product
286
00:15:08,400 --> 00:15:10,560
side. Maybe
I can get more involved on
287
00:15:10,560 --> 00:15:13,180
the loyalty and retention
side. Or
288
00:15:13,180 --> 00:15:16,340
maybe I can really, you
know, become a manager and
289
00:15:16,340 --> 00:15:19,260
manage these AI tools that
we're bringing to the team.
290
00:15:19,300 --> 00:15:21,900
So
I feel like it just brings
291
00:15:21,900 --> 00:15:25,560
a lot of opportunity more,
it opens more doors that
292
00:15:25,560 --> 00:15:27,820
it closes in my opinion.
And
293
00:15:27,820 --> 00:15:33,300
yeah, it's, it has, it has
a lot of opportunities for
294
00:15:33,300 --> 00:15:35,380
teams. I
feel like, yes, there is,
295
00:15:36,200 --> 00:15:38,500
I feel like the adoption
curve is still, is still
296
00:15:38,500 --> 00:15:42,100
slow. You
do need to be looking at
297
00:15:42,100 --> 00:15:44,860
it with an open mind and
embrace it in order for
298
00:15:44,860 --> 00:15:47,500
you to get the full
benefits and just take
299
00:15:47,500 --> 00:15:52,060
advantage of this new wave
of innovation that's, you
300
00:15:52,060 --> 00:15:55,000
know, here. I
think it's an interesting
301
00:15:55,000 --> 00:15:59,100
point you might catch about
the, the mood consistency,
302
00:15:59,100 --> 00:16:01,340
because people do, we're
all human, we have off
303
00:16:01,340 --> 00:16:04,040
days, we have days when we
cannot be asked to deal
304
00:16:04,040 --> 00:16:07,600
with that complaint. All
that angry power, whatever
305
00:16:07,600 --> 00:16:10,660
it might be, and being
able to take the emotion
306
00:16:10,660 --> 00:16:13,200
out of a response and have
consistency. Yeah,
307
00:16:13,260 --> 00:16:15,580
I hadn't, I hadn't focused
on that point, but that's
308
00:16:15,580 --> 00:16:18,020
a really important one.
Yeah,
309
00:16:18,060 --> 00:16:20,880
and I feel like empathic
AI in particular really
310
00:16:20,880 --> 00:16:24,480
cuts through the biases and
the emotional swings that
311
00:16:24,480 --> 00:16:28,040
often just fog up for
human judgment when we,
312
00:16:28,260 --> 00:16:30,080
when we engage with people.
I
313
00:16:30,080 --> 00:16:32,780
mean, you know, I would
love to have a Sienna. I
314
00:16:32,788 --> 00:16:35,940
mean, I'm not in the best
mood every single day, even
315
00:16:35,940 --> 00:16:38,380
on Slack. So,
you know, just having a
316
00:16:38,380 --> 00:16:41,620
Sienna for myself, which
we're actually doing that.
317
00:16:41,740 --> 00:16:43,540
But
yeah, it would be, it
318
00:16:43,540 --> 00:16:47,180
would be fun. I've
got a kind of like
319
00:16:47,180 --> 00:16:49,280
tangential question related
to that. So
320
00:16:49,280 --> 00:16:52,740
there was a, there was a
high profile case of a
321
00:16:52,740 --> 00:16:55,140
chat board that got
hijacked by people messing
322
00:16:55,140 --> 00:16:57,720
around with it in the UK,
because whoever had set up
323
00:16:57,720 --> 00:16:59,760
obviously hadn't put the
right safeguards in place.
324
00:17:00,540 --> 00:17:03,760
It
was a carrier and somebody
325
00:17:03,760 --> 00:17:07,040
had basically tricked it
into swearing, tricked it
326
00:17:07,040 --> 00:17:10,780
into calling it, calling
the company shit, etc. So
327
00:17:10,780 --> 00:17:12,660
it was very entertaining.
But
328
00:17:12,660 --> 00:17:15,680
what, okay, it would be
really great to, it's
329
00:17:15,680 --> 00:17:18,780
specifically in the context
of Sienna, how do people
330
00:17:18,780 --> 00:17:21,000
get grants and safeguarding
it? How
331
00:17:21,000 --> 00:17:24,540
do people stop the chatbot
from saying bad things or
332
00:17:24,540 --> 00:17:26,640
using inappropriate language?
At
333
00:17:26,640 --> 00:17:29,080
what level of control and
how do you support people
334
00:17:29,080 --> 00:17:34,020
to get that right? Yeah,
yeah, for sure. So
335
00:17:34,020 --> 00:17:38,580
I think one of the best
parts about Sienna is that
336
00:17:38,580 --> 00:17:42,160
you are in full control
over what she responds to
337
00:17:42,160 --> 00:17:45,580
versus what she doesn't
respond to, what types of,
338
00:17:45,580 --> 00:17:48,800
you know, use cases she
handles and what doesn't,
339
00:17:48,860 --> 00:17:52,560
and channels as well. So
there's a lot of guard
340
00:17:52,560 --> 00:17:56,680
rail, guard trails embedded
into the model that we're,
341
00:17:57,100 --> 00:18:00,820
they're just there. But
there is, there's also some
342
00:18:00,820 --> 00:18:03,640
features that they can
leverage like our content
343
00:18:03,640 --> 00:18:08,240
moderation feature that
really tells Sienna, okay,
344
00:18:08,300 --> 00:18:12,620
if you see these keywords
or held this tags or users
345
00:18:12,620 --> 00:18:15,780
or domains, then you
completely ignore those
346
00:18:15,780 --> 00:18:17,920
conversations and those
customers. So
347
00:18:17,928 --> 00:18:20,040
you just do not touch
those tickets. So
348
00:18:20,040 --> 00:18:22,660
this is just one example.
Another
349
00:18:22,660 --> 00:18:28,160
one is during the
implementation phase, we
350
00:18:28,160 --> 00:18:30,740
take a very gradual
approach. So
351
00:18:30,740 --> 00:18:35,000
we first in the first week
or two, we start going
352
00:18:35,000 --> 00:18:38,940
live with CNN, but we only
go live with drafted
353
00:18:38,940 --> 00:18:41,780
responses, we call them
internal notes. So
354
00:18:41,788 --> 00:18:44,520
basically, they will see
what responses Sienna
355
00:18:44,520 --> 00:18:46,740
provide, the agents will
see them in their help
356
00:18:46,740 --> 00:18:49,620
desk, but they're not going
live to the customer. So
357
00:18:49,620 --> 00:18:53,000
it's an extra step of, you
know, confidence that they
358
00:18:53,000 --> 00:18:55,660
can build and just make
sure, okay, this is a
359
00:18:55,660 --> 00:18:58,340
response that meets our
expectation. And
360
00:18:58,340 --> 00:19:00,940
then we can even copy
paste and use it as the
361
00:19:00,940 --> 00:19:05,960
best, like I feel
validation for us when
362
00:19:05,960 --> 00:19:08,480
agents are actually starting
to copy paste and use
363
00:19:08,480 --> 00:19:10,860
Sienna's responses, then
okay, you're ready to go
364
00:19:10,860 --> 00:19:13,500
live. That's
like the best, you know,
365
00:19:13,540 --> 00:19:16,080
signal. And
also, he tells you, okay,
366
00:19:16,140 --> 00:19:18,980
if this response is not
the way I want it to be,
367
00:19:19,020 --> 00:19:21,960
let's go back and do some
fight tuning, maybe give
368
00:19:21,960 --> 00:19:25,920
CNN some more information,
add more data, provide more
369
00:19:25,920 --> 00:19:28,540
context or add another
instruction. So
370
00:19:28,540 --> 00:19:31,040
it's a really good step
that builds their trust
371
00:19:31,040 --> 00:19:34,780
before going fully
autonomous. And
372
00:19:34,780 --> 00:19:37,020
there's a bunch of other
things that you can set
373
00:19:37,020 --> 00:19:38,440
up. So
it's highly customizable.
374
00:19:39,540 --> 00:19:42,180
That's
one of the best parts. And
00:19:44,440
it really ensures that it
376
00:19:44,440 --> 00:19:47,720
just sticks within what you
train it. Also
377
00:19:47,720 --> 00:19:50,320
another cool thing is it
works all even real time
378
00:19:50,320 --> 00:19:54,180
data. So
right now, you will import,
379
00:19:54,320 --> 00:19:56,580
you can import documents,
you can sing your Shopify
380
00:19:56,580 --> 00:19:59,340
product catalog, you can
sing your website. And
381
00:19:59,340 --> 00:20:02,900
then you can also build
automations, like building
382
00:20:02,900 --> 00:20:04,860
processes. So
for example, process for
383
00:20:04,860 --> 00:20:07,360
returns, you just tell CNN
how returns should be
384
00:20:07,360 --> 00:20:10,440
managed a process for order
status, a process for
385
00:20:10,440 --> 00:20:13,900
changing in, you know,
subscription edits and all
386
00:20:13,900 --> 00:20:16,200
of that. So,
you know, you train it and
387
00:20:16,200 --> 00:20:20,540
that's what she will
follow, nothing else. Great.
00:20:22,100
And I guess on top of
389
00:20:22,100 --> 00:20:24,180
that, there's some really
good use cases there. But
390
00:20:24,180 --> 00:20:27,200
can you just talk us
through any like further
391
00:20:27,200 --> 00:20:31,280
really interesting kind of
applications of Sienna that
392
00:20:31,280 --> 00:20:34,580
some of your customers or
clients have kind of
393
00:20:34,580 --> 00:20:38,140
implemented where they've
made use of the chat bar
394
00:20:38,140 --> 00:20:41,600
a little bit more kind of
engaging. So
395
00:20:41,600 --> 00:20:45,640
this perspective. Yeah.
Yeah, I feel like one of
396
00:20:45,640 --> 00:20:50,760
the biggest goals that I
see brands have when they
397
00:20:50,760 --> 00:20:55,280
join Sienna is they say we
want to enable live chat
398
00:20:55,280 --> 00:20:58,240
and we want to have 20 %
support, but we haven't
399
00:20:58,240 --> 00:21:00,520
been able to do it so far
because we just don't have
400
00:21:00,520 --> 00:21:03,639
the bandit and resources in
house. So
401
00:21:03,639 --> 00:21:06,380
I feel like the chat
application is really,
402
00:21:06,380 --> 00:21:09,000
really cool. And
we're actually working with
403
00:21:09,000 --> 00:21:12,940
two beauty brands who have
such a cool use case. So
404
00:21:12,940 --> 00:21:15,080
basically what they want to
do is they want to help
405
00:21:15,080 --> 00:21:20,880
their customers find the
best product and match the
406
00:21:20,880 --> 00:21:24,260
shade, you know, of the
product that based on their
407
00:21:24,260 --> 00:21:27,480
skin type. So
we have, we're working with
408
00:21:27,480 --> 00:21:31,680
this brand. And
every time someone asks,
409
00:21:31,800 --> 00:21:33,280
Hey, what do you recommend?
You
410
00:21:33,288 --> 00:21:35,780
know, my skin type is very
pale. And
411
00:21:35,780 --> 00:21:37,920
I would say, Hey, like,
can you send us a picture?
412
00:21:38,280 --> 00:21:40,460
I
want to see your skin tone
413
00:21:40,460 --> 00:21:43,480
and type. So
they would send a selfie.
414
00:21:43,920 --> 00:21:45,840
Sienna
has image recognition. It's
415
00:21:45,840 --> 00:21:50,500
called Sienna vision. And
she was trained on what to
416
00:21:50,500 --> 00:21:52,580
look for in that picture.
And
417
00:21:52,588 --> 00:21:55,500
basically she was trained
that, Okay, if, you know,
418
00:21:55,540 --> 00:21:57,820
their skin type is very
pale, you should recommend
419
00:21:57,820 --> 00:21:59,540
this shade. If
it's a little bit more
420
00:21:59,540 --> 00:22:02,260
darker, you should recommend
this straight. So
421
00:22:02,260 --> 00:22:04,760
that's a super cool
approach. And
422
00:22:04,760 --> 00:22:07,580
basically Sienna now can
really provide broader
423
00:22:07,580 --> 00:22:13,800
recommendations and, you
know, really take care of
424
00:22:13,800 --> 00:22:16,820
this entire use case on
her own without necessarily
425
00:22:16,820 --> 00:22:20,000
having like a human needed
to get involved. So
426
00:22:20,000 --> 00:22:23,520
it's really cool. And
that leads obviously to
427
00:22:23,520 --> 00:22:27,200
revenue, right? To
more purchases. Yeah,
428
00:22:27,240 --> 00:22:29,020
great. And
I think, I mean, this is
429
00:22:29,020 --> 00:22:31,440
the, I think a really
interesting question. I'm
430
00:22:31,440 --> 00:22:33,500
sure you get all the time.
But
431
00:22:33,500 --> 00:22:37,900
some of the kind of CX
solutions and live chat
432
00:22:37,900 --> 00:22:41,660
vendors are releasing their
own kind of AI offerings,
433
00:22:42,020 --> 00:22:45,340
I guess. So
what's your view on, I
434
00:22:45,340 --> 00:22:50,340
guess, one, how Sienna will
differentiate against those,
435
00:22:50,580 --> 00:22:53,680
but also like how well
placed they are to kind of
436
00:22:53,680 --> 00:22:56,020
build this type of stuff?
Yeah,
437
00:22:56,180 --> 00:22:59,920
of course, of course. So
I think when you look at
438
00:22:59,920 --> 00:23:04,860
help desks that are adding
some sort of AI on top of
439
00:23:04,860 --> 00:23:08,220
what they already have, you
know, their main offering,
440
00:23:08,220 --> 00:23:11,420
their main product is being
a ticketing system. That's
441
00:23:11,420 --> 00:23:13,400
what they do best, right?
So
442
00:23:13,400 --> 00:23:16,640
they have their very like,
solid foundation
443
00:23:16,640 --> 00:23:19,680
infrastructure of being a
ticketing system that's, you
444
00:23:19,680 --> 00:23:22,700
know, like built years ago.
So
445
00:23:22,700 --> 00:23:25,760
a lot of these are adding
some sort of AI
446
00:23:25,760 --> 00:23:28,620
capabilities, whether they
call it right now, you
447
00:23:28,620 --> 00:23:33,940
know, like a AI agent or,
you know, AI assistant,
448
00:23:34,160 --> 00:23:38,460
whatever that might be. But
it's very different being
449
00:23:38,460 --> 00:23:43,900
like having like a
foundation that's completely
450
00:23:43,900 --> 00:23:47,700
different not powered by
Generative AI versus a
451
00:23:47,700 --> 00:23:51,840
fully powered AI agent that
was built on top of large
452
00:23:51,840 --> 00:23:54,440
language models from the
ground up. And
453
00:23:54,440 --> 00:23:57,960
the main difference I would
say between these two is
454
00:23:57,960 --> 00:24:01,680
the fact that when you
look at Sienna, she was
455
00:24:01,680 --> 00:24:04,400
really designed and trained
like a human. So
456
00:24:04,400 --> 00:24:07,340
what we did before we
actually even started to
457
00:24:07,340 --> 00:24:10,200
write a line of code, we
always try to look, okay,
458
00:24:10,240 --> 00:24:13,240
how do the human processes
look today? How
459
00:24:13,240 --> 00:24:16,500
does a human interact with
the customer today? What
460
00:24:16,500 --> 00:24:18,460
are the steps that the
human takes? And
461
00:24:18,460 --> 00:24:21,200
we worked backwards from
the human experience to
462
00:24:21,200 --> 00:24:23,900
actually building this AI
agent and how we should
463
00:24:23,900 --> 00:24:29,840
just like navigate. So
there's multiple like, you
464
00:24:29,840 --> 00:24:32,540
know, agent frameworks that
we're using an agent that
465
00:24:32,540 --> 00:24:36,560
does, you know, order,
order management, an agent
466
00:24:36,560 --> 00:24:38,900
that does subscription and
agent that, you know,
467
00:24:39,180 --> 00:24:43,220
provides like a super nice
response, because it's very
468
00:24:43,220 --> 00:24:45,820
empathic and personalized.
And
469
00:24:45,820 --> 00:24:49,720
then it really applies that
three main principles in
470
00:24:49,720 --> 00:24:53,280
customer communication, which
is comprehension, reasoning,
471
00:24:53,440 --> 00:24:56,160
and decision making. So
first, it understands
472
00:24:56,160 --> 00:24:58,880
exactly what the customer
wants and needs, whether
473
00:24:58,880 --> 00:25:02,900
they are sending us like
just one message with one
474
00:25:02,900 --> 00:25:05,480
question or like a huge
paragraph with a bunch of
475
00:25:05,480 --> 00:25:08,800
questions, it's able to
understand exactly, you
476
00:25:08,800 --> 00:25:13,800
know, really go beyond the
lines. Second,
477
00:25:14,100 --> 00:25:18,080
it's able to reason. So
it's really able to
478
00:25:18,080 --> 00:25:21,020
understand, okay, so I have
this data point, the
479
00:25:21,020 --> 00:25:23,300
customer said that I need
to go into shop, I need
480
00:25:23,300 --> 00:25:26,180
to go here and just like,
okay, I will make sure
481
00:25:26,180 --> 00:25:28,260
this is the best course of
action, and then we'll
482
00:25:28,260 --> 00:25:30,980
actually take that action,
which is the decision,
483
00:25:31,060 --> 00:25:35,240
which can be either just a
response or it might
484
00:25:35,240 --> 00:25:38,460
involve, you know, doing
some changes in another
485
00:25:38,460 --> 00:25:42,140
platform. So
I think I feel like these
486
00:25:42,140 --> 00:25:45,020
three principles are what
makes the difference. And
487
00:25:45,028 --> 00:25:50,200
most chatbots, AI chatbots
that are advertised today,
488
00:25:50,380 --> 00:25:53,360
unfortunately, they're not
built on top of generative
489
00:25:53,360 --> 00:25:56,680
AI, and they're non
-generative AI chatbots,
490
00:25:57,220 --> 00:26:02,460
they're decision, they're
behind, they're, you know,
491
00:26:03,720 --> 00:26:08,280
nice UX or UI, they're
basically built on top of
492
00:26:08,280 --> 00:26:10,300
decision trees. So,
you know, if this do that,
493
00:26:10,360 --> 00:26:14,060
so it's rule based, rule
based engines that are
494
00:26:14,060 --> 00:26:17,680
powered by keywords, and
then if you ask something
495
00:26:17,680 --> 00:26:20,940
that's not built out, the
customer, basically the
496
00:26:20,940 --> 00:26:23,780
customer will not get a
good response, just will be
497
00:26:23,780 --> 00:26:28,640
lost. Yeah,
that's interesting. I
498
00:26:28,640 --> 00:26:31,280
had a question around the
differentiation from between
499
00:26:31,280 --> 00:26:33,680
AI and just normal
chatbots. I'd
500
00:26:33,680 --> 00:26:35,960
be useful to just probe a
bit more on that because
501
00:26:35,960 --> 00:26:40,000
they're obviously you, your
product has the ability to,
502
00:26:40,120 --> 00:26:42,340
like, preprogram, have
specific commands that can
503
00:26:42,340 --> 00:26:45,980
follow specific paths versus
having that AI capability
504
00:26:45,980 --> 00:26:50,040
to discuss on a more human
level. What
505
00:26:50,040 --> 00:26:53,260
are the, what are some of
the most productive
506
00:26:53,260 --> 00:26:56,960
preprogram paths that your
customers are using within
507
00:26:56,960 --> 00:27:00,320
the overall CNN experience
to get people quickly into
508
00:27:00,320 --> 00:27:03,460
routes, whether that's like
tracking order, statuses,
509
00:27:03,460 --> 00:27:05,020
etc. Love
you to share with
510
00:27:05,020 --> 00:27:07,520
customers, with everyone
listening, like what are
511
00:27:07,520 --> 00:27:09,700
the essentials, what should
people be doing, and what
512
00:27:09,700 --> 00:27:14,400
can you do? Yeah,
yeah. So,
513
00:27:14,440 --> 00:27:20,200
I feel like the essentials
should be always based on
514
00:27:20,200 --> 00:27:25,580
what is the highest
percentage of tickets that
515
00:27:25,580 --> 00:27:28,260
you're getting on a
specific topic. So,
516
00:27:28,280 --> 00:27:31,660
if you look in your help
desk and you see that 60 %
517
00:27:31,660 --> 00:27:34,800
of our tickets are about
where is my order, then
518
00:27:34,800 --> 00:27:37,660
you should start automating
where is my order. And
519
00:27:37,660 --> 00:27:39,800
that literally you can do
it in like two clicks
520
00:27:39,800 --> 00:27:42,880
inside of CNN because all
of these modules with
521
00:27:42,880 --> 00:27:45,720
Shopify would recharge with
all of our integrations
522
00:27:45,720 --> 00:27:47,920
partners were built in the
backend by us. So,
523
00:27:47,980 --> 00:27:50,860
you just need to literally
connect Shopify and then
524
00:27:50,860 --> 00:27:53,720
select get order status and
then CNN will know what to
525
00:27:53,720 --> 00:27:57,520
do. If,
you know, I don't know, 70
526
00:27:57,520 --> 00:28:00,280
% of your inquiries that
are about returns, then we
527
00:28:00,280 --> 00:28:02,600
should do something about
returns, right? Maybe
528
00:28:02,600 --> 00:28:04,280
we want to mitigate
returns. So,
529
00:28:04,300 --> 00:28:07,840
we're trying to use CNN to
offer them an alternative
530
00:28:07,840 --> 00:28:11,180
to returning, or maybe we
actually want to automate
531
00:28:11,180 --> 00:28:14,000
and have like a self -serve
return experience versus
532
00:28:14,000 --> 00:28:17,940
having our agents manually
deal with returns. So,
533
00:28:18,220 --> 00:28:21,860
I feel like it's first
look at, you know, the
534
00:28:21,860 --> 00:28:24,800
percentages and what are
those use cases that you're
535
00:28:24,800 --> 00:28:27,420
seeing over and over in
your ticketing system and
536
00:28:27,420 --> 00:28:30,880
start automating those
because you'll see like a
537
00:28:30,880 --> 00:28:38,000
huge drop in your tickets
that need to be handled by
538
00:28:38,000 --> 00:28:42,180
humans and then you'll see
like a being like a super
539
00:28:42,180 --> 00:28:45,400
fast when in the automation
rate with CNN. So,
540
00:28:45,420 --> 00:28:48,120
I think that's how our
customers typically start.
541
00:28:48,360 --> 00:28:49,400
Just
like what are the most
542
00:28:49,400 --> 00:28:52,440
pressing things that we
need to tackle on the
543
00:28:52,440 --> 00:28:58,860
lowest hanging fruit? Yeah,
I've been working with a
544
00:28:58,860 --> 00:29:02,360
customer recently and the
biggest issue they have is
545
00:29:02,360 --> 00:29:05,300
people keep asking for
tracking information because
546
00:29:05,300 --> 00:29:07,320
they're not very good at
surfacing it and actually
547
00:29:07,320 --> 00:29:10,620
it's quicker randomly
quicker for them to get
548
00:29:10,620 --> 00:29:13,280
that info plugged into a
very, they've got a very
549
00:29:13,280 --> 00:29:15,420
basic chatbot for a POC.
It's
550
00:29:15,420 --> 00:29:17,520
quicker to push that in
from their WMS and it was
551
00:29:17,520 --> 00:29:19,800
to get integrated with
their e -commerce platform
552
00:29:19,800 --> 00:29:23,060
because they've got quite a
custom theme. So,
553
00:29:23,140 --> 00:29:26,860
I'm not surprised. The
question that I've heard a
554
00:29:26,860 --> 00:29:30,260
few people mention is, I
stopped a few people going
555
00:29:30,260 --> 00:29:33,440
live as well or trying is
how do you track
556
00:29:33,440 --> 00:29:35,380
performance? How
do you validate it's having
557
00:29:35,380 --> 00:29:39,140
an impact other than
reducing tickets? So,
558
00:29:39,200 --> 00:29:42,920
how do you measure the
impact of CNN and how do
559
00:29:42,920 --> 00:29:44,940
you work with customers to
help them show what the
560
00:29:44,940 --> 00:29:47,420
real impact is rather than
just okay, we've saved X
561
00:29:47,420 --> 00:29:51,820
thousand tickets. Yeah,
great question. So,
562
00:29:51,880 --> 00:29:57,820
I feel like what's really
cool about CNN is the fact
563
00:29:57,820 --> 00:30:00,860
that she's literally an
agent, they takes a user's
564
00:30:00,860 --> 00:30:03,800
instead of your ticketing
system. So,
565
00:30:03,840 --> 00:30:07,020
you can track her
performance exactly how you
566
00:30:07,020 --> 00:30:08,800
track it for your team.
So,
567
00:30:08,840 --> 00:30:12,860
in your help test, you can
go and see, you know, CNN
568
00:30:12,860 --> 00:30:15,580
as far as response time,
CNN as resolution time, the
569
00:30:15,580 --> 00:30:19,140
ticket conversion rate, the
CSAT score of CNN, the
570
00:30:19,140 --> 00:30:22,180
reviews that CNN received
after interacting with
571
00:30:22,180 --> 00:30:24,460
customers. So,
that's on one side, you
572
00:30:24,460 --> 00:30:27,700
can see her as like a
real agent, how she's
573
00:30:27,700 --> 00:30:30,000
performing. And
the other side, when you
574
00:30:30,000 --> 00:30:32,640
want to look at like,
okay, how is the AI agent
575
00:30:32,640 --> 00:30:35,360
and the automation
performing, you go into
576
00:30:35,360 --> 00:30:39,380
CNN, you see the automation
rate, you see the time
577
00:30:39,380 --> 00:30:41,960
saved. And
another very cool metric
578
00:30:41,960 --> 00:30:46,140
that we recently introduced,
it's the CXP. So,
579
00:30:46,180 --> 00:30:48,600
what this means is the
customer experience score.
580
00:30:49,220 --> 00:30:51,260
So,
the customer experience
581
00:30:51,260 --> 00:30:54,700
score is like a CSAT on
steroids, and it really
582
00:30:54,700 --> 00:30:58,940
tells you, okay, is my AI
agent actually performing?
583
00:30:59,260 --> 00:31:00,900
Is
it solving problems? Is
584
00:31:00,900 --> 00:31:03,840
are the tickets being
closed with happy customers?
585
00:31:04,040 --> 00:31:05,600
Or
am I just seeing a very
586
00:31:05,600 --> 00:31:08,780
high automation rate that
is just doesn't, you know,
587
00:31:08,840 --> 00:31:11,080
tell me anything, because
you can have 90 %
588
00:31:11,080 --> 00:31:14,180
automation rate if the
responses are shitty and
589
00:31:14,180 --> 00:31:16,780
your customers are pissed.
So,
590
00:31:16,820 --> 00:31:20,040
we introduced the CXP. So,
now, every time you look
591
00:31:20,040 --> 00:31:21,980
at the automation rate, you
need to correlate it with
592
00:31:21,980 --> 00:31:24,340
the CXP. So,
what does this mean? It
593
00:31:24,340 --> 00:31:29,160
takes into consideration six
data points, the CSAT
594
00:31:29,160 --> 00:31:33,540
scores, effort index
resolution rate, the
595
00:31:33,540 --> 00:31:40,720
empathy, and then the also
the, I think it was the
596
00:31:40,720 --> 00:31:45,600
emotional aspect of the
conversation. So,
597
00:31:45,740 --> 00:31:48,600
what's the sentiment? So,
by having all of these
598
00:31:48,600 --> 00:31:52,200
metrics, you can rest for,
okay, like actually CNN did
599
00:31:52,200 --> 00:31:54,880
a really good job here.
And
600
00:31:54,880 --> 00:32:00,400
we typically say, if you
have a CXP above 90%, then
601
00:32:00,400 --> 00:32:03,180
it means you have a very
good strategy, like really,
602
00:32:03,180 --> 00:32:05,740
really good agent or
former. But
603
00:32:05,740 --> 00:32:09,060
then if it's, you know,
between 70 and 90, then
604
00:32:09,060 --> 00:32:12,460
okay, let's look at what
type of intents are
605
00:32:12,460 --> 00:32:17,020
actually the ones that are
not that high in the CXP
606
00:32:17,020 --> 00:32:18,520
score. And
let's double down, like
607
00:32:18,520 --> 00:32:22,040
optimize fine tune those
automations. And
608
00:32:22,040 --> 00:32:24,360
then if it's below, okay,
it's a red flag, we really
609
00:32:24,360 --> 00:32:27,140
need to dive deeper into
those and see what's going
610
00:32:27,140 --> 00:32:32,380
on, like what's, you know,
blocker in, you know, in
611
00:32:32,380 --> 00:32:35,560
us getting such a low
score. Also,
612
00:32:35,660 --> 00:32:38,400
another very important thing
that we are tracking
613
00:32:38,400 --> 00:32:41,440
internally. So,
this is our North Star
614
00:32:41,440 --> 00:32:44,160
metric with our customer's
existing. Every
615
00:32:44,160 --> 00:32:47,460
time we sync with a
customer, we ask them on a
616
00:32:47,460 --> 00:32:51,920
scale from one to 10, if
CNN where you are an
617
00:32:51,920 --> 00:32:54,880
agent, like a real agent,
how would you rate it?
618
00:32:55,300 --> 00:32:57,480
One,
CNN is useless, and 10,
619
00:32:57,620 --> 00:33:00,420
CNN is the best performer
we've ever had. So
620
00:33:00,420 --> 00:33:04,040
that also tells us exactly,
okay, if it's around seven
621
00:33:04,040 --> 00:33:06,960
or eight, how can we make
it a 10? And
622
00:33:06,960 --> 00:33:10,500
we look exactly at those
areas that need improvement.
623
00:33:13,180 --> 00:33:14,880
Great.
And then I guess last
624
00:33:14,880 --> 00:33:17,380
couple of questions. And
I think this one's
625
00:33:17,380 --> 00:33:21,060
particularly interesting for
me is if you're working
626
00:33:21,060 --> 00:33:25,460
with like a large retailer,
how would your average
627
00:33:25,460 --> 00:33:28,140
client approach a pre for
concert? So
628
00:33:28,140 --> 00:33:31,000
that what's the typical
process for kind of
629
00:33:31,000 --> 00:33:34,040
trialing this, and maybe
presenting different options
630
00:33:34,040 --> 00:33:36,420
to the customer, that kind
of stuff? Yeah,
631
00:33:36,720 --> 00:33:39,460
yeah. So
what we've been doing with
632
00:33:39,460 --> 00:33:44,360
enterprise, enterprise
customers, or businesses is
633
00:33:44,360 --> 00:33:48,720
we were putting together
some documentation of like,
634
00:33:48,840 --> 00:33:53,220
okay, what is like the POC
documentation, like what are
635
00:33:53,220 --> 00:33:55,620
our goals? What
are the use cases that
636
00:33:55,620 --> 00:33:57,820
you'd like to test to see
if this would work for
637
00:33:57,820 --> 00:34:00,540
your business? What
are the you know, some
638
00:34:00,540 --> 00:34:02,900
like first metrics that
you'd see that you'd like
639
00:34:02,900 --> 00:34:07,680
to see being impacted? And
we obviously work backwards
640
00:34:07,680 --> 00:34:11,820
from their goals and their
use cases, and we'll just
641
00:34:11,820 --> 00:34:14,860
put together a testing
account for them. And
642
00:34:14,860 --> 00:34:17,860
we keep a very close
collaboration. We
643
00:34:17,860 --> 00:34:21,100
go through, you know, those
use cases with Trains C
644
00:34:21,100 --> 00:34:24,179
and we choose like between
five to 10, not more. We
645
00:34:24,179 --> 00:34:27,679
want to make sure we stay
very focused because, you
646
00:34:27,679 --> 00:34:33,440
know, like doing testing,
obviously, it's just, it's
647
00:34:33,440 --> 00:34:37,080
more scrappy, more
superficial is just for you
648
00:34:37,080 --> 00:34:40,260
to gauge the potential and
having this kind of AI
649
00:34:40,260 --> 00:34:43,020
agent into your team, into
your operations. I
650
00:34:43,020 --> 00:34:46,600
think what I see teams
being very interested in is
651
00:34:46,600 --> 00:34:48,780
like, how does the handoff
process look like? Okay,
652
00:34:48,840 --> 00:34:51,600
what happens when Sienna
maybe doesn't have enough
653
00:34:51,600 --> 00:34:53,820
knowledge or data to
respond? Like
654
00:34:53,820 --> 00:34:56,199
how does it go to the
team? How
655
00:34:56,199 --> 00:34:58,940
does you know, the is it
seamless? Do
656
00:34:58,940 --> 00:35:00,500
we need to do something?
So
657
00:35:00,500 --> 00:35:02,800
I think that's what they
want to see, like that
658
00:35:02,800 --> 00:35:05,300
experience working with
their teams. And
659
00:35:05,300 --> 00:35:09,120
then just like, I feel
like they're very, very,
660
00:35:09,940 --> 00:35:14,200
particular about the brand
side of things. So
661
00:35:14,200 --> 00:35:18,040
how can we make sure that
we, you know, automate, but
662
00:35:18,040 --> 00:35:21,200
we still keep our brand
voice tone and personality,
663
00:35:21,520 --> 00:35:24,220
like, you know, imagine you
spend years and years
664
00:35:24,220 --> 00:35:27,380
building that you do not
want to just use an AI
665
00:35:27,380 --> 00:35:30,240
tool that will ruin
everything. So
666
00:35:30,240 --> 00:35:34,320
they look very closely to
the quality of responses
667
00:35:34,320 --> 00:35:37,460
and the engagements and
just seeing, okay, will
668
00:35:37,460 --> 00:35:40,700
this actually fit? Like
how would this agent and
669
00:35:40,700 --> 00:35:43,840
the way it interacts will
integrate into everything
670
00:35:43,840 --> 00:35:46,660
that we stand for and into
our business and the way
671
00:35:46,660 --> 00:35:49,680
we communicate with our
customers? So
672
00:35:49,680 --> 00:35:52,980
yeah, it's like, you know,
to sum it up, it's very
673
00:35:52,980 --> 00:35:58,600
focused on, you know,
having some quick wins and
674
00:35:58,600 --> 00:36:01,300
just gauging the potential
based on their use cases.
675
00:36:01,760 --> 00:36:04,160
And
yeah. Great.
676
00:36:04,500 --> 00:36:06,200
And I guess last question.
So
677
00:36:06,200 --> 00:36:10,920
I was recently chatting to
my mobile phone for Budda
678
00:36:10,920 --> 00:36:13,860
and they presented two
options. So
679
00:36:13,860 --> 00:36:16,940
they're basically like, you
can chat with a real human
680
00:36:16,940 --> 00:36:19,840
agent, but it's going to
take 30 minutes before
681
00:36:19,840 --> 00:36:22,760
they're available. Or
you can chat with our AI
682
00:36:22,760 --> 00:36:24,940
agent and they can
essentially process
683
00:36:24,940 --> 00:36:27,720
information based on, you
know, their language model,
684
00:36:27,800 --> 00:36:31,560
etc. So
they obviously disclose the
685
00:36:31,560 --> 00:36:34,600
usage of AI. A
lot of brands don't. So
686
00:36:34,600 --> 00:36:36,340
what's your view on that?
Like,
687
00:36:36,440 --> 00:36:38,700
do you think your best
kind of presenting options,
688
00:36:39,040 --> 00:36:41,000
do you think your best
kind of saying that AI is
689
00:36:41,000 --> 00:36:43,300
partially used? Or
do you think it's a case
690
00:36:43,300 --> 00:36:45,800
of, yeah, using it in a
different way so you don't
691
00:36:45,800 --> 00:36:49,920
necessarily need to tell
the user that is the case?
692
00:36:50,700 --> 00:36:54,140
Yeah.
So I think there's no
693
00:36:54,140 --> 00:36:56,660
right or wrong answer, to
be honest. It
694
00:36:56,660 --> 00:37:01,480
all comes down to the
business preferences. So
695
00:37:01,480 --> 00:37:05,840
the majority of our
customers, they do not
696
00:37:05,840 --> 00:37:08,040
disclose it. For
example, if they have like
697
00:37:08,040 --> 00:37:11,440
a live chat or on email
or on social media, most
698
00:37:11,440 --> 00:37:14,780
of them, they, you know,
create CNN as a real
699
00:37:14,780 --> 00:37:17,860
person, they give it a
real name, your real
700
00:37:17,860 --> 00:37:21,800
attributes and really give,
give life to CNN in a way
701
00:37:21,800 --> 00:37:24,280
that you have no idea that
you're talking to an AI.
702
00:37:24,980 --> 00:37:27,040
And
most like 99 % of customers
703
00:37:27,040 --> 00:37:29,240
have no idea they're
actually talking to CNN.
704
00:37:30,100 --> 00:37:33,600
But
also I feel like there's
705
00:37:33,600 --> 00:37:36,660
customers who just have
like transparency as a core
706
00:37:36,660 --> 00:37:40,000
value and they might want
to add it in their, you
707
00:37:40,000 --> 00:37:42,600
know, privacy policy or
terms and conditions or
708
00:37:42,600 --> 00:37:45,380
just like as a disclaimer
there, when they, when
709
00:37:45,380 --> 00:37:48,520
someone opens the chat, or
maybe, you know, a
710
00:37:48,520 --> 00:37:51,800
disclaimer after they are
being sent an email that
711
00:37:51,800 --> 00:37:55,340
was a par by AI. So
there's no right or wrong.
712
00:37:55,540 --> 00:37:57,720
I
feel like I would even
713
00:37:57,720 --> 00:38:00,540
encourage customers to
disclose it because it
714
00:38:00,540 --> 00:38:03,100
makes the experience much
more fun. I
715
00:38:03,100 --> 00:38:06,160
feel like it also like
reduces the pressure on,
716
00:38:06,580 --> 00:38:09,740
okay, you know, if the AI
just doesn't have enough
717
00:38:09,740 --> 00:38:12,960
data, like at least the
customer, you know, has had
718
00:38:12,960 --> 00:38:15,380
a good experience so far
and they're more like can
719
00:38:15,380 --> 00:38:20,000
be more, you know, patient
and just waiting for a
720
00:38:20,000 --> 00:38:22,960
real human to take over.
And
721
00:38:22,960 --> 00:38:25,060
also you can really make
it fun. So
722
00:38:25,060 --> 00:38:28,660
we've had some customers
who have, you know, from
723
00:38:28,660 --> 00:38:31,520
the human name, they've
turned it into a super fun
724
00:38:31,520 --> 00:38:34,040
like AI barista or stuff
like that. So
725
00:38:34,040 --> 00:38:37,720
I just feel like at the
end of the day, the
726
00:38:37,720 --> 00:38:40,940
customer, they just care
about getting their problems
727
00:38:40,940 --> 00:38:43,380
solved or getting the good
answer and having a good
728
00:38:43,380 --> 00:38:46,540
experience. It
doesn't really matter, you
729
00:38:46,540 --> 00:38:47,940
know, who they're speaking
with. That's
730
00:38:47,940 --> 00:38:51,620
what we're seeing so far
at least. Yeah,
731
00:38:51,680 --> 00:38:53,840
I think that's a really
nice way of framing it
732
00:38:53,840 --> 00:38:56,460
because I've had that
experience as a shopper.
733
00:38:56,820 --> 00:38:58,020
And
I've heard that from other
734
00:38:58,020 --> 00:39:00,260
people who have said the
feedback they've had for
735
00:39:00,260 --> 00:39:03,640
customers, which is we're
all we're all impatient. We
00:39:04,560
hate waiting. We
737
00:39:04,560 --> 00:39:07,620
can't be asked. And
if you're stuck waiting for
738
00:39:07,620 --> 00:39:10,860
a live chat agent, and
it's your 18th in the
739
00:39:10,860 --> 00:39:13,220
queue, so we'll be with
you. Oh,
740
00:39:13,320 --> 00:39:15,400
that's not bad. Yeah.
All right, it's just
741
00:39:15,400 --> 00:39:17,120
pissing. I've
had this on live chat with
742
00:39:17,120 --> 00:39:20,900
with a web hosting company
is you are connected to
743
00:39:20,900 --> 00:39:24,000
Derek. Derek,
where's Derek? Derek
744
00:39:24,000 --> 00:39:26,380
hasn't replied to me for
15 minutes. Is
745
00:39:26,380 --> 00:39:28,800
Derek connected to me or
is he pissed off on break?
746
00:39:29,020 --> 00:39:30,040
I
don't know. So
747
00:39:30,040 --> 00:39:32,580
I think you're right. I've
been having the ability to
748
00:39:32,580 --> 00:39:35,600
pre pre prompted of what
do you want to do track
749
00:39:35,600 --> 00:39:38,280
and order, which can just
be scripted. Having
750
00:39:38,280 --> 00:39:41,220
a chatbot that can say
hey, we're busy right now,
751
00:39:41,240 --> 00:39:42,920
well, let me help you
answer and do a really
752
00:39:42,920 --> 00:39:45,580
nice on brand really
positive time of the way.
753
00:39:45,980 --> 00:39:47,560
But
with a full back, which is
754
00:39:47,560 --> 00:39:49,800
that I'm really sorry,
can't help it. Let
755
00:39:49,800 --> 00:39:52,900
us let us send a call
back request that ability
756
00:39:52,900 --> 00:39:55,320
where you can control the
customer experience rather
757
00:39:55,320 --> 00:39:57,520
than make them sit there
waiting. I
758
00:39:57,520 --> 00:40:01,140
just think it's it's a no
brainer. Yeah,
759
00:40:01,140 --> 00:40:03,420
absolutely. An
acknowledgement, even one
760
00:40:03,420 --> 00:40:05,740
acknowledgement go so far.
Yep.
761
00:40:07,020 --> 00:40:09,900
Yeah, exactly. It's
about anything that takes
762
00:40:09,900 --> 00:40:12,220
friction when set
expectations because my
763
00:40:12,220 --> 00:40:14,780
biggest frustration at any
e -commerce or non e
764
00:40:14,780 --> 00:40:16,620
-commerce interaction is when
I don't know what's
765
00:40:16,620 --> 00:40:18,780
happening or how long I've
got away, or whether
766
00:40:18,780 --> 00:40:21,040
whether actually anything
will ever happen. So
767
00:40:21,880 --> 00:40:24,260
that's been really really
enjoyed that. I
768
00:40:24,260 --> 00:40:25,880
really enjoyed the chat we
had. I
769
00:40:25,880 --> 00:40:27,120
think it's a few weeks
back when you first
770
00:40:27,120 --> 00:40:28,740
introduced me to the
product. It's
771
00:40:28,740 --> 00:40:30,420
really cool. And
I can see why this is
772
00:40:30,420 --> 00:40:34,160
going to get more traction
with with ecom teams. If
773
00:40:34,160 --> 00:40:38,460
anyone listening wants to
get a demo or if they've
774
00:40:38,460 --> 00:40:40,340
got anything to challenge
you on in terms of goal,
775
00:40:40,400 --> 00:40:42,520
is this really going to
work for my business? How
776
00:40:42,520 --> 00:40:45,580
do they reach out? Yeah,
of course. So
777
00:40:45,580 --> 00:40:48,980
they can go on our
websites, yana .cx from
778
00:40:48,980 --> 00:40:52,060
customer experience. They
can book a time with us
779
00:40:52,060 --> 00:40:55,140
there or they can find us
on LinkedIn. I
780
00:40:55,140 --> 00:40:58,940
am Lisa Popovich on
LinkedIn or Twitter. They
781
00:40:58,940 --> 00:41:00,680
can find our team there as
well. But
782
00:41:00,680 --> 00:41:03,760
yeah, you can always reach
out to me. Cool.
783
00:41:04,680 --> 00:41:05,960
Excellent. Well,
thanks so much for taking
784
00:41:05,960 --> 00:41:07,360
the time. I
really enjoyed that. Good
785
00:41:07,360 --> 00:41:10,300
luck with the the future
plans. I
786
00:41:10,300 --> 00:41:13,540
look forward to seeing more
success from Sienna. I've
787
00:41:13,540 --> 00:41:15,320
everyone who's listened. I
hope you enjoyed that. It
788
00:41:15,328 --> 00:41:17,620
was a great episode. Make
sure you're following inside
789
00:41:17,620 --> 00:41:19,600
commerce on LinkedIn and
subscribe to our newsletter.
790
00:41:20,240 --> 00:41:21,880
And
feel free to reach out to
791
00:41:21,880 --> 00:41:24,140
Lisa if you want to find
out more about Sienna. Take
00:41:26,700
care, everybody. You're
793
00:41:26,700 --> 00:41:28,860
listening to the inside
commerce podcast hosted by
794
00:41:28,860 --> 00:41:31,300
experienced consultants James
Gert and Paul Rogers.
795
00:41:31,980 --> 00:41:33,500
Inside
commerce is your fast track
796
00:41:33,500 --> 00:41:35,740
to digital knowledge,
featuring interviews with
797
00:41:35,740 --> 00:41:37,480
leading e -commerce
retailers, industry
798
00:41:37,480 --> 00:41:39,960
thoughties, and the C -suite
key tech companies to
799
00:41:39,960 --> 00:41:42,520
explore the latest digital
strategy customer experience
800
00:41:42,520 --> 00:41:44,820
and tech. Discover
what the latest trends
801
00:41:44,820 --> 00:41:47,480
really mean and get
practical advice based on
802
00:41:47,480 --> 00:41:49,200
the experience to help find
you in your e -commerce
803
00:41:49,200 --> 00:41:51,320
strategy and taste them.
Follow
804
00:41:51,320 --> 00:41:53,880
us on LinkedIn and reach
out to James and Paul if
805
00:41:53,880 --> 00:41:56,200
you'd like to discuss an e
-commerce project. We
806
00:41:56,200 --> 00:41:57,660
hope you enjoy the latest
episode.