Description of the video:
WEBVTT
1
00:00:00.000 --> 00:00:02.880
Hello.
2
00:00:02.880 --> 00:00:04.580
And welcome, everyone.
3
00:00:04.580 --> 00:00:08.600
This is Paul Acito, coming
to you live from Minneapolis.
4
00:00:08.600 --> 00:00:10.480
And there we go again.
5
00:00:10.480 --> 00:00:14.720
And welcome to the
Kelley ExecEd Programs.
6
00:00:14.720 --> 00:00:19.040
This is a free webinar.
7
00:00:19.040 --> 00:00:20.760
But there's really
two objectives here
8
00:00:20.760 --> 00:00:24.220
that I wanted you to be
transparently aware of.
9
00:00:24.220 --> 00:00:28.280
One is that we do these
in order to create demand
10
00:00:28.280 --> 00:00:31.025
for our other courses,
in particular,
11
00:00:31.025 --> 00:00:33.400
a course called AI Applications
in Marketing, which we're
12
00:00:33.400 --> 00:00:35.580
doing for the fourth time.
13
00:00:35.580 --> 00:00:38.560
We've had about 100 executives
go through this program already.
14
00:00:38.560 --> 00:00:40.220
So it's a very
successful program.
15
00:00:40.220 --> 00:00:43.140
We'd invite you to
join us starting May 6.
16
00:00:43.140 --> 00:00:44.140
We still have some room.
17
00:00:44.140 --> 00:00:46.480
I think we're about 70% full.
18
00:00:46.480 --> 00:00:48.520
So act quickly.
19
00:00:48.520 --> 00:00:51.040
This is an online
course, instructor led.
20
00:00:51.040 --> 00:00:52.820
And we'll talk more
about it at the end,
21
00:00:52.820 --> 00:00:54.703
but I wanted to get
that out of the way
22
00:00:54.703 --> 00:00:56.620
because I know many of
you have busy schedules
23
00:00:56.620 --> 00:00:58.980
and I wanted to
put that in there.
24
00:00:58.980 --> 00:01:02.890
But here's an offer
that I don't think Kim--
25
00:01:02.890 --> 00:01:05.730
Dr. Allison-- knows that I
was going to make this offer.
26
00:01:05.730 --> 00:01:12.030
But if you can, text to Kim what
the name of the music school
27
00:01:12.030 --> 00:01:15.430
is, the top-ranked music
school at Indiana University,
28
00:01:15.430 --> 00:01:19.350
or who was the composer--
this is an easy one,
29
00:01:19.350 --> 00:01:22.790
but what the
symphony number was.
30
00:01:22.790 --> 00:01:25.050
And this is a trick question.
31
00:01:25.050 --> 00:01:28.813
So what TV show used
that as an opening
32
00:01:28.813 --> 00:01:31.230
for years and years and years,
probably before most of you
33
00:01:31.230 --> 00:01:32.430
were born?
34
00:01:32.430 --> 00:01:35.330
And try not to use the internet.
35
00:01:35.330 --> 00:01:36.330
Some of these you can't.
36
00:01:36.330 --> 00:01:38.070
But anyway, a little bit of fun.
37
00:01:38.070 --> 00:01:39.870
And Kim's got a
deal for you if get
38
00:01:39.870 --> 00:01:43.710
any of those right at the end.
39
00:01:43.710 --> 00:01:48.430
So we're here to discuss AI
agents and agentic workflows.
40
00:01:48.430 --> 00:01:51.990
And there's clearly a great
deal of discussion around
41
00:01:51.990 --> 00:01:53.890
what are AI agents.
42
00:01:53.890 --> 00:01:56.203
So I thought we would get
that cleared up right away.
43
00:01:56.203 --> 00:01:56.870
[VIDEO PLAYBACK]
44
00:01:56.870 --> 00:01:58.250
- So what are AI agents?
45
00:01:58.250 --> 00:01:59.790
- What is an AI agent?
46
00:01:59.790 --> 00:02:00.710
- What's an agent?
47
00:02:00.710 --> 00:02:03.050
- What is that and
what does it mean?
48
00:02:03.050 --> 00:02:03.670
- All right.
49
00:02:03.670 --> 00:02:06.190
So what is an AI agent?
50
00:02:06.190 --> 00:02:09.729
- What are AI agents and how are
they different than typical AIs?
51
00:02:09.729 --> 00:02:12.810
- What exactly are AI
agents and why is everybody
52
00:02:12.810 --> 00:02:13.990
talking about them?
53
00:02:13.990 --> 00:02:16.990
- What are AI agents and why is
everybody talking about them?
54
00:02:16.990 --> 00:02:19.010
- Agents are the new app.
55
00:02:19.010 --> 00:02:21.010
- AI agents are
simply programs--
56
00:02:21.010 --> 00:02:21.510
- Cool.
57
00:02:21.510 --> 00:02:23.850
Let's talk about
generative AI agents.
58
00:02:23.850 --> 00:02:26.790
Kevin threw this
statement up earlier.
59
00:02:26.790 --> 00:02:29.670
- No one seems to agree
what exactly an agent is.
60
00:02:29.670 --> 00:02:31.330
[END PLAYBACK]
61
00:02:31.330 --> 00:02:35.570
So if you have questions
about what AI agents are,
62
00:02:35.570 --> 00:02:39.530
you've come to the right place,
but you're also not alone.
63
00:02:39.530 --> 00:02:41.910
This is a definition
that I'm going to adopt.
64
00:02:41.910 --> 00:02:44.330
This is off of a
brand new publication
65
00:02:44.330 --> 00:02:46.530
that if you click
through, Kim will
66
00:02:46.530 --> 00:02:49.730
have this available to you
on the Kelley ExecEd site.
67
00:02:49.730 --> 00:02:53.030
This is OpenAI's A Practical
Guide to Building Agents,
68
00:02:53.030 --> 00:02:54.510
and this is how they start out.
69
00:02:54.510 --> 00:02:56.090
Page one title.
70
00:02:56.090 --> 00:03:00.600
So AI agents are systems that
independently accomplish tasks
71
00:03:00.600 --> 00:03:01.500
on your behalf.
72
00:03:01.500 --> 00:03:05.040
So these are what we will
be talking about in terms
73
00:03:05.040 --> 00:03:07.400
of AI agents today.
74
00:03:07.400 --> 00:03:11.040
A little more granularity
here, what an agent is.
75
00:03:11.040 --> 00:03:14.400
It is a goal-driven
system that can reason.
76
00:03:14.400 --> 00:03:16.020
That's hugely important.
77
00:03:16.020 --> 00:03:17.920
Plan and take action.
78
00:03:17.920 --> 00:03:24.080
So those three verbs are
extremely important-- reasoning,
79
00:03:24.080 --> 00:03:26.840
planning, and acting.
80
00:03:26.840 --> 00:03:29.220
This is very different
than what you've seen.
81
00:03:29.220 --> 00:03:30.648
And we've got some
great examples.
82
00:03:30.648 --> 00:03:32.440
And I'm even going to
show you how to build
83
00:03:32.440 --> 00:03:35.680
a simple agent at the very end.
84
00:03:35.680 --> 00:03:38.320
An agent is a dynamic
assistant, and it
85
00:03:38.320 --> 00:03:43.680
adapts based on your input,
context and then feedback.
86
00:03:43.680 --> 00:03:45.860
It's a bridge between tools.
87
00:03:45.860 --> 00:03:47.160
This is critical.
88
00:03:47.160 --> 00:03:48.620
It uses tools.
89
00:03:48.620 --> 00:03:51.100
It can write, analyze,
speak, trigger actions,
90
00:03:51.100 --> 00:03:55.280
and in some cases, even
create its own tools.
91
00:03:55.280 --> 00:03:57.820
It's a co-pilot, not
in the Microsoft sense,
92
00:03:57.820 --> 00:04:03.210
but a co-worker who works with
you to solve complex problems,
93
00:04:03.210 --> 00:04:07.390
but importantly,
multistep problems.
94
00:04:07.390 --> 00:04:11.830
And it's a scalable system that
can learn from your workflows,
95
00:04:11.830 --> 00:04:15.870
and it can get better over time.
96
00:04:15.870 --> 00:04:19.709
An AI agent is not simple.
97
00:04:19.709 --> 00:04:23.390
It's not a macro.
98
00:04:23.390 --> 00:04:25.550
It's not an assistant.
99
00:04:25.550 --> 00:04:27.670
It's not a static
chatbot, which is
100
00:04:27.670 --> 00:04:30.310
what we're used to using
up until this point.
101
00:04:30.310 --> 00:04:31.850
It's not siloed.
102
00:04:31.850 --> 00:04:37.150
It performs with sometimes
called a swarm of other agents.
103
00:04:37.150 --> 00:04:40.710
It's not a digital admin
that just executes commands
104
00:04:40.710 --> 00:04:43.470
without judgment or escalation.
105
00:04:43.470 --> 00:04:46.270
One of the defining
features of an AI agent
106
00:04:46.270 --> 00:04:50.510
is that it makes
decisions on your behalf,
107
00:04:50.510 --> 00:04:52.270
sometimes without asking.
108
00:04:52.270 --> 00:04:54.350
And it's important
when you get into this
109
00:04:54.350 --> 00:04:58.370
that you start to guardrail
action, or excuse me, agents.
110
00:04:58.370 --> 00:05:00.890
It's not a
set-it-and-forget-it tool.
111
00:05:00.890 --> 00:05:04.150
If anyone has tried this type
of agentic workflow and said,
112
00:05:04.150 --> 00:05:07.410
hey, it'd be nice if I
posted every week at 8 p.m.
113
00:05:07.410 --> 00:05:12.070
on a Tuesday and no need to send
a draft to my inbox, watch out.
114
00:05:12.070 --> 00:05:16.270
I did that and it was sending
out some not terrible stuff,
115
00:05:16.270 --> 00:05:17.950
but just real repetitive stuff.
116
00:05:17.950 --> 00:05:21.530
So it's something
that-- you're already
117
00:05:21.530 --> 00:05:24.150
being targeted with AI
agents all the time,
118
00:05:24.150 --> 00:05:29.050
especially if you're on
LinkedIn, Insta, or TikTok.
119
00:05:29.050 --> 00:05:32.410
I don't know if I love this
quote, but I use it a lot.
120
00:05:32.410 --> 00:05:36.770
And this is from Jensen
Huang, the CEO of NVIDIA.
121
00:05:36.770 --> 00:05:38.790
And I don't agree with
this, first of all.
122
00:05:38.790 --> 00:05:40.810
The IT department
of every company
123
00:05:40.810 --> 00:05:44.290
is going to be the HR department
of AI agents in the future.
124
00:05:44.290 --> 00:05:47.370
I don't believe that because
who makes the hiring decisions
125
00:05:47.370 --> 00:05:48.510
for the IT department?
126
00:05:48.510 --> 00:05:49.510
The HR department.
127
00:05:49.510 --> 00:05:51.930
So I think that they'll
probably pay attention.
128
00:05:51.930 --> 00:05:56.493
But let's take a look at one of
his most recent annual meetings
129
00:05:56.493 --> 00:05:57.160
[VIDEO PLAYBACK]
130
00:05:57.160 --> 00:06:00.480
- AI has been advancing
at an incredible pace.
131
00:06:00.480 --> 00:06:03.200
It started with perception AI.
132
00:06:03.200 --> 00:06:05.240
We now can understand
images and words
133
00:06:05.240 --> 00:06:08.720
and sounds, to generative AI.
134
00:06:08.720 --> 00:06:12.160
We can generate images
and text and sounds.
135
00:06:12.160 --> 00:06:15.600
And now agentic AI.
136
00:06:15.600 --> 00:06:20.840
AIs that can perceive,
reason, plan, and act.
137
00:06:20.840 --> 00:06:21.680
[END PLAYBACK]
138
00:06:21.680 --> 00:06:24.920
Reason, plan, and act.
139
00:06:24.920 --> 00:06:27.560
Sounds familiar.
140
00:06:27.560 --> 00:06:30.600
And this is a friend of mine.
141
00:06:30.600 --> 00:06:32.340
I'm not sure if
Nufar is joining us.
142
00:06:32.340 --> 00:06:34.360
She's in European
Central time zone.
143
00:06:34.360 --> 00:06:36.400
But she is one of our
colleagues and has
144
00:06:36.400 --> 00:06:39.360
taught the AI Applications
Marketing course with us.
145
00:06:39.360 --> 00:06:42.320
And spoiler alert, we're
going to be announcing a brand
146
00:06:42.320 --> 00:06:44.920
new course, probably
launched in early summer.
147
00:06:44.920 --> 00:06:47.720
And Nufar and I are putting that
together with other colleagues
148
00:06:47.720 --> 00:06:49.640
here at the Kelley School.
149
00:06:49.640 --> 00:06:51.820
And she is a bit of an expert.
150
00:06:51.820 --> 00:06:55.680
And I've got another link to
some of Nufar's online stuff
151
00:06:55.680 --> 00:06:56.360
in a second.
152
00:06:56.360 --> 00:07:00.740
But let's see what Nufar says
about the new IT department
153
00:07:00.740 --> 00:07:01.593
here.
154
00:07:01.593 --> 00:07:02.260
[VIDEO PLAYBACK]
155
00:07:02.260 --> 00:07:08.080
- I brought a quote that I find
to be very thought provoking,
156
00:07:08.080 --> 00:07:11.020
and that's from Jensen
Huang, NVIDIA's CEO.
157
00:07:11.020 --> 00:07:16.320
He said in the last CES keynote
that IT will become the new HR.
158
00:07:16.320 --> 00:07:17.940
And his intent was
that there will
159
00:07:17.940 --> 00:07:22.500
be so many agents to manage
that, as part of the workforce,
160
00:07:22.500 --> 00:07:25.700
someone will need to manage
them similarly to how
161
00:07:25.700 --> 00:07:29.580
we manage a human employees.
162
00:07:29.580 --> 00:07:32.540
And thereby, what
I wanted to do now
163
00:07:32.540 --> 00:07:35.620
is to talk about some--
let's call it even
164
00:07:35.620 --> 00:07:40.000
near future of work, and stuff
that we need to be aware of.
165
00:07:40.000 --> 00:07:43.300
So first and foremost,
everyone will manage AI agents.
166
00:07:43.300 --> 00:07:47.300
You will probably have between
a handful to a few dozen agents
167
00:07:47.300 --> 00:07:50.220
working for you in various
capacities for doing
168
00:07:50.220 --> 00:07:53.100
various elements of
your work at some point
169
00:07:53.100 --> 00:07:55.850
in the not-so-far
future, meaning
170
00:07:55.850 --> 00:08:00.910
that you need to have the
skills to manage these agents,
171
00:08:00.910 --> 00:08:06.195
and also to work alongside
agents as a team member.
172
00:08:06.195 --> 00:08:07.570
With all that said
and done, I do
173
00:08:07.570 --> 00:08:11.130
believe that human oversight
becomes even more critical
174
00:08:11.130 --> 00:08:12.470
and remains very important.
175
00:08:12.470 --> 00:08:15.290
And you just need to figure out
what's the right intervention
176
00:08:15.290 --> 00:08:18.870
points and what are the things
that still has to remain humans,
177
00:08:18.870 --> 00:08:23.570
and what are the things that we
can offload and how we offload.
178
00:08:23.570 --> 00:08:26.250
And the main skills that
you need for the future
179
00:08:26.250 --> 00:08:28.850
are really human skills.
180
00:08:28.850 --> 00:08:31.170
So those are communication
skills, your ability
181
00:08:31.170 --> 00:08:33.250
to really clearly
set and articulate
182
00:08:33.250 --> 00:08:37.809
your goals, your ability
to delegate work to others
183
00:08:37.809 --> 00:08:42.130
and manage the process
in a cost-effective way,
184
00:08:42.130 --> 00:08:45.530
as well as orchestrating the
work of multiple agents working
185
00:08:45.530 --> 00:08:48.230
for you and for
others all together.
186
00:08:48.230 --> 00:08:51.990
So these are all skills that if
you are not that good at them,
187
00:08:51.990 --> 00:08:55.200
now is the right time to
skill up and make sure
188
00:08:55.200 --> 00:08:57.600
that you're becoming an
excellent communicator
189
00:08:57.600 --> 00:09:01.240
and a very sharp minded
about what the work
190
00:09:01.240 --> 00:09:04.560
that needs to be done
and how to manage that.
191
00:09:04.560 --> 00:09:09.060
And these are all
going, to some extent,
192
00:09:09.060 --> 00:09:12.520
have you future proof
to what's coming.
193
00:09:12.520 --> 00:09:14.840
[END PLAYBACK]
194
00:09:14.840 --> 00:09:18.040
So one of the things I love
about Nufar and our relationship
195
00:09:18.040 --> 00:09:21.440
in terms of AI agents
is that we remain
196
00:09:21.440 --> 00:09:25.500
pliable in how we're looking
at this because this is,
197
00:09:25.500 --> 00:09:31.660
of course, evolving extremely
rapidly, but evolving,
198
00:09:31.660 --> 00:09:32.660
and we're adapting.
199
00:09:32.660 --> 00:09:35.580
And Nufar and I are both
working with clients.
200
00:09:35.580 --> 00:09:39.300
We're deploying agents, and we
can tell you the ups and downs.
201
00:09:39.300 --> 00:09:43.520
And again, I'll refer you later
to some additional information.
202
00:09:43.520 --> 00:09:45.400
And this is my
particular viewpoint,
203
00:09:45.400 --> 00:09:50.940
which will evolve, but is that
AI agents, assistants, avatars,
204
00:09:50.940 --> 00:09:56.740
automations, and even APIs
have a good deal of overlap.
205
00:09:56.740 --> 00:09:59.820
They depend on each other
in some circumstances,
206
00:09:59.820 --> 00:10:04.380
and that while it's important
to have a definition of an AI
207
00:10:04.380 --> 00:10:06.740
agent that will
continue to evolve,
208
00:10:06.740 --> 00:10:08.280
and we will see
where this lands.
209
00:10:08.280 --> 00:10:11.700
But the good news is that you're
on this call, so you're curious
210
00:10:11.700 --> 00:10:14.340
and you're exploring,
and you're probably
211
00:10:14.340 --> 00:10:16.620
wondering how either
you can apply these
212
00:10:16.620 --> 00:10:19.020
in your personal
workflows or how
213
00:10:19.020 --> 00:10:22.100
you can respond in helping
improve, take cycle
214
00:10:22.100 --> 00:10:25.300
time out, or improve the
growth of your companies
215
00:10:25.300 --> 00:10:29.700
or firm's processes.
216
00:10:29.700 --> 00:10:31.160
So what I thought
I would do now,
217
00:10:31.160 --> 00:10:35.980
and I sometimes get criticized
for doing too much too fast,
218
00:10:35.980 --> 00:10:39.960
but AI is too much too fast.
219
00:10:39.960 --> 00:10:43.800
So we're going to roll down and
take a look at research agents.
220
00:10:43.800 --> 00:10:46.780
Some of you are going to
say, I know all about that--
221
00:10:46.780 --> 00:10:49.560
coding agents, voice
agents, agent orchestration.
222
00:10:49.560 --> 00:10:54.170
So we had 800 people plus
sign up for this, which
223
00:10:54.170 --> 00:10:56.090
is typical for these webinars.
224
00:10:56.090 --> 00:10:58.810
There's almost no
way we can fine tune
225
00:10:58.810 --> 00:11:01.438
for the folks on the phone.
226
00:11:01.438 --> 00:11:03.730
So we hope there will be a
little bit of something here
227
00:11:03.730 --> 00:11:05.530
for everyone, whether
you're extremely
228
00:11:05.530 --> 00:11:09.290
advanced on the subject
or you're a beginner
229
00:11:09.290 --> 00:11:10.910
and/or somewhere in the middle.
230
00:11:10.910 --> 00:11:14.350
This is a great way to do this.
231
00:11:14.350 --> 00:11:16.093
And I apologize in
advance if you already
232
00:11:16.093 --> 00:11:18.010
know more about AI agents
than I'll ever know.
233
00:11:18.010 --> 00:11:21.330
But this is just
some good examples
234
00:11:21.330 --> 00:11:25.650
that really struck me and helped
generate some additional ideas
235
00:11:25.650 --> 00:11:27.390
on where I can apply
them in my work,
236
00:11:27.390 --> 00:11:30.090
and I hope you find the same.
237
00:11:30.090 --> 00:11:33.130
So let's take a look
at research agents.
238
00:11:33.130 --> 00:11:36.410
Now, research agents are
all what you've heard about.
239
00:11:36.410 --> 00:11:39.730
It's Gemini Deep Research.
240
00:11:39.730 --> 00:11:43.370
It's OpenAI Deep Research.
241
00:11:43.370 --> 00:11:47.250
It's Perplexity Deep Research.
242
00:11:47.250 --> 00:11:50.410
It's xAI.
243
00:11:50.410 --> 00:11:53.160
It's not deep research,
but it's something similar.
244
00:11:53.160 --> 00:11:55.880
So apparently, the only
thing that it can't research
245
00:11:55.880 --> 00:11:59.360
is brand names because
they see-- or maybe they
246
00:11:59.360 --> 00:12:03.400
did use research and it came
up with the same brand name
247
00:12:03.400 --> 00:12:04.760
every time.
248
00:12:04.760 --> 00:12:11.220
But what you're seeing here is
a way that you set it in motion.
249
00:12:11.220 --> 00:12:14.400
OpenAI did something that the
other two didn't, so I gave it
250
00:12:14.400 --> 00:12:17.600
all the same prompt, which was
generate a two-page executive
251
00:12:17.600 --> 00:12:18.980
summary, which they all failed.
252
00:12:18.980 --> 00:12:22.540
I got two pages or three pages,
six pages, and nine pages
253
00:12:22.540 --> 00:12:25.960
because there's probably
so much data out there.
254
00:12:25.960 --> 00:12:29.240
And the quality was excellent.
255
00:12:29.240 --> 00:12:31.260
This is one of the
strongest ones.
256
00:12:31.260 --> 00:12:33.600
And what makes
this agentic is you
257
00:12:33.600 --> 00:12:37.040
can see how it's
plugging through this.
258
00:12:37.040 --> 00:12:40.162
In the case of OpenAI, it
actually comes back and says,
259
00:12:40.162 --> 00:12:41.620
hey, answer these
questions for me.
260
00:12:41.620 --> 00:12:44.200
So there's some
human-in-the-loop activity.
261
00:12:44.200 --> 00:12:50.900
And then with Gemini,
it did the same thing.
262
00:12:50.900 --> 00:12:54.980
So those are some of the
most easily accessible
263
00:12:54.980 --> 00:12:59.300
agentic workflows, and
they're available on almost
264
00:12:59.300 --> 00:13:02.800
every subscription level
in those programs--
265
00:13:02.800 --> 00:13:09.300
ChatGPT, Gemini, and Perplexity.
266
00:13:09.300 --> 00:13:11.120
Now, if you haven't
tried coding agents,
267
00:13:11.120 --> 00:13:14.220
this sounds more
complicated than it is,
268
00:13:14.220 --> 00:13:16.380
but there are a
number of companies
269
00:13:16.380 --> 00:13:21.260
out there with game-changing
coding platforms.
270
00:13:21.260 --> 00:13:24.438
This is where you use
English and express an idea--
271
00:13:24.438 --> 00:13:26.980
and we'll have an example at
the end on how to build an agent
272
00:13:26.980 --> 00:13:28.380
like this--
273
00:13:28.380 --> 00:13:30.820
and it writes code for you.
274
00:13:30.820 --> 00:13:33.740
And you can see, in this
case, a little more clearly
275
00:13:33.740 --> 00:13:37.220
than in the research agents
and agentic workflows,
276
00:13:37.220 --> 00:13:40.300
how it goes to work
doing the coding.
277
00:13:40.300 --> 00:13:43.100
So here, this is
Anthropic's Claude.
278
00:13:43.100 --> 00:13:44.780
And here's a survey.
279
00:13:44.780 --> 00:13:49.060
It's a 200-person survey
that we had responses to.
280
00:13:49.060 --> 00:13:51.690
And this is the thing
that is most impressive
281
00:13:51.690 --> 00:13:53.365
about these large
language models.
282
00:13:53.365 --> 00:13:55.490
If you don't know it already,
large language models
283
00:13:55.490 --> 00:13:59.273
kind of stink at math, but
they're really good at English,
284
00:13:59.273 --> 00:14:00.690
and they're good
enough at English
285
00:14:00.690 --> 00:14:07.010
to translate your English
into Python or JSON or YAML,
286
00:14:07.010 --> 00:14:12.330
or you name it, and produce
code and then run and execute
287
00:14:12.330 --> 00:14:12.830
this code.
288
00:14:12.830 --> 00:14:14.770
One of the things-- and
again, I keep referring
289
00:14:14.770 --> 00:14:15.728
to this one at the end.
290
00:14:15.728 --> 00:14:17.810
I hope it lives up
to your expectations,
291
00:14:17.810 --> 00:14:20.470
is this was create
a research report.
292
00:14:20.470 --> 00:14:22.690
Now, this was a
200-respondent survey.
293
00:14:22.690 --> 00:14:24.970
Probably about 60 questions.
294
00:14:24.970 --> 00:14:27.330
And Bob's your uncle.
295
00:14:27.330 --> 00:14:28.690
There it is.
296
00:14:28.690 --> 00:14:31.530
Not only does it produce this
in an agentic fashion, where
297
00:14:31.530 --> 00:14:35.090
it makes the decisions that
go along with the analysis,
298
00:14:35.090 --> 00:14:38.450
but then it offers you an
opportunity to publish this.
299
00:14:38.450 --> 00:14:43.370
So that is a very, very simple--
both times, very simple agents
300
00:14:43.370 --> 00:14:45.450
so far.
301
00:14:45.450 --> 00:14:48.490
Now, I'm going to give you
an example of a voice agent.
302
00:14:48.490 --> 00:14:50.470
And this is a
little bit different
303
00:14:50.470 --> 00:14:56.430
because it involves spoken
and understood human voice.
304
00:14:56.430 --> 00:15:01.370
And this is something that is
probably coming on very fast.
305
00:15:01.370 --> 00:15:02.690
It wasn't around a year ago.
306
00:15:02.690 --> 00:15:04.610
You couldn't do some of
this stuff a year ago.
307
00:15:04.610 --> 00:15:06.910
Some of this you couldn't
do six months ago.
308
00:15:06.910 --> 00:15:10.250
And these voice
agents are incredible.
309
00:15:10.250 --> 00:15:13.590
They're going to change customer
service, service operations,
310
00:15:13.590 --> 00:15:15.970
and, I believe, market research.
311
00:15:15.970 --> 00:15:17.710
Here's an agent that
we built to collect
312
00:15:17.710 --> 00:15:22.950
some data on our students and
help them by producing a custom
313
00:15:22.950 --> 00:15:26.750
report on an assessment
that was given
314
00:15:26.750 --> 00:15:29.590
by an artificial intelligence.
315
00:15:29.590 --> 00:15:31.170
So let's look and listen.
316
00:15:31.170 --> 00:15:35.174
This is my recorded
voice, by the way.
317
00:15:35.174 --> 00:15:35.950
[AUDIO PLAYBACK]
318
00:15:35.950 --> 00:15:36.450
- Hello.
319
00:15:36.450 --> 00:15:37.210
This is Bianca.
320
00:15:37.210 --> 00:15:38.090
Thanks for calling.
321
00:15:38.090 --> 00:15:40.710
To whom do I have the pleasure
of speaking with today?
322
00:15:40.710 --> 00:15:41.850
- This is Paul.
323
00:15:41.850 --> 00:15:42.350
- Hi.
324
00:15:42.350 --> 00:15:44.210
We are looking forward to
having you in the course.
325
00:15:44.210 --> 00:15:45.790
We have seven
questions we'd like
326
00:15:45.790 --> 00:15:48.790
to ask to help us better tailor
the course to your learning
327
00:15:48.790 --> 00:15:49.290
goals.
328
00:15:49.290 --> 00:15:51.180
Let's get started, Paul.
329
00:15:51.180 --> 00:15:52.820
So first off, why
don't you tell me
330
00:15:52.820 --> 00:15:55.140
about how you're thinking
about artificial intelligence
331
00:15:55.140 --> 00:15:55.640
these days?
332
00:15:55.640 --> 00:15:58.380
Are you excited, worried,
or some combination
333
00:15:58.380 --> 00:16:00.400
regarding the prospects of AI?
334
00:16:00.400 --> 00:16:01.900
- Well, it's always
some combination
335
00:16:01.900 --> 00:16:03.800
of excited and worried.
336
00:16:03.800 --> 00:16:08.580
It's unstoppable technical
force, which is troubling.
337
00:16:08.580 --> 00:16:12.220
But at the same time, it's
exciting to think about some
338
00:16:12.220 --> 00:16:13.980
of the possibilities.
339
00:16:13.980 --> 00:16:14.900
[END PLAYBACK]
340
00:16:14.900 --> 00:16:16.580
So that was the
data collection--
341
00:16:16.580 --> 00:16:16.700
[AUDIO PLAYBACK]
342
00:16:16.700 --> 00:16:17.825
--you get back to your day.
343
00:16:17.825 --> 00:16:21.020
What, if anything else,
would you like to add?
344
00:16:21.020 --> 00:16:22.480
- Just excited about the course.
345
00:16:22.480 --> 00:16:24.240
Looking forward to
meeting everyone.
346
00:16:24.240 --> 00:16:24.740
- You too.
347
00:16:24.740 --> 00:16:26.100
Talk soon, Paul.
348
00:16:26.100 --> 00:16:34.380
[END PLAYBACK]
349
00:16:34.380 --> 00:16:37.260
So the data are captured.
350
00:16:37.260 --> 00:16:40.460
The data are
interpreted by an AI.
351
00:16:40.460 --> 00:16:43.740
A report based on
that student interview
352
00:16:43.740 --> 00:16:49.930
is generated, formatted,
and emailed to the student
353
00:16:49.930 --> 00:16:51.470
without a human in the loop.
354
00:16:51.470 --> 00:16:56.250
Now, I just got through teaching
a principled leadership course,
355
00:16:56.250 --> 00:17:00.370
and one of my messages to them
was to be the human in the loop.
356
00:17:00.370 --> 00:17:03.590
You do not want to delegate
this stuff first try.
357
00:17:03.590 --> 00:17:07.170
And I've got some pointers on
what my personal philosophy is
358
00:17:07.170 --> 00:17:10.290
on these agentic workflows
and incorporating these
359
00:17:10.290 --> 00:17:11.770
into your work.
360
00:17:11.770 --> 00:17:14.410
But for the first time,
these are poking--
361
00:17:14.410 --> 00:17:15.990
just like automations have.
362
00:17:15.990 --> 00:17:18.650
We've got some
great, great players.
363
00:17:18.650 --> 00:17:23.369
Just like automations have, have
poked through the corporate IT
364
00:17:23.369 --> 00:17:26.530
area and now can be deployed
at the individual level
365
00:17:26.530 --> 00:17:29.850
to improve your own
personal workflows.
366
00:17:29.850 --> 00:17:32.210
Next, I want to call your
attention to something
367
00:17:32.210 --> 00:17:33.870
called agent orchestration.
368
00:17:33.870 --> 00:17:39.130
And this is broad
windows of orchestration
369
00:17:39.130 --> 00:17:41.370
that come from
places like OpenAI
370
00:17:41.370 --> 00:17:46.770
and a company called Manus,
where they order around
371
00:17:46.770 --> 00:17:49.170
and orchestrate other agents.
372
00:17:49.170 --> 00:17:52.670
And I have to tell you,
I've been very impressed.
373
00:17:52.670 --> 00:17:54.330
I have found some use cases.
374
00:17:54.330 --> 00:17:57.590
I think you'll have to noodle
a bit before you understand
375
00:17:57.590 --> 00:17:59.510
what the use cases are.
376
00:17:59.510 --> 00:18:04.110
And I will tell you that the
OpenAI access to their OpenAI
377
00:18:04.110 --> 00:18:08.770
Operator is only available to
the $200-a-month subscriber
378
00:18:08.770 --> 00:18:09.270
level.
379
00:18:09.270 --> 00:18:11.910
380
00:18:11.910 --> 00:18:15.910
But Manus-- it's also in beta
and a little hard to get,
381
00:18:15.910 --> 00:18:18.530
but I was fortunate enough
to be included in their beta,
382
00:18:18.530 --> 00:18:19.690
and I've been impressed.
383
00:18:19.690 --> 00:18:20.523
I've been impressed.
384
00:18:20.523 --> 00:18:22.230
Let's take a look.
385
00:18:22.230 --> 00:18:24.590
So this is Operator.
386
00:18:24.590 --> 00:18:29.370
And here, we're talking
about Bianca, again, who was,
387
00:18:29.370 --> 00:18:31.570
of course, our interview agent.
388
00:18:31.570 --> 00:18:32.810
I used that name a lot.
389
00:18:32.810 --> 00:18:37.230
Turns out Bianca is easily
recognized by the AI voice.
390
00:18:37.230 --> 00:18:41.150
And I said, go sign her up
for a Kelley ExecEd course.
391
00:18:41.150 --> 00:18:43.350
That's all I said.
392
00:18:43.350 --> 00:18:45.870
And at this point,
thankfully, it
393
00:18:45.870 --> 00:18:49.620
can get all the way to when you
got to put your credit card in.
394
00:18:49.620 --> 00:18:52.780
And then it comes back and
asks you for permission.
395
00:18:52.780 --> 00:18:54.640
You know and I know
that'll change.
396
00:18:54.640 --> 00:18:57.180
I can remember when
one click on Amazon
397
00:18:57.180 --> 00:19:00.660
was new behavior
that we wouldn't
398
00:19:00.660 --> 00:19:05.480
trust an online vendor to
have your credit card number.
399
00:19:05.480 --> 00:19:07.620
And, of course, now,
I think we're all
400
00:19:07.620 --> 00:19:10.360
doing that on a pretty
much daily basis.
401
00:19:10.360 --> 00:19:20.447
402
00:19:20.447 --> 00:19:22.280
I'm sorry if it's a
little too slow for you,
403
00:19:22.280 --> 00:19:26.680
but how impatient have we
gotten with this technology?
404
00:19:26.680 --> 00:19:29.920
How quickly have we incorporated
it into our day-to-day workflows
405
00:19:29.920 --> 00:19:32.260
to point where we don't
want to wait around
406
00:19:32.260 --> 00:19:34.780
for 15 minutes
for something that
407
00:19:34.780 --> 00:19:38.200
would have taken us two weeks,
or an hour, or some other thing?
408
00:19:38.200 --> 00:19:42.580
So this is part and parcel of
the experience with these apps
409
00:19:42.580 --> 00:19:43.780
as they come out.
410
00:19:43.780 --> 00:19:48.470
We are on the jagged
frontier of AI with agents,
411
00:19:48.470 --> 00:19:49.950
but they're coming fast.
412
00:19:49.950 --> 00:19:54.370
And as you can see, every
company is focused on these.
413
00:19:54.370 --> 00:19:58.670
And this is going to have a huge
impact on the business world,
414
00:19:58.670 --> 00:20:00.690
on marketing jobs, and the like.
415
00:20:00.690 --> 00:20:04.650
So long story short on this,
it does a pretty good job
416
00:20:04.650 --> 00:20:06.530
of handling this.
417
00:20:06.530 --> 00:20:11.370
So this is OpenAI Operator.
418
00:20:11.370 --> 00:20:12.750
I'll let it finish up here.
419
00:20:12.750 --> 00:20:15.490
And you'll see that
it terminates in--
420
00:20:15.490 --> 00:20:20.890
oh, I can't go any further,
which again, is a positive.
421
00:20:20.890 --> 00:20:22.850
This is Manus.
422
00:20:22.850 --> 00:20:28.150
And here, I said, look, I've
got this Sam Altman article.
423
00:20:28.150 --> 00:20:30.250
I want you to generate
a summary for me.
424
00:20:30.250 --> 00:20:34.610
And it goes at it.
425
00:20:34.610 --> 00:20:40.370
And I said, take a look at
his last three blog posts,
426
00:20:40.370 --> 00:20:44.690
summarize them and
analyze it and tell me
427
00:20:44.690 --> 00:20:46.430
where you think AI is--
428
00:20:46.430 --> 00:20:50.350
excuse me-- Sam Altman
is taking OpenAI next.
429
00:20:50.350 --> 00:20:52.550
And it did exactly that.
430
00:20:52.550 --> 00:20:55.670
You can ask it to generate
three different reports, one
431
00:20:55.670 --> 00:20:59.450
on each blog post, do the
analysis, all with a click.
432
00:20:59.450 --> 00:21:04.070
Go get a coffee, come
back, and it's done.
433
00:21:04.070 --> 00:21:06.710
As Sam Altman
himself said, we're
434
00:21:06.710 --> 00:21:08.710
going to need to get
used to these things
435
00:21:08.710 --> 00:21:14.310
clicking around the internet
without supervision.
436
00:21:14.310 --> 00:21:16.090
So that's agent orchestration.
437
00:21:16.090 --> 00:21:19.990
That's OpenAI Operator,
and that's Manus.
438
00:21:19.990 --> 00:21:22.610
Let's take a look
at agent building.
439
00:21:22.610 --> 00:21:26.643
Now, this is the promised
demonstration that I've got.
440
00:21:26.643 --> 00:21:28.310
And there's a couple
of different things
441
00:21:28.310 --> 00:21:30.110
we can talk about here.
442
00:21:30.110 --> 00:21:33.030
My favorite, my recommendation
to you for today
443
00:21:33.030 --> 00:21:35.910
is to start with
Anthropic's Claude.
444
00:21:35.910 --> 00:21:39.670
And you can see what we
can do here is-- again,
445
00:21:39.670 --> 00:21:43.830
these LLMs are not great at math
but they're perfect at English,
446
00:21:43.830 --> 00:21:46.940
and they really excel
at writing code.
447
00:21:46.940 --> 00:21:50.520
So they provide that bridge,
which, if you think about it,
448
00:21:50.520 --> 00:21:52.240
is what's been going
on all the time.
449
00:21:52.240 --> 00:21:54.460
It's just been
static up til now.
450
00:21:54.460 --> 00:21:59.500
You would enter a number or
a formula in an Excel cell,
451
00:21:59.500 --> 00:22:02.500
and it would have
code underlying that
452
00:22:02.500 --> 00:22:03.900
and go execute some function.
453
00:22:03.900 --> 00:22:06.300
Here, it's the same process.
454
00:22:06.300 --> 00:22:10.020
You put in a prompt in English,
ask it to do something, perhaps
455
00:22:10.020 --> 00:22:14.160
math, and it goes off,
writes the Python code--
456
00:22:14.160 --> 00:22:16.180
and you should be aware of this.
457
00:22:16.180 --> 00:22:18.360
Every time you ask
it to do something,
458
00:22:18.360 --> 00:22:22.140
it's writing the Python
code over and over again.
459
00:22:22.140 --> 00:22:26.140
So you may try this and get
a really good result at first
460
00:22:26.140 --> 00:22:28.940
and go to show
somebody, do it again,
461
00:22:28.940 --> 00:22:31.460
and get a different result,
maybe not substantially
462
00:22:31.460 --> 00:22:32.160
different.
463
00:22:32.160 --> 00:22:33.980
And then the third
time, it may come back
464
00:22:33.980 --> 00:22:36.220
and say, yeah, I got hung
up, I made a mistake.
465
00:22:36.220 --> 00:22:37.620
Do you want me to fix it?
466
00:22:37.620 --> 00:22:39.660
So again, it's
trying to write code,
467
00:22:39.660 --> 00:22:42.300
and it's like an eager
intern right now.
468
00:22:42.300 --> 00:22:45.320
The predictions are
that 90% of all code
469
00:22:45.320 --> 00:22:48.000
will be done, within
three to six months,
470
00:22:48.000 --> 00:22:50.200
by these AI coding agents.
471
00:22:50.200 --> 00:22:54.580
And Dario Amodei, the Chief
Technical Officer at Anthropic,
472
00:22:54.580 --> 00:22:58.400
predicts that within 12 months
to two years, that almost
473
00:22:58.400 --> 00:23:00.000
and essentially all
of the code will
474
00:23:00.000 --> 00:23:01.700
be written by these AI agents.
475
00:23:01.700 --> 00:23:04.500
Let's take a look at one
that you can build at home,
476
00:23:04.500 --> 00:23:05.620
and I'll show you how.
477
00:23:05.620 --> 00:23:08.240
478
00:23:08.240 --> 00:23:10.740
So I'll give you a
flash warning here.
479
00:23:10.740 --> 00:23:17.560
I started this so early in the
morning that when it came on,
480
00:23:17.560 --> 00:23:20.520
the screen shifts
to bright white
481
00:23:20.520 --> 00:23:22.500
because it was dark
when I started.
482
00:23:22.500 --> 00:23:23.860
So I had a dark screen.
483
00:23:23.860 --> 00:23:24.940
So that's not a glitch.
484
00:23:24.940 --> 00:23:26.732
That's just me
being up, probably,
485
00:23:26.732 --> 00:23:28.940
and working on this stuff
earlier than I should have.
486
00:23:28.940 --> 00:23:32.000
487
00:23:32.000 --> 00:23:34.160
But here, you can
see the thinking.
488
00:23:34.160 --> 00:23:37.000
These are called artifacts.
489
00:23:37.000 --> 00:23:40.157
And this is a little
accelerated, but not by much.
490
00:23:40.157 --> 00:23:41.240
And it's writing the code.
491
00:23:41.240 --> 00:23:44.230
There, you can see
my computer woke up.
492
00:23:44.230 --> 00:23:46.950
And look at the beautiful
graphic user interface
493
00:23:46.950 --> 00:23:48.590
that it produces.
494
00:23:48.590 --> 00:23:54.350
And this you can publish and
share with others instantly.
495
00:23:54.350 --> 00:23:57.510
So if you have an app you've
always wanted to write,
496
00:23:57.510 --> 00:23:59.150
it's the way to do it.
497
00:23:59.150 --> 00:24:00.650
And you can see the
pull-down menus.
498
00:24:00.650 --> 00:24:02.270
Everything works fine.
499
00:24:02.270 --> 00:24:07.410
And it's not a coincidence
that I chose this Machines
500
00:24:07.410 --> 00:24:08.770
of Beauty and Grace.
501
00:24:08.770 --> 00:24:12.910
One is because it's
a 16,000-word essay.
502
00:24:12.910 --> 00:24:15.270
Some might call it a manifesto.
503
00:24:15.270 --> 00:24:17.570
And the other is
that, of course,
504
00:24:17.570 --> 00:24:21.010
Dario Amodei is the Chief
Technical Officer at Anthropic,
505
00:24:21.010 --> 00:24:26.230
which is a fantastic AI company
and brings you that capability.
506
00:24:26.230 --> 00:24:28.590
And try it yourself.
507
00:24:28.590 --> 00:24:35.070
So this is a link, I
think, to that app.
508
00:24:35.070 --> 00:24:39.550
And you can go on
to Anthropic Claude
509
00:24:39.550 --> 00:24:41.768
and try to build your
own agent just like that.
510
00:24:41.768 --> 00:24:44.060
Give it an English prompt,
tell it what you want to do.
511
00:24:44.060 --> 00:24:46.900
And you can figure
out how to publish up
512
00:24:46.900 --> 00:24:52.860
in the upper right hand corner
and develop an AI agent app.
513
00:24:52.860 --> 00:24:57.420
So some advice as you venture
out and create your first AI
514
00:24:57.420 --> 00:24:58.860
agents.
515
00:24:58.860 --> 00:25:01.660
If you don't have a
process written down,
516
00:25:01.660 --> 00:25:03.520
you don't understand
your workflows,
517
00:25:03.520 --> 00:25:07.580
or Mary and Joe have
been doing this for years
518
00:25:07.580 --> 00:25:09.380
and it's all in their
head, you probably
519
00:25:09.380 --> 00:25:10.760
are going to need a process map.
520
00:25:10.760 --> 00:25:12.940
So companies that don't
have good process discipline
521
00:25:12.940 --> 00:25:15.360
and process excellence are
going to be caught off.
522
00:25:15.360 --> 00:25:17.420
But those who understand
their process,
523
00:25:17.420 --> 00:25:21.340
are constantly improving
their processes are really
524
00:25:21.340 --> 00:25:25.180
in good shape to start
off with creating
525
00:25:25.180 --> 00:25:29.380
agentic workflows within
their important processes.
526
00:25:29.380 --> 00:25:35.540
Don't start off with an entire
organizational chart of agents.
527
00:25:35.540 --> 00:25:37.540
Start with one.
528
00:25:37.540 --> 00:25:39.160
Keep the human in the loop.
529
00:25:39.160 --> 00:25:42.420
In other words, don't just
have it go out and publish,
530
00:25:42.420 --> 00:25:46.200
even though things like
Make.com make that very easy.
531
00:25:46.200 --> 00:25:48.840
They also make it very easy for
it to put a draft in your inbox
532
00:25:48.840 --> 00:25:52.320
or send you an email before
you hit click and publish.
533
00:25:52.320 --> 00:25:53.900
Stay internal at first.
534
00:25:53.900 --> 00:25:56.260
Why expose your customers
to your experimentation?
535
00:25:56.260 --> 00:25:59.480
I just think this is
good alpha testing.
536
00:25:59.480 --> 00:26:02.560
And it's a great
idea in terms of how
537
00:26:02.560 --> 00:26:08.240
to work your way up the
ladder of expertise in agents.
538
00:26:08.240 --> 00:26:11.640
And try to balance
that lowest risk.
539
00:26:11.640 --> 00:26:13.977
Don't put your financial
data and your Social Security
540
00:26:13.977 --> 00:26:15.060
number and all that stuff.
541
00:26:15.060 --> 00:26:18.440
Have it do something simple,
like summarize a report
542
00:26:18.440 --> 00:26:21.720
or do some analysis.
543
00:26:21.720 --> 00:26:24.600
I've done a very good job using
this in some of the classes
544
00:26:24.600 --> 00:26:27.180
I teach, where it can
help me grade essays,
545
00:26:27.180 --> 00:26:29.840
but I remain the
human in the loop.
546
00:26:29.840 --> 00:26:31.340
But it helps me organize.
547
00:26:31.340 --> 00:26:34.160
It'll put all the
essays, like, 38 of them
548
00:26:34.160 --> 00:26:37.080
into a single Excel
spreadsheet, each one getting
549
00:26:37.080 --> 00:26:39.520
its own worksheet,
and then give me
550
00:26:39.520 --> 00:26:41.920
preliminary thoughts
and grades, which
551
00:26:41.920 --> 00:26:45.070
turn out to be, more
cases than not, right.
552
00:26:45.070 --> 00:26:47.412
So where can you get quick wins?
553
00:26:47.412 --> 00:26:48.870
You can demonstrate
internally what
554
00:26:48.870 --> 00:26:54.030
the value of these agents
on what are low-risk ideas.
555
00:26:54.030 --> 00:26:57.510
So with that, I'm going to
pitch once again-- so bookmark
556
00:26:57.510 --> 00:27:01.070
or bookend, I guess, the Kelley
ExecEd Program, AI Applications
557
00:27:01.070 --> 00:27:01.730
in Marketing.
558
00:27:01.730 --> 00:27:03.570
This QR code will
take you there.
559
00:27:03.570 --> 00:27:04.990
We start May 6.
560
00:27:04.990 --> 00:27:09.270
And I'm going to show you the
exciting faculty that we've got.
561
00:27:09.270 --> 00:27:11.070
It's online, instructor led.
562
00:27:11.070 --> 00:27:12.110
And I don't know--
563
00:27:12.110 --> 00:27:15.590
I'm sure there are some
of my former students--
564
00:27:15.590 --> 00:27:18.730
we call them executive learners
because they are, in many cases,
565
00:27:18.730 --> 00:27:21.510
chief marketing officers,
marketing directors,
566
00:27:21.510 --> 00:27:26.430
and online marketing leaders.
567
00:27:26.430 --> 00:27:27.930
And we meet Tuesdays
and Thursdays.
568
00:27:27.930 --> 00:27:30.190
It's at 6 to 7:30 p.m.
569
00:27:30.190 --> 00:27:31.170
Eastern time.
570
00:27:31.170 --> 00:27:34.010
And take a look at this faculty.
571
00:27:34.010 --> 00:27:36.410
So we just left
Caroline Ylitalo,
572
00:27:36.410 --> 00:27:38.490
who is the 3M
scientist-- or excuse me,
573
00:27:38.490 --> 00:27:40.830
the Minnesota
Scientist of the Year.
574
00:27:40.830 --> 00:27:43.720
Just gave a luncheon speech
at our principal's class
575
00:27:43.720 --> 00:27:46.600
here, is one of the speakers.
576
00:27:46.600 --> 00:27:47.840
And she's fascinating.
577
00:27:47.840 --> 00:27:51.160
Over 126 patents because--
she just told us that.
578
00:27:51.160 --> 00:27:53.440
And I thought it
was 100 patents,
579
00:27:53.440 --> 00:27:56.220
but, apparently, she's
got 26 more already.
580
00:27:56.220 --> 00:27:58.900
And Tim Lemper, an
expert on copyright law.
581
00:27:58.900 --> 00:28:02.420
We have Sarah Bellamy, who
comes in and talks about justice
582
00:28:02.420 --> 00:28:06.020
and what she sees from a
humanity standpoint on this.
583
00:28:06.020 --> 00:28:10.080
Alain Barker, who is a
PhD flutist or flautist.
584
00:28:10.080 --> 00:28:11.860
I'm not sure which
way that goes.
585
00:28:11.860 --> 00:28:14.340
And some very, very
special new additions
586
00:28:14.340 --> 00:28:16.500
that we used last
time that were a huge
587
00:28:16.500 --> 00:28:18.880
hit and handle an entire module.
588
00:28:18.880 --> 00:28:22.260
Darin Patterson from Make.ai.
589
00:28:22.260 --> 00:28:25.420
This is one of the engines,
and they have a brand new agent
590
00:28:25.420 --> 00:28:26.480
product out there.
591
00:28:26.480 --> 00:28:28.000
And I know he's excited.
592
00:28:28.000 --> 00:28:29.952
In fact, it wasn't
even released when he
593
00:28:29.952 --> 00:28:31.160
presented at our last course.
594
00:28:31.160 --> 00:28:33.110
So he'll be giving
some new material.
595
00:28:33.110 --> 00:28:34.860
Mark Beitz, who is
Chief Marketing Officer
596
00:28:34.860 --> 00:28:36.900
here in Minnesota at fun.com.
597
00:28:36.900 --> 00:28:39.620
Nufar Gaspar, who
you may know already
598
00:28:39.620 --> 00:28:42.000
from some of her podcasts.
599
00:28:42.000 --> 00:28:45.760
She is the former director
of AI Everywhere for Intel.
600
00:28:45.760 --> 00:28:50.120
Cailin Rogers and Kelly King are
both founders of ad agencies.
601
00:28:50.120 --> 00:28:52.560
Danielle Amfahr, longtime
colleague and former VP
602
00:28:52.560 --> 00:28:54.400
of Strategy from 3M.
603
00:28:54.400 --> 00:28:55.840
And Frank Acito--
and yes, there's
604
00:28:55.840 --> 00:28:59.080
a resemblance there-- professor
emeritus of business analytics
605
00:28:59.080 --> 00:29:02.520
and marketing at
Indiana University.
606
00:29:02.520 --> 00:29:06.840
And a little plug here,
there's Nufar and Nathaniel
607
00:29:06.840 --> 00:29:09.000
from the AI Daily Brief.
608
00:29:09.000 --> 00:29:10.540
There's a link to her podcast.
609
00:29:10.540 --> 00:29:14.440
And she's got, I think, a double
episode out there right now
610
00:29:14.440 --> 00:29:17.520
where she talks about some of
the do's and don'ts and the
611
00:29:17.520 --> 00:29:21.440
risks and what she's seeing in
the field on agentic workflow
612
00:29:21.440 --> 00:29:23.840
and agent deployment.
613
00:29:23.840 --> 00:29:26.547
And one other link, I thought--
again, if you're anxious to get
614
00:29:26.547 --> 00:29:28.880
started and you can't wait
to get into our course, where
615
00:29:28.880 --> 00:29:32.240
you'll learn way
more than this, this
616
00:29:32.240 --> 00:29:37.240
is a link to Make.com, which
is a powerhouse, in my view,
617
00:29:37.240 --> 00:29:40.670
and very, very much
in reach on things.
618
00:29:40.670 --> 00:29:43.090
So with that, Kim,
that's what I had.
619
00:29:43.090 --> 00:29:47.390
I said I'd be done by
1:30 and it's 1:28.
620
00:29:47.390 --> 00:29:48.590
Oh, 1:29.
621
00:29:48.590 --> 00:29:50.210
So I think I did my job.
622
00:29:50.210 --> 00:29:51.990
Why don't you take us home?
623
00:29:51.990 --> 00:29:52.490
OK.
624
00:29:52.490 --> 00:29:53.070
Sure.
625
00:29:53.070 --> 00:29:55.870
Paul, if you don't mind, we
do have a couple of questions
626
00:29:55.870 --> 00:29:58.890
that came in that
I'd love to address,
627
00:29:58.890 --> 00:30:00.790
if you want to cover
them for a few minutes
628
00:30:00.790 --> 00:30:01.890
before we wrap it up.
629
00:30:01.890 --> 00:30:06.030
630
00:30:06.030 --> 00:30:06.870
OK.
631
00:30:06.870 --> 00:30:09.710
So the first
question that came in
632
00:30:09.710 --> 00:30:12.910
was, which industries
are you seeing the most
633
00:30:12.910 --> 00:30:16.790
value from deploying AI agents?
634
00:30:16.790 --> 00:30:18.990
So the way I'd
answer that, Kim, is
635
00:30:18.990 --> 00:30:21.550
that I would look
at it as industry
636
00:30:21.550 --> 00:30:24.510
and I would look
at it as function.
637
00:30:24.510 --> 00:30:28.210
And this is not new thinking,
but it's also somewhat obvious.
638
00:30:28.210 --> 00:30:31.510
So if you look at functions,
the McKinsey reports
639
00:30:31.510 --> 00:30:36.830
and plus what we're seeing, is
that customer service, coding--
640
00:30:36.830 --> 00:30:40.170
obviously, if they're
predicting coding
641
00:30:40.170 --> 00:30:43.630
is going to be hit
with 90% replacement.
642
00:30:43.630 --> 00:30:46.530
Now, again, don't worry
your friends and family
643
00:30:46.530 --> 00:30:48.888
who are studying
coding or are coders.
644
00:30:48.888 --> 00:30:50.430
There's always going
to be-- in fact,
645
00:30:50.430 --> 00:30:55.410
there's a surge of an improved
demand for really good coders.
646
00:30:55.410 --> 00:30:57.450
But those are the
kinds of functions that
647
00:30:57.450 --> 00:31:00.170
are getting treated right now.
648
00:31:00.170 --> 00:31:02.570
It will blanket every function.
649
00:31:02.570 --> 00:31:05.450
And the order in
which that will happen
650
00:31:05.450 --> 00:31:11.530
is, how risky is the
privacy and the security
651
00:31:11.530 --> 00:31:13.370
concern around the data?
652
00:31:13.370 --> 00:31:17.130
So if you're a bank where you
have some banking clients,
653
00:31:17.130 --> 00:31:20.690
you're probably not going
to be experimenting with AI
654
00:31:20.690 --> 00:31:23.770
internally, deploying it
to everyone because you've
655
00:31:23.770 --> 00:31:26.490
got very sensitive information.
656
00:31:26.490 --> 00:31:28.290
If you are a
healthcare organization
657
00:31:28.290 --> 00:31:30.650
with a lot of HIPAA
data, you're probably
658
00:31:30.650 --> 00:31:33.410
not going to be doing a
lot of experimentation.
659
00:31:33.410 --> 00:31:35.410
Will it affect those industries?
660
00:31:35.410 --> 00:31:36.170
Absolutely.
661
00:31:36.170 --> 00:31:38.010
But there's a lot
of work being done
662
00:31:38.010 --> 00:31:41.200
in a number of
different countries
663
00:31:41.200 --> 00:31:42.780
around some of those risks.
664
00:31:42.780 --> 00:31:47.040
So I see that for
now, the types of--
665
00:31:47.040 --> 00:31:50.680
certainly, tech is an early
adopter on all of this stuff.
666
00:31:50.680 --> 00:31:53.240
Anywhere where you
have generative,
667
00:31:53.240 --> 00:31:57.120
predictive, data-driven,
or repetitive workflows,
668
00:31:57.120 --> 00:32:00.360
you're going to see AI.
669
00:32:00.360 --> 00:32:00.880
Fantastic.
670
00:32:00.880 --> 00:32:02.280
Thank you.
671
00:32:02.280 --> 00:32:04.840
One other really great
question that came through
672
00:32:04.840 --> 00:32:07.800
was, can you share
real-world examples
673
00:32:07.800 --> 00:32:12.240
of how AI agents can improve
operational efficiency
674
00:32:12.240 --> 00:32:14.140
and decision-making?
675
00:32:14.140 --> 00:32:14.640
Yeah.
676
00:32:14.640 --> 00:32:17.500
So the way that I look at
operational efficiency,
677
00:32:17.500 --> 00:32:19.740
and for those of you who--
678
00:32:19.740 --> 00:32:21.740
I didn't really spend any
time on my background.
679
00:32:21.740 --> 00:32:23.532
That's not what you
came here, but I'm also
680
00:32:23.532 --> 00:32:25.800
a certified Lean Six
Sigma Master Black
681
00:32:25.800 --> 00:32:29.360
Belt. So everything I do goes
through a filter of cost, cash,
682
00:32:29.360 --> 00:32:30.080
growth.
683
00:32:30.080 --> 00:32:33.960
So I'm constantly looking for,
how do these improve processes
684
00:32:33.960 --> 00:32:36.440
and how do they
improve efficiencies?
685
00:32:36.440 --> 00:32:40.710
So for growth, which is
usually the last place
686
00:32:40.710 --> 00:32:43.150
that these technologies
are applied to,
687
00:32:43.150 --> 00:32:46.710
it just enables you, especially
in marketing, sales, customer
688
00:32:46.710 --> 00:32:50.510
service, to accelerate
your customer engagement
689
00:32:50.510 --> 00:32:52.870
and really double down on it.
690
00:32:52.870 --> 00:32:58.430
For cost-- and it's obvious--
if you're removing or augmenting
691
00:32:58.430 --> 00:33:02.750
humans in their
workflows with AI agents,
692
00:33:02.750 --> 00:33:03.930
you can take costs out.
693
00:33:03.930 --> 00:33:09.710
Now, that agent may in
fact handle and hand off
694
00:33:09.710 --> 00:33:10.750
to the human.
695
00:33:10.750 --> 00:33:12.730
And where we've seen
successful case studies--
696
00:33:12.730 --> 00:33:16.850
and I'm going to draw a blank
on the famous case studies,
697
00:33:16.850 --> 00:33:21.270
but there have been reports of
companies reducing 700 customer
698
00:33:21.270 --> 00:33:22.310
service agents.
699
00:33:22.310 --> 00:33:25.350
I can promise you this is going
to be one of the first areas.
700
00:33:25.350 --> 00:33:31.390
And the reason that is is
because the agentic voice agents
701
00:33:31.390 --> 00:33:35.470
and the workflows that are being
put in place can in many times
702
00:33:35.470 --> 00:33:39.050
improve customer
satisfaction with that as you
703
00:33:39.050 --> 00:33:41.170
remove or augment humans.
704
00:33:41.170 --> 00:33:43.630
So those are big ones.
705
00:33:43.630 --> 00:33:47.230
And then in terms of
cash or cycle time,
706
00:33:47.230 --> 00:33:49.410
just think about response
rates and how much faster
707
00:33:49.410 --> 00:33:51.090
you can get things done.
708
00:33:51.090 --> 00:33:53.870
So with agents, they
don't take Fridays off.
709
00:33:53.870 --> 00:33:55.350
They don't take weekends off.
710
00:33:55.350 --> 00:34:00.550
They don't sleep, so you
can improve things 24/7.
711
00:34:00.550 --> 00:34:06.010
So even with small businesses
like healthcare services,
712
00:34:06.010 --> 00:34:10.170
offices, appointment setting
is a huge application.
713
00:34:10.170 --> 00:34:11.929
Callbacks.
714
00:34:11.929 --> 00:34:14.927
I tend to not recommend
the outbound calls
715
00:34:14.927 --> 00:34:16.969
where certainly you can
do prospecting and things
716
00:34:16.969 --> 00:34:19.090
like that, but there's
probably no better way
717
00:34:19.090 --> 00:34:22.770
to put your customers off than
to have a bot give them a phone
718
00:34:22.770 --> 00:34:23.610
call.
719
00:34:23.610 --> 00:34:25.139
So I think those
are the answers.
720
00:34:25.139 --> 00:34:27.389
Again, that's probably more
than you were looking for,
721
00:34:27.389 --> 00:34:29.310
but in terms of
real-life answers,
722
00:34:29.310 --> 00:34:35.590
I would Google it or frankly, go
ask ChatGPT to search for you.
723
00:34:35.590 --> 00:34:36.090
Great.
724
00:34:36.090 --> 00:34:37.120
Thank you.
725
00:34:37.120 --> 00:34:43.060
Paul, another question regarding
the AI Agent voice recording.
726
00:34:43.060 --> 00:34:45.739
Someone was wondering,
is that an AI agent,
727
00:34:45.739 --> 00:34:48.040
the voice that we heard
in the voice recording?
728
00:34:48.040 --> 00:34:50.600
Was it actually an AI agent,
and can you expand a little bit
729
00:34:50.600 --> 00:34:53.159
about that technology?
730
00:34:53.159 --> 00:34:53.800
Yeah.
731
00:34:53.800 --> 00:34:55.739
So these are agents
that talk to you.
732
00:34:55.739 --> 00:34:58.880
That was a recording because
my colleagues don't like
733
00:34:58.880 --> 00:35:01.960
me to do live demos because--
734
00:35:01.960 --> 00:35:03.260
you ever heard of Murphy's law?
735
00:35:03.260 --> 00:35:06.020
What can go wrong will go wrong?
736
00:35:06.020 --> 00:35:08.320
So that was not a live
demo, but it just as easily
737
00:35:08.320 --> 00:35:09.140
could have been.
738
00:35:09.140 --> 00:35:12.220
So I could give you a
phone number right now.
739
00:35:12.220 --> 00:35:17.320
You could call Bianca, as
we affectionately call her--
740
00:35:17.320 --> 00:35:22.000
it-- and she'll pick up, ask
you a bunch of questions.
741
00:35:22.000 --> 00:35:25.320
And what makes that agentic?
742
00:35:25.320 --> 00:35:27.280
You've already had
these interactions,
743
00:35:27.280 --> 00:35:28.780
and it could have
been a recording.
744
00:35:28.780 --> 00:35:30.360
What makes it
agentic is that she
745
00:35:30.360 --> 00:35:35.240
will change her questioning
based on your instructions
746
00:35:35.240 --> 00:35:39.310
and dig deeper
into the questions.
747
00:35:39.310 --> 00:35:42.870
And one of the challenges
here that we've solved
748
00:35:42.870 --> 00:35:45.950
is they can't ask a lot.
749
00:35:45.950 --> 00:35:47.830
They don't do a very
good job natively
750
00:35:47.830 --> 00:35:51.110
on multiple-choice questions
and scales and things like that.
751
00:35:51.110 --> 00:35:52.950
But on open-ended,
they do a great job.
752
00:35:52.950 --> 00:35:55.870
We've figured out how to do
multiple choice and scales
753
00:35:55.870 --> 00:35:56.870
and things.
754
00:35:56.870 --> 00:35:59.790
But that's reserved for
our clients at this point.
755
00:35:59.790 --> 00:36:06.630
And the idea is that these
expand on that question that's
756
00:36:06.630 --> 00:36:08.270
then transcribed.
757
00:36:08.270 --> 00:36:11.370
I suppose it's recorded too,
but it's really a transcription,
758
00:36:11.370 --> 00:36:13.670
meaning it's voice to text.
759
00:36:13.670 --> 00:36:17.230
That's then put in to
another agentic step, which
760
00:36:17.230 --> 00:36:20.690
is the analysis by a
language model which says,
761
00:36:20.690 --> 00:36:23.110
hey, here's the
sentiment analysis
762
00:36:23.110 --> 00:36:25.690
of how Paul was
feeling about AI.
763
00:36:25.690 --> 00:36:28.950
He was both excited and
not excited about it.
764
00:36:28.950 --> 00:36:33.870
And then it made some decisions
on recommended further reading
765
00:36:33.870 --> 00:36:37.470
for that coursework before
they get to our class.
766
00:36:37.470 --> 00:36:41.250
767
00:36:41.250 --> 00:36:41.910
Wonderful.
768
00:36:41.910 --> 00:36:43.210
Thank you guys.
769
00:36:43.210 --> 00:36:44.630
Any other questions?
770
00:36:44.630 --> 00:36:46.290
Just looking in the
chat real quick.
771
00:36:46.290 --> 00:36:47.350
One second.
772
00:36:47.350 --> 00:36:52.410
773
00:36:52.410 --> 00:36:55.050
Are there any examples
outside of customer
774
00:36:55.050 --> 00:36:56.890
outreach where AI
agents should be
775
00:36:56.890 --> 00:37:01.370
avoided that you can think of?
776
00:37:01.370 --> 00:37:01.870
Oh, boy.
777
00:37:01.870 --> 00:37:03.230
Yeah, I can think of a lot.
778
00:37:03.230 --> 00:37:06.530
779
00:37:06.530 --> 00:37:09.050
There's two levels
to think about this.
780
00:37:09.050 --> 00:37:10.630
It's like the health
of the industry.
781
00:37:10.630 --> 00:37:12.270
If we spoil it before
we get a chance,
782
00:37:12.270 --> 00:37:14.010
we're going to get
regulated out of the--
783
00:37:14.010 --> 00:37:17.890
I mean, it's already very
difficult for those of you
784
00:37:17.890 --> 00:37:20.370
who know about customer
outreach to get access
785
00:37:20.370 --> 00:37:26.230
to these VoIP protocols
and phone numbers
786
00:37:26.230 --> 00:37:29.290
to do outreach or texting.
787
00:37:29.290 --> 00:37:36.280
And I would just urge you not to
spam people with these requests.
788
00:37:36.280 --> 00:37:39.680
But anytime there is PII,
personal identification
789
00:37:39.680 --> 00:37:42.560
information, anytime
there's HIPAA information,
790
00:37:42.560 --> 00:37:47.920
anytime that you have a lot
of ambient, competitive noise
791
00:37:47.920 --> 00:37:50.460
in the background, you're not
going to get a good result.
792
00:37:50.460 --> 00:37:52.200
So there's still a
good deal of work
793
00:37:52.200 --> 00:37:55.940
to be done around the
agent technical capability.
794
00:37:55.940 --> 00:37:59.040
But in terms of use cases, the
way I would look at it, again,
795
00:37:59.040 --> 00:38:02.400
is with that filter of what's
the lowest risk, highest
796
00:38:02.400 --> 00:38:06.600
probability of success, because
you're probably at some point
797
00:38:06.600 --> 00:38:09.120
selling this to
your organization
798
00:38:09.120 --> 00:38:12.500
or to your customers in
terms of how you use it.
799
00:38:12.500 --> 00:38:18.680
But for generative, repetitive,
data-driven, and predictive
800
00:38:18.680 --> 00:38:23.760
workflows, this can
be very powerful.
801
00:38:23.760 --> 00:38:25.440
Great, great.
802
00:38:25.440 --> 00:38:26.000
All right.
803
00:38:26.000 --> 00:38:29.360
I think that's most
of the questions.
804
00:38:29.360 --> 00:38:29.940
One more.
805
00:38:29.940 --> 00:38:33.460
What are your thoughts on AI
and the augmented connected
806
00:38:33.460 --> 00:38:34.280
workforce?
807
00:38:34.280 --> 00:38:38.180
808
00:38:38.180 --> 00:38:38.700
Yeah.
809
00:38:38.700 --> 00:38:41.363
810
00:38:41.363 --> 00:38:42.780
The first thing
is, if you haven't
811
00:38:42.780 --> 00:38:47.380
tried some of these
note-taking AIs like Fireflies,
812
00:38:47.380 --> 00:38:52.720
Otter, and others, literally,
you can agentically just say,
813
00:38:52.720 --> 00:38:54.260
attend all my meetings.
814
00:38:54.260 --> 00:38:58.700
I had one where I clicked
the wrong menu button
815
00:38:58.700 --> 00:39:01.800
and I got a call while I
was doing something else.
816
00:39:01.800 --> 00:39:05.720
And they said, it shows
that you're on the meeting,
817
00:39:05.720 --> 00:39:07.400
but you don't
appear to be there.
818
00:39:07.400 --> 00:39:08.200
What's happening?
819
00:39:08.200 --> 00:39:11.360
So my Fireflies app
attended a meeting for me,
820
00:39:11.360 --> 00:39:14.500
took and summarized
notes, and sent it to me.
821
00:39:14.500 --> 00:39:17.700
You can literally be in
more than one place at once.
822
00:39:17.700 --> 00:39:22.340
So for remote work,
augmented work--
823
00:39:22.340 --> 00:39:26.220
mark my words, you will be
going into ChatGPT or Claude
824
00:39:26.220 --> 00:39:29.500
or Copilot in 12 to 18
months more than you're
825
00:39:29.500 --> 00:39:31.960
going into Microsoft Office.
826
00:39:31.960 --> 00:39:37.090
827
00:39:37.090 --> 00:39:39.610
This will be a tool-- just like
as you moved up the learning
828
00:39:39.610 --> 00:39:42.570
curve to expert level
on Excel and PowerPoint
829
00:39:42.570 --> 00:39:45.250
and Word and your company ERP.
830
00:39:45.250 --> 00:39:47.490
This will be a similar
kind of a lever
831
00:39:47.490 --> 00:39:51.530
in your day-to-day productivity.
832
00:39:51.530 --> 00:39:52.070
Great.
833
00:39:52.070 --> 00:39:53.410
Thank you so much.
834
00:39:53.410 --> 00:39:53.910
All right.
835
00:39:53.910 --> 00:39:56.810
So I think that's
most of the questions.
836
00:39:56.810 --> 00:39:58.410
You all can feel
free to email me
837
00:39:58.410 --> 00:40:00.170
if you have any
specific questions
838
00:40:00.170 --> 00:40:02.910
or if you need
additional information.
839
00:40:02.910 --> 00:40:05.250
We'll do our best
to provide that.
840
00:40:05.250 --> 00:40:08.790
But I just wanted to thank
you all for attending today.
841
00:40:08.790 --> 00:40:10.890
We hope you enjoyed
this session.
842
00:40:10.890 --> 00:40:12.110
My name is Kim Allison.
843
00:40:12.110 --> 00:40:14.847
I'm with the Kelley School of
Business Executive Education.
844
00:40:14.847 --> 00:40:16.430
If you're not familiar
with our group,
845
00:40:16.430 --> 00:40:19.130
we provide a wide range
of courses and certificate
846
00:40:19.130 --> 00:40:24.170
programs designed for working
professionals and organizations.
847
00:40:24.170 --> 00:40:26.690
Our courses cover
a number of topics
848
00:40:26.690 --> 00:40:29.650
such as AI, leadership
and management,
849
00:40:29.650 --> 00:40:33.120
operational excellence,
finance, and more.
850
00:40:33.120 --> 00:40:35.960
If you have any questions,
you can reach out to me.
851
00:40:35.960 --> 00:40:38.240
If you want to learn more
about the AI applications
852
00:40:38.240 --> 00:40:40.360
for marketing course,
which Paul was discussing
853
00:40:40.360 --> 00:40:42.600
a little bit during
this session,
854
00:40:42.600 --> 00:40:45.440
we have a QR code
here that you can
855
00:40:45.440 --> 00:40:47.920
scan for detailed information.
856
00:40:47.920 --> 00:40:50.320
But it launches May 6.
857
00:40:50.320 --> 00:40:54.560
It's completely online with
meetings Tuesdays and Thursday
858
00:40:54.560 --> 00:40:55.300
evenings.
859
00:40:55.300 --> 00:40:58.440
It's instructor
led, 6 to 7:30 p.m.
860
00:40:58.440 --> 00:41:01.440
We have about 10 seats
left for this session.
861
00:41:01.440 --> 00:41:05.078
And we'll be opening
another one in the fall too.
862
00:41:05.078 --> 00:41:06.620
If you have any
questions about that,
863
00:41:06.620 --> 00:41:09.040
feel free to reach
out to either of us.
864
00:41:09.040 --> 00:41:12.080
For attending this
session, we are giving you
865
00:41:12.080 --> 00:41:16.720
guys a discount code to
register for the course.
866
00:41:16.720 --> 00:41:19.665
So I'll put that into the
chat, and that will give you
867
00:41:19.665 --> 00:41:21.040
a little bit of
a discount if you
868
00:41:21.040 --> 00:41:24.230
want to register ahead of time.
869
00:41:24.230 --> 00:41:26.480
I'm trying to think if there's
any other information I
870
00:41:26.480 --> 00:41:27.400
can share.
871
00:41:27.400 --> 00:41:32.200
Any other questions
before we wrap it up?
872
00:41:32.200 --> 00:41:34.720
Kim, I'd just like to
give a special welcome.
873
00:41:34.720 --> 00:41:37.960
We tend to overindex on
IU and Kelley alumni.
874
00:41:37.960 --> 00:41:40.260
Thank you, everyone, for coming.
875
00:41:40.260 --> 00:41:43.300
And if you've got
any questions, you
876
00:41:43.300 --> 00:41:47.700
can reach out to me on LinkedIn
or, obviously, reach out to Kim.
877
00:41:47.700 --> 00:41:49.400
This is a great
course coming up.
878
00:41:49.400 --> 00:41:50.940
We have a lot of fun.
879
00:41:50.940 --> 00:41:53.740
And believe it or not,
a lot of our students
880
00:41:53.740 --> 00:41:56.678
say, when can we do this again?
881
00:41:56.678 --> 00:41:57.720
They don't want to leave.
882
00:41:57.720 --> 00:41:59.940
Boy, that went by
fast, which is,
883
00:41:59.940 --> 00:42:01.680
in my view, a very
high compliment.
884
00:42:01.680 --> 00:42:03.580
So thank you all for
your kind attention,
885
00:42:03.580 --> 00:42:07.980
and hope to see you
in the online course.
886
00:42:07.980 --> 00:42:08.480
Great.
887
00:42:08.480 --> 00:42:08.960
Thank you all.
888
00:42:08.960 --> 00:42:11.543
If you have any questions, we'll
hang around for a few minutes
889
00:42:11.543 --> 00:42:13.400
so that we can answer
as best we can.
890
00:42:13.400 --> 00:42:15.020
But thank you all
for your time today,
891
00:42:15.020 --> 00:42:17.940
and we appreciate
you being here.
892
00:42:17.940 --> 00:42:18.520
Take care.
893
00:42:18.520 --> 00:42:20.070
Bye.
894
00:42:20.070 --> 00:42:32.000
