WEBVTT

1
00:00:00.240 --> 00:00:03.840
European retail is in a margin war. It didn't choose.

2
00:00:04.080 --> 00:00:07.760
Assortments move faster than teams can analyze data

3
00:00:07.760 --> 00:00:11.520
quality collapses on the scale and global platforms like

4
00:00:11.520 --> 00:00:15.200
Amazon and Temu operate on a level of speed that

5
00:00:15.200 --> 00:00:17.920
exposes every structural weakness.

6
00:00:18.400 --> 00:00:22.000
Abba SETI built Zenline AI

7
00:00:22.000 --> 00:00:25.840
to close that gap. Agents that act not

8
00:00:25.840 --> 00:00:29.160
observe systems that detect margin leaks,

9
00:00:29.480 --> 00:00:32.680
resolve clutter and surface opportunities

10
00:00:33.000 --> 00:00:36.360
humans never see. Today we examine why

11
00:00:36.440 --> 00:00:39.960
autonomous decision agents are becoming to operate the system

12
00:00:40.200 --> 00:00:43.720
of modern retail. And what happens when a retailer

13
00:00:43.800 --> 00:00:47.400
shifts from dashboard analyzers to product level action

14
00:00:47.400 --> 00:00:50.960
in seconds. By the end, you'll understand why the next

15
00:00:50.960 --> 00:00:54.340
competitive edge isn't scale its velocity.

16
00:01:00.180 --> 00:01:03.220
Welcome to StartUpLead IO,

17
00:01:03.780 --> 00:01:07.300
your podcast and YouTube blog covering the German

18
00:01:07.300 --> 00:01:10.820
startup scene with news interviews and

19
00:01:11.060 --> 00:01:12.260
live events.

20
00:01:14.580 --> 00:01:18.420
Hello and welcome everybody to Startup Rio. Today my guest is

21
00:01:18.420 --> 00:01:21.620
Arba Seti, the CEO and co founder of

22
00:01:21.620 --> 00:01:24.620
Zenline AI from Switzerland, Goodsy

23
00:01:25.500 --> 00:01:29.260
Goody, a company that is redefining how European

24
00:01:29.260 --> 00:01:32.860
retailers decide what to sell, what to move and where

25
00:01:32.940 --> 00:01:36.660
the real margin sits. With experience at BCG,

26
00:01:36.660 --> 00:01:40.379
Procter Gamble, ETH Zurich and Research work at

27
00:01:40.379 --> 00:01:44.140
Cambridge, he brings both the strategic clarity and the

28
00:01:44.140 --> 00:01:47.660
technical depth required to rebuild assortment logic.

29
00:01:47.980 --> 00:01:51.020
From first principles. Zenline agents

30
00:01:51.570 --> 00:01:55.170
clean messy catalog data, detect cannibalization,

31
00:01:55.650 --> 00:01:59.250
map substitutes and reveal trends before they

32
00:01:59.250 --> 00:02:02.850
become mainstream. They already process more than a million

33
00:02:03.090 --> 00:02:06.850
SKUs and deliver recommendations with measurable

34
00:02:06.850 --> 00:02:10.410
financial impact enable retailers to operate with the

35
00:02:10.410 --> 00:02:13.810
speed of global platforms without the overhead.

36
00:02:14.050 --> 00:02:17.890
Aber is here because his work signals a structural change.

37
00:02:18.520 --> 00:02:22.080
Decisions that took once weeks now happen in

38
00:02:22.080 --> 00:02:25.600
minutes and the retailers who adapt will redefine

39
00:02:25.600 --> 00:02:29.080
competitiveness in Europe's Euro trillion

40
00:02:29.400 --> 00:02:33.200
market. That was an intro, don't you

41
00:02:33.200 --> 00:02:36.840
think? Welcome to the show and let us

42
00:02:37.240 --> 00:02:40.760
get right into it. You said the real power

43
00:02:41.000 --> 00:02:44.770
center of retail isn't price or promotion, but

44
00:02:44.840 --> 00:02:48.160
but the assortment itself. When did that

45
00:02:48.160 --> 00:02:51.840
insight crystallize for you? In my previous

46
00:02:51.840 --> 00:02:55.640
job with bcg, we were advising retailers with assortment optimization.

47
00:02:55.640 --> 00:02:59.400
And what becomes very obvious when you

48
00:02:59.400 --> 00:03:03.000
work with retailers on the assortment is that the assortment touches the entire

49
00:03:03.000 --> 00:03:06.680
organization. So it's not only the top line or the bottom line as is

50
00:03:06.680 --> 00:03:08.680
for pricing and for promotion, but

51
00:03:10.200 --> 00:03:13.920
certainly it also touches the top line. So if a customer, for example,

52
00:03:13.920 --> 00:03:17.500
doesn't find a product that it wants in the stor stores or on the

53
00:03:17.500 --> 00:03:21.100
shelves, that's lost revenue. However. Moreover, it

54
00:03:21.100 --> 00:03:24.620
really defines your brand. So the brand is typically

55
00:03:24.620 --> 00:03:28.260
not only shaped by the marketing, but

56
00:03:28.260 --> 00:03:32.060
it's mostly shaped by the products you offer. So you can imagine when

57
00:03:32.060 --> 00:03:35.820
a consumer has a demand, what do they first Think of what's the

58
00:03:35.820 --> 00:03:39.460
brand, what's the retailer that they think of to go and

59
00:03:39.460 --> 00:03:43.180
buy the product from? And that's so important to be top of mind for the

60
00:03:43.180 --> 00:03:46.840
consumers. And, and that only works with an assortment. That consistency

61
00:03:46.840 --> 00:03:50.480
delivers what the consumers actually want. And then the

62
00:03:50.480 --> 00:03:54.320
last one is the assortment. And the depth of the, and the

63
00:03:54.320 --> 00:03:57.480
breadth of the assortment influences the

64
00:03:57.560 --> 00:04:01.120
complexity of your operations, the logistics, the supply chain, the

65
00:04:01.120 --> 00:04:04.840
inventory management. So the broader the assortment,

66
00:04:05.320 --> 00:04:09.000
the higher the complexity and the more difficult it is to maintain

67
00:04:09.000 --> 00:04:12.280
lean operations. When you talk about

68
00:04:12.280 --> 00:04:15.440
assortment, you have something in mind. For example, when I was

69
00:04:16.560 --> 00:04:20.280
thinking about this interview, what I had in mind when I was reading this,

70
00:04:20.280 --> 00:04:23.040
at first I was thinking, okay,

71
00:04:23.840 --> 00:04:26.960
option one is I look for the

72
00:04:27.120 --> 00:04:30.920
basics. That is a retailer with a yellow logo. Then I look

73
00:04:30.920 --> 00:04:34.400
for some more fancy stuff. Then that comes in a

74
00:04:34.400 --> 00:04:38.160
retailer with the orange logo or a red logo.

75
00:04:38.160 --> 00:04:41.970
Is that something that you're thinking about here? Yes,

76
00:04:41.970 --> 00:04:45.810
exactly, that's exactly it. What comes to your mind when you want to

77
00:04:45.810 --> 00:04:49.610
buy a product, what is it? Is it someone who's super, super trendy

78
00:04:50.250 --> 00:04:54.090
and always knows what the consumers want and can offer it, offers it

79
00:04:54.090 --> 00:04:57.810
fast, or is it typically 6 to 12 months behind

80
00:04:57.810 --> 00:05:01.530
the schedule? Your early

81
00:05:01.850 --> 00:05:05.130
discovery work exposed how little transparency

82
00:05:05.370 --> 00:05:09.050
retailers had into their own margin drivers, what

83
00:05:09.050 --> 00:05:12.300
revealed the debt of that blind spot.

84
00:05:13.180 --> 00:05:16.860
So the importance of that blind spot comes basically

85
00:05:16.860 --> 00:05:20.220
what you said in your intro. It's a super, super, super competitive market

86
00:05:20.460 --> 00:05:24.220
with Temu and Shein entering the European market not only in

87
00:05:24.220 --> 00:05:28.020
fashion, which a lot of people think, but it's across all categories from electronics to

88
00:05:28.020 --> 00:05:31.580
fashion to pet supplies. And retailers

89
00:05:31.580 --> 00:05:35.300
are too slow because of the analysis, just take too

90
00:05:35.300 --> 00:05:39.040
long, which should be done automatically and systematically with

91
00:05:39.040 --> 00:05:42.600
AI. So basically we're entering

92
00:05:42.600 --> 00:05:46.240
this competitive, basically war

93
00:05:46.320 --> 00:05:49.840
in retail coming from different countries from different angles.

94
00:05:50.320 --> 00:05:54.080
And if retailers don't move fast

95
00:05:54.080 --> 00:05:57.600
enough, don't filter out products that have bad

96
00:05:57.600 --> 00:06:01.400
margins and focus on the ones, focus on the entire assortment. Instead

97
00:06:01.400 --> 00:06:05.240
of managing just 10 to 20%, European retailers will

98
00:06:05.240 --> 00:06:08.580
not be able to cope with the speed and with the

99
00:06:08.580 --> 00:06:10.780
velocity of these big platform players.

100
00:06:13.100 --> 00:06:16.900
So that you're mostly targeting

101
00:06:16.900 --> 00:06:20.340
this towards retailers that would

102
00:06:20.340 --> 00:06:23.260
compete with Amazon, Temu, Shine,

103
00:06:24.220 --> 00:06:26.380
Alibaba Express and so on and so forth.

104
00:06:27.900 --> 00:06:31.420
Yes, we're trying to basically help them to have the

105
00:06:31.820 --> 00:06:35.500
same data, science and AI capabilities as

106
00:06:35.500 --> 00:06:39.260
these platforms have, who have actually their own internal assortment

107
00:06:39.260 --> 00:06:42.920
teams to make these decisions. And you're

108
00:06:42.920 --> 00:06:46.720
now entering a market that is basically shaped

109
00:06:46.720 --> 00:06:50.440
by Amazon level speed. You order one day, you get it,

110
00:06:50.440 --> 00:06:54.280
the Next day or a little bit later. What convinced you that

111
00:06:54.280 --> 00:06:57.680
Europe's response had to be agentic and not

112
00:06:57.680 --> 00:07:01.400
incremental? So what's happening right now is a perfect storm.

113
00:07:01.720 --> 00:07:04.840
We're seeing years of inflation, declining consumer confidence,

114
00:07:05.160 --> 00:07:08.990
increased saving rates and what's making it now

115
00:07:08.990 --> 00:07:12.750
very hard for retailers to stay ahead the curve and stay

116
00:07:12.750 --> 00:07:16.390
profitable. And this does not only apply to

117
00:07:16.390 --> 00:07:20.150
retail, right? We're speaking about all industries that are

118
00:07:20.150 --> 00:07:23.670
basically undergoing this wave of the industrial

119
00:07:23.990 --> 00:07:27.630
revolution or let's say revolution led by general purpose

120
00:07:27.630 --> 00:07:31.230
technologies. And in these times of general purpose

121
00:07:31.230 --> 00:07:34.390
technology is changing not only one but multitude of industries,

122
00:07:35.650 --> 00:07:39.410
the answer has never been incremental innovation. And US and China will

123
00:07:39.410 --> 00:07:43.170
move ahead and they will basically equip millions of white

124
00:07:43.170 --> 00:07:46.650
collar workers with agentic AI. And speed will be the

125
00:07:46.650 --> 00:07:50.450
differentiating factor, launching new products, understanding

126
00:07:50.530 --> 00:07:54.090
the assortment, understanding relationships in the assortments much, much, much

127
00:07:54.090 --> 00:07:57.810
deeper, much earlier than all competitors will.

128
00:07:59.010 --> 00:08:02.760
And again just to to basically highlight that if we

129
00:08:02.760 --> 00:08:06.440
still think we can do the same, the

130
00:08:06.440 --> 00:08:10.200
same thing that we did for the past 20 years, which is incrementally

131
00:08:11.080 --> 00:08:14.520
increase efficiency, these companies will

132
00:08:14.840 --> 00:08:18.520
basically frog leap the European retail industry.

133
00:08:21.240 --> 00:08:24.920
In the entrepreneur's world, we'll be talked

134
00:08:24.920 --> 00:08:28.570
about the moment that data forced a decision

135
00:08:28.890 --> 00:08:32.410
you didn't expect. Stay tuned and learn

136
00:08:32.410 --> 00:08:34.570
more down here in the link.

137
00:08:36.570 --> 00:08:40.370
We've been talking category teams already that some of our large

138
00:08:40.370 --> 00:08:43.810
competitors have. They focus on top

139
00:08:43.810 --> 00:08:47.650
sellers and ignore the long tail. What breaks inside

140
00:08:47.650 --> 00:08:51.410
an organization when the long tail suddenly becomes

141
00:08:51.410 --> 00:08:55.170
visible through good data, through AI, real time

142
00:08:55.170 --> 00:08:58.810
observations? I would say the first reaction is

143
00:08:58.810 --> 00:09:02.610
inconvenience. There's a lot of things that get exposed, like

144
00:09:02.610 --> 00:09:06.250
data flaws, wrong images linked to wrong products which

145
00:09:06.250 --> 00:09:10.010
then lead to very high returns because consumers got the

146
00:09:10.010 --> 00:09:13.330
wrong product which they actually didn't want to order. But the product

147
00:09:13.890 --> 00:09:16.370
image on the website was leading to the wrong product

148
00:09:17.330 --> 00:09:20.730
attributes that are misleading or misconfigured or inconsistent

149
00:09:20.730 --> 00:09:24.440
pricing and all of these errors, or

150
00:09:24.520 --> 00:09:27.560
let's say inconsistencies,

151
00:09:27.640 --> 00:09:30.600
misconfigurations in the assortment become

152
00:09:31.240 --> 00:09:34.400
very obvious when you start focusing not only on the

153
00:09:34.400 --> 00:09:37.920
1015 to 20% of the assortment that is typically

154
00:09:37.920 --> 00:09:41.320
managed very well, but then

155
00:09:41.560 --> 00:09:44.680
surfaces and this then creates, let's say

156
00:09:45.480 --> 00:09:48.600
inconvenience at the beginning. But

157
00:09:49.030 --> 00:09:52.630
organizations then learn to accept the fact that it was

158
00:09:52.630 --> 00:09:56.430
very, very hard to manually be able to manage

159
00:09:56.430 --> 00:10:00.230
hundreds of thousands, if not millions of SKUs. But now with the

160
00:10:00.230 --> 00:10:04.070
power of AI, you can actually do what retail

161
00:10:05.110 --> 00:10:08.950
wasn't able to do for the past. Basically for the past decades.

162
00:10:11.510 --> 00:10:14.990
You'Re the agents you provide for the companies, they do

163
00:10:14.990 --> 00:10:18.680
challenge internal narratives. Which narrative

164
00:10:19.000 --> 00:10:22.800
collapses the fastest once real data

165
00:10:22.800 --> 00:10:26.400
pressure was applied? You said first the flaws are exposed and

166
00:10:26.400 --> 00:10:30.160
then I think the next step is their internal narratives breaking

167
00:10:30.160 --> 00:10:33.960
down. Yeah, very, very, very good question. And I would

168
00:10:33.960 --> 00:10:37.720
say the most common narrative is the intern.

169
00:10:37.720 --> 00:10:41.440
The most common internal narrative is that the consumers know

170
00:10:41.440 --> 00:10:45.120
exactly what they want to buy. And this consequentially leads to

171
00:10:45.120 --> 00:10:48.820
an increased breadth of the assortment. And we as read retailers,

172
00:10:48.820 --> 00:10:52.580
we don't have power to because the consumers have all the power

173
00:10:52.580 --> 00:10:56.220
and they will either way just buy what they want. If you don't have

174
00:10:56.220 --> 00:10:59.860
it, they will not buy. However, data that very, very

175
00:11:00.500 --> 00:11:04.100
early on will show that curated

176
00:11:04.100 --> 00:11:07.700
assortments is actually what consumers want. They don't want

177
00:11:07.700 --> 00:11:10.980
decision paralysis, but what they want is a curated

178
00:11:10.980 --> 00:11:14.500
assortment. Exactly. Curated to their

179
00:11:14.500 --> 00:11:18.170
demands and their basically the consumer behavior. And

180
00:11:18.170 --> 00:11:21.570
what you typically see immediately is that revenue and margin

181
00:11:21.810 --> 00:11:25.170
goes up significantly. We're speaking 1, 2, 3 percentage points

182
00:11:25.650 --> 00:11:28.930
of margin when the

183
00:11:28.930 --> 00:11:32.770
retailer makes the decision of what should be in the store, in the stores

184
00:11:32.930 --> 00:11:36.730
instead of basically giving up and saying it's always, it's the

185
00:11:36.730 --> 00:11:40.530
consumer who's always right and we need to offer everything that they

186
00:11:40.530 --> 00:11:41.810
can possibly imagine.

187
00:11:46.300 --> 00:11:49.820
Part of your thesis or your main

188
00:11:49.820 --> 00:11:53.260
part is speed. What's the cost of

189
00:11:53.260 --> 00:11:56.620
operating a full year behind the real markets decision

190
00:11:56.620 --> 00:12:00.340
Tempo. I just mentioned it. We're speaking 1, 2 percentage

191
00:12:00.340 --> 00:12:03.780
points in margin improvement and we're speaking 1 to 3% in

192
00:12:03.780 --> 00:12:07.260
revenue increase. Very often we're not even

193
00:12:07.340 --> 00:12:10.800
one year behind the real market, especially in the private

194
00:12:10.800 --> 00:12:14.560
label segment of the businesses. And in Switzerland, I think for the

195
00:12:14.560 --> 00:12:18.160
first time, private label revenues surpassed 50% of the market.

196
00:12:18.160 --> 00:12:21.840
So private label is now making more revenue than the branded market

197
00:12:21.840 --> 00:12:25.559
for retailers. And in Germany, in the rest of Europe,

198
00:12:25.559 --> 00:12:28.920
I think it's close to between 30 to 50%.

199
00:12:29.720 --> 00:12:33.400
And there we're speaking very often about 18 to 24 months

200
00:12:34.040 --> 00:12:37.770
of let's say distance to real consumer,

201
00:12:37.770 --> 00:12:41.530
consumer changes. I do have

202
00:12:41.530 --> 00:12:45.370
in mind that basically we talked about here the

203
00:12:45.370 --> 00:12:48.810
less fancy products, especially talked about food here

204
00:12:48.970 --> 00:12:52.650
and clothing. There's a big online retailer who offers

205
00:12:52.650 --> 00:12:56.450
essentials and, and that's kind of stuff we talked about here,

206
00:12:56.450 --> 00:13:00.210
that that's retail brands, that's the brands they come up with themselves

207
00:13:00.210 --> 00:13:04.010
and for some larger specialty food retailer, they are meant by, they

208
00:13:04.010 --> 00:13:07.510
are produced by household names. So that' so

209
00:13:07.590 --> 00:13:11.270
pretty interesting area to watch. But let us go a little bit further

210
00:13:11.350 --> 00:13:12.470
and talk about

211
00:13:15.190 --> 00:13:18.750
what impact the data had on teams because you've Watched

212
00:13:18.750 --> 00:13:22.070
messy data, paralyzed teams. What's the silent

213
00:13:22.310 --> 00:13:24.790
limiter they underestimate most?

214
00:13:29.350 --> 00:13:32.310
So I would like to answer slightly

215
00:13:33.120 --> 00:13:36.960
different question which is, and I'll get back to yours, which is

216
00:13:37.760 --> 00:13:41.440
companies very often think that data overall

217
00:13:41.520 --> 00:13:45.280
is the limiter of, for the generation of the value

218
00:13:46.480 --> 00:13:50.320
margin, improvement, revenue increases, making faster

219
00:13:50.320 --> 00:13:54.040
decisions, and then they very often focus on the 5

220
00:13:54.040 --> 00:13:57.760
to 10% of the data that is imperfect, where the master data

221
00:13:57.760 --> 00:14:01.460
is not clean, where the attribute data is not clean. And then

222
00:14:01.460 --> 00:14:05.300
organizations often stagnate and move very, very, very slow

223
00:14:06.020 --> 00:14:09.140
because they think of, oh, data cleaning being this one

224
00:14:09.780 --> 00:14:13.540
incredibly big IT exercise that they need to perform.

225
00:14:13.780 --> 00:14:16.260
However, what should happen is that

226
00:14:17.300 --> 00:14:21.020
while doing these use cases, you clean the data as

227
00:14:21.020 --> 00:14:24.580
you need them because you will figure out where the data leads to. Wrong

228
00:14:24.900 --> 00:14:28.660
suggestions, wrong recommendations, and then it

229
00:14:28.660 --> 00:14:32.500
makes sense. And then there's also a business case to clean up

230
00:14:32.500 --> 00:14:35.980
the data and make it worth this exercise.

231
00:14:36.140 --> 00:14:39.980
And what a lot of companies and retailers have not realized yet

232
00:14:39.980 --> 00:14:43.820
is that agentic AI is making data cleaning and insight generation

233
00:14:44.140 --> 00:14:47.580
so feasible. The cost to or the cost and the time

234
00:14:48.060 --> 00:14:51.420
to get cleaner data is reduced on a daily basis

235
00:14:51.740 --> 00:14:55.220
and retailers can already save millions in margin

236
00:14:55.620 --> 00:14:59.420
without having the perfect data. So instead of focusing too much on the

237
00:14:59.420 --> 00:15:03.140
limiter, I think companies do that either way all the time.

238
00:15:03.700 --> 00:15:07.460
The message should be although data can be messy

239
00:15:07.460 --> 00:15:11.140
and too much data can paralyze, the

240
00:15:11.300 --> 00:15:14.420
limiter is very often the mindset towards the data.

241
00:15:19.300 --> 00:15:22.970
When we talk about agents, many people

242
00:15:23.210 --> 00:15:26.650
have absolute neutral

243
00:15:27.850 --> 00:15:31.530
AI in mind. But every agent encodes

244
00:15:31.770 --> 00:15:35.570
a judgment, either by the person who decides what to

245
00:15:35.570 --> 00:15:39.370
encode or the person who does the encoding. Which

246
00:15:39.450 --> 00:15:42.890
judgment was the most difficult to translate into

247
00:15:43.210 --> 00:15:44.970
an autonomous logic?

248
00:15:47.700 --> 00:15:50.100
Very good question. And I have to say

249
00:15:51.780 --> 00:15:55.620
the entire space is very, very, very early and it will change in a

250
00:15:55.620 --> 00:15:56.580
year from now. But

251
00:15:59.300 --> 00:16:03.100
what is happening right now is that agents are very, very, very

252
00:16:03.100 --> 00:16:06.740
sensitive to a few sentences in the prompt in how they

253
00:16:06.740 --> 00:16:10.380
make the final decision. So because of

254
00:16:10.380 --> 00:16:13.780
that, as you were saying, rightfully, the

255
00:16:13.940 --> 00:16:17.740
design team of the agent, of the system prompt of

256
00:16:17.740 --> 00:16:21.460
the context engineering that allows the agent to make a decision

257
00:16:21.460 --> 00:16:25.060
of which products to recommend. One very

258
00:16:25.060 --> 00:16:28.500
simple example might be that one of our customers

259
00:16:28.740 --> 00:16:32.420
doesn't want to be the most trendy firm. So they don't want to just launch

260
00:16:32.420 --> 00:16:36.220
products that are, let's say, that

261
00:16:36.220 --> 00:16:40.020
are just coming to the market, but they want some validation, some proof points,

262
00:16:40.250 --> 00:16:43.920
and to know that this product is actually performing and will be here two, three,

263
00:16:43.920 --> 00:16:47.520
four years from now. And now the question is, what is A trendy product,

264
00:16:47.520 --> 00:16:50.880
when does the trend turn into something more sustainable?

265
00:16:51.520 --> 00:16:55.320
And in the end, the makers, which we are

266
00:16:55.320 --> 00:16:58.880
here with Zenline, need to be very, very, very thoughtful

267
00:16:58.880 --> 00:17:02.480
of how do we encode that? And I think

268
00:17:03.040 --> 00:17:06.880
the most difficult part of this is the memory layer. So

269
00:17:06.880 --> 00:17:10.660
you do one recommendation and then the question is next week,

270
00:17:10.660 --> 00:17:14.380
next month, next quarter, how do you encode all the recommendations

271
00:17:14.380 --> 00:17:17.940
that you made? How do you encode, let's say, the

272
00:17:17.940 --> 00:17:21.700
feedback from the real world? Did the product launchers

273
00:17:21.700 --> 00:17:25.380
succeed or didn't they succeed? How do you bring that in and how

274
00:17:25.380 --> 00:17:29.220
do you feedback that into something that the agent

275
00:17:29.220 --> 00:17:32.740
can use to then be able to make better

276
00:17:32.740 --> 00:17:36.530
recommendations next time? Because this is something that is incredibly natural to

277
00:17:36.530 --> 00:17:40.250
us as humans. We do something, we, we observe what happens

278
00:17:40.330 --> 00:17:43.530
and we have an initial gut feeling of whether we should do that again or

279
00:17:43.530 --> 00:17:46.770
what we need to adjust. And this is something that's very hard to, to encode

280
00:17:46.770 --> 00:17:50.170
into. Into autonomous systems. Yes,

281
00:17:51.050 --> 00:17:54.690
there's a lot of stuff to really hard to encode in

282
00:17:54.690 --> 00:17:58.450
autonomous systems. And I think the more systems

283
00:17:58.450 --> 00:18:02.300
we'll have, the more it will be

284
00:18:02.300 --> 00:18:06.100
of importance. But let us

285
00:18:06.580 --> 00:18:10.260
go to the next question, because I was wondering,

286
00:18:11.460 --> 00:18:14.340
your launch validator runs on social signals,

287
00:18:14.580 --> 00:18:18.260
competitive shifts and emerging demand. When

288
00:18:18.260 --> 00:18:22.020
did you trust its signals over the human

289
00:18:22.020 --> 00:18:25.780
instinct? So what's

290
00:18:25.780 --> 00:18:29.490
important to distinguish is whenever you launch a product,

291
00:18:29.490 --> 00:18:32.850
which is what we aim to support with the launch validator

292
00:18:33.410 --> 00:18:37.250
is we want to strengthen the human

293
00:18:37.250 --> 00:18:40.770
instincts by giving it more data. However,

294
00:18:41.570 --> 00:18:45.210
as every single human being, category managers, for example, in the

295
00:18:45.210 --> 00:18:48.970
fashion world, are biased by what they see on their

296
00:18:48.970 --> 00:18:52.530
channels, their social media channels, by walking through the city, watching

297
00:18:52.530 --> 00:18:56.110
colleagues, friends and so on, what they wear, and

298
00:18:56.110 --> 00:18:59.870
then using hundreds of thousands of data points

299
00:19:00.510 --> 00:19:03.950
actually allows them to double check your gut feelings

300
00:19:04.350 --> 00:19:08.190
and use these data points from social media, from Google shopping,

301
00:19:08.190 --> 00:19:11.550
from competitors, bestseller data, to see whether

302
00:19:11.870 --> 00:19:14.270
your initial gut feeling was right

303
00:19:15.630 --> 00:19:19.110
or whether you were missing something. And it can be much, much, much more

304
00:19:19.110 --> 00:19:22.930
nuanced, especially in the fashion world, where new designs

305
00:19:22.930 --> 00:19:26.130
from quarter to quarter, from year to year very often change

306
00:19:27.010 --> 00:19:30.690
in subtle ways which are very hard to grasp. So it's not

307
00:19:31.890 --> 00:19:35.610
trusting the signals over human instinct, but it is challenging

308
00:19:35.610 --> 00:19:39.410
the human instinct, which can be biased by their own world, which is very often

309
00:19:39.410 --> 00:19:43.130
not representative of, let's say, the

310
00:19:43.130 --> 00:19:46.610
consumers, what they want in certain categories. Always,

311
00:19:47.980 --> 00:19:51.580
always when it's working in consulting and

312
00:19:51.980 --> 00:19:55.820
always if you do have strategy project at one point, there's always

313
00:19:56.060 --> 00:19:59.780
the question of internal cannibalization. So

314
00:19:59.780 --> 00:20:03.260
if the, if I'M adding product

315
00:20:03.420 --> 00:20:06.860
A, does it take away from B, does it take away from C?

316
00:20:07.260 --> 00:20:10.700
And your consolidation agent exposes

317
00:20:10.780 --> 00:20:13.820
internal cannibalization. What's the most

318
00:20:14.370 --> 00:20:17.450
counterintuitive callization call that

319
00:20:17.450 --> 00:20:20.770
surfaced? There's always revenue

320
00:20:20.770 --> 00:20:24.330
cannibalization and there's margin cannibalization. And the

321
00:20:24.330 --> 00:20:27.970
revenue cannibalization is basically where a lower priced product

322
00:20:28.290 --> 00:20:31.490
is purchased by the consumer because

323
00:20:31.810 --> 00:20:35.170
although there is a duplicate which is maybe higher priced and

324
00:20:36.210 --> 00:20:39.970
is the one product that you as a retailer would want the consumer to

325
00:20:40.520 --> 00:20:44.120
to purchase. And the issue here is that

326
00:20:44.600 --> 00:20:48.440
revenue cannibalization is very often not counterintuitive.

327
00:20:48.440 --> 00:20:52.240
It's very intuitive. However, it's never in the data. I haven't seen a

328
00:20:52.240 --> 00:20:55.880
single retailer that has good relationship model across

329
00:20:55.959 --> 00:20:59.640
tens of thousands, hundreds of thousands of products to know which products

330
00:20:59.640 --> 00:21:03.080
are very similar because you would need a very good

331
00:21:03.080 --> 00:21:06.800
relationship mapping of all products, similarity scores,

332
00:21:06.800 --> 00:21:10.330
quite good attribute data to figure out which products are

333
00:21:10.330 --> 00:21:13.690
closely related to each other and what is the impact of one to the other.

334
00:21:13.690 --> 00:21:16.810
However, margin cannibalization is very often

335
00:21:16.810 --> 00:21:20.610
counterintuitive. You take for example two

336
00:21:21.330 --> 00:21:24.290
examples that we had are drinks where

337
00:21:24.850 --> 00:21:28.290
branded products that are higher priced

338
00:21:28.690 --> 00:21:32.330
had significantly worse margins than the product than the

339
00:21:32.330 --> 00:21:35.970
private label brands that are cheaper and have higher margins.

340
00:21:37.010 --> 00:21:40.770
So in these situations what happens very

341
00:21:40.770 --> 00:21:44.370
often is that companies

342
00:21:44.450 --> 00:21:48.130
do not understand sufficiently the

343
00:21:48.130 --> 00:21:51.850
impact that let's say price differences and then the relative

344
00:21:51.850 --> 00:21:55.330
margin of these products has on in the end the

345
00:21:55.330 --> 00:21:59.090
purchase decision of the consumer and then incentivize

346
00:21:59.170 --> 00:22:02.580
the consumer to purchase the wrong product

347
00:22:03.140 --> 00:22:06.580
by for example, reducing prices by watching

348
00:22:06.740 --> 00:22:10.580
the market too closely or trying to increase

349
00:22:10.580 --> 00:22:14.340
velocity of the product. We'll be back after a

350
00:22:14.340 --> 00:22:17.659
short ad break. And after the break we step into the

351
00:22:17.659 --> 00:22:21.460
strategic misread that holds retailers back and

352
00:22:21.460 --> 00:22:25.220
where the next advantage isn't data but decisiveness.

353
00:22:31.710 --> 00:22:35.310
Hey guys, welcome back to our interview with Arbor from

354
00:22:35.310 --> 00:22:39.070
Zenline. Helping retailers to increase revenue

355
00:22:39.070 --> 00:22:42.590
and margins. Working with AI here and I was

356
00:22:42.590 --> 00:22:46.270
wondering, your weekly decision cycles replace

357
00:22:46.430 --> 00:22:49.870
seasonal ones. What operational structure

358
00:22:49.950 --> 00:22:53.620
fails first when a retailer attempts that

359
00:22:53.620 --> 00:22:57.340
transition? And also a little bit tongue in

360
00:22:57.340 --> 00:23:00.940
cheek question, when are we at daily

361
00:23:00.940 --> 00:23:04.740
decision cycles? Very

362
00:23:04.740 --> 00:23:08.540
good question. I have to say we

363
00:23:08.540 --> 00:23:12.340
have spoken to many retailers and we're working with many retailers where the decision

364
00:23:12.420 --> 00:23:16.100
cycles can sometimes take more than 12 months. We were

365
00:23:16.100 --> 00:23:19.620
seeing very good examples and we're seeing bad ones.

366
00:23:20.630 --> 00:23:24.470
The operational structure that fails very often is the

367
00:23:24.470 --> 00:23:27.590
relationship, especially in the private label business to the

368
00:23:27.590 --> 00:23:31.190
suppliers where it can take up to 12 to 18

369
00:23:31.350 --> 00:23:34.870
months to actually supply the product. So from

370
00:23:34.870 --> 00:23:38.390
let's say initial ideation to concept development

371
00:23:38.630 --> 00:23:42.390
to iteration to receiving the stock that's required

372
00:23:42.390 --> 00:23:46.150
to sell the product, sometimes it takes between 12 to 18 months, which

373
00:23:46.150 --> 00:23:49.860
in the end makes AI driven quick decisions

374
00:23:49.860 --> 00:23:53.580
completely obsolete because you don't need it if your supply chain process in the end

375
00:23:53.580 --> 00:23:57.300
can't deliver what you're trying to do organizationally in

376
00:23:57.300 --> 00:24:00.780
terms of commercial output, delivering the assortment that consumers want.

377
00:24:01.420 --> 00:24:05.260
However, with

378
00:24:05.420 --> 00:24:09.020
many of our retail customers, we're also seeing very good examples where the

379
00:24:09.020 --> 00:24:12.780
operational architecture actually is already in place. But the

380
00:24:12.780 --> 00:24:16.540
bottleneck was never the supply chain delivery but, but it was the

381
00:24:16.540 --> 00:24:20.100
analysis and the inside gathering because it's bottlenecked at the category

382
00:24:20.100 --> 00:24:23.700
managers. And if the category managers simply can output

383
00:24:23.700 --> 00:24:27.380
more decisions, the let's say supply chain

384
00:24:27.380 --> 00:24:31.140
behind it is already supportive enough, is already, let's say

385
00:24:31.140 --> 00:24:34.900
established so that they can quickly start to

386
00:24:34.900 --> 00:24:38.700
see, see these products being in the

387
00:24:38.700 --> 00:24:42.340
shelves within the weeks. On the question when are we on daily decisions? I

388
00:24:42.340 --> 00:24:45.990
think daily decisions will not be the bottleneck,

389
00:24:46.070 --> 00:24:49.790
but daily supply chains will be very hard to achieve for

390
00:24:49.790 --> 00:24:53.470
retailers in most industries. If we wouldn't be on a time constraint,

391
00:24:53.470 --> 00:24:54.550
I would go into

392
00:24:57.430 --> 00:25:00.950
deviate a little bit and go into some thoughts,

393
00:25:00.950 --> 00:25:04.390
experiments, what we could do with Jones here. But let's take

394
00:25:04.550 --> 00:25:07.590
our main line of interview questions here.

395
00:25:08.460 --> 00:25:11.740
Rejected recommendations reveal a lot about the

396
00:25:11.740 --> 00:25:15.540
organizational psychology. What patterns do you see in

397
00:25:15.540 --> 00:25:19.300
teams that aren't ready for agentic systems? I would

398
00:25:19.300 --> 00:25:22.940
say organizations that focus too much on the errors that happen,

399
00:25:23.580 --> 00:25:26.980
or let's say the parts of the

400
00:25:26.980 --> 00:25:30.820
recommendations that are misleading, that are error prone, that

401
00:25:30.820 --> 00:25:34.460
are just simply not useful at the beginning. And

402
00:25:35.230 --> 00:25:38.750
organizations that focus too much on what's not working

403
00:25:39.150 --> 00:25:42.710
and where the data is not sufficiently good are

404
00:25:42.710 --> 00:25:46.350
the ones that will become the laggards

405
00:25:46.350 --> 00:25:49.910
of the, of the decade. So however, on the other

406
00:25:49.910 --> 00:25:53.670
side, the teams that are utilizing agentic

407
00:25:53.670 --> 00:25:57.390
systems most effectively basically treat them as junior

408
00:25:57.390 --> 00:26:01.140
analysts, I would say. So you have a junior analyst. What happens if

409
00:26:01.290 --> 00:26:04.970
you receive a lot of input that will be good, some will be exceptionally good

410
00:26:05.210 --> 00:26:08.770
and some will be bad simply because it misses context, it doesn't know

411
00:26:08.770 --> 00:26:11.690
exactly what has happened in the last 12 months of the organization.

412
00:26:12.250 --> 00:26:15.930
Sometimes recommendations are bad because they have been done,

413
00:26:15.930 --> 00:26:19.370
they have been tried in the past, maybe 12 months ago, 24 months ago.

414
00:26:19.610 --> 00:26:23.330
And certain products, certain product types or categories just don't

415
00:26:23.330 --> 00:26:27.180
work with the retailer. And a

416
00:26:27.180 --> 00:26:31.020
junior analyst also wouldn't know them. So what very

417
00:26:31.020 --> 00:26:34.180
good, what patterns of very good teams are

418
00:26:34.580 --> 00:26:38.380
is the willingness to give feedback, give

419
00:26:38.380 --> 00:26:42.220
nuanced, and detailed feedback to advance the memory of the system and

420
00:26:42.220 --> 00:26:45.860
give more context into the system, define the strategy, the company

421
00:26:45.940 --> 00:26:49.660
strategy, the category strategy, and thus be able then

422
00:26:49.660 --> 00:26:53.500
to improve it over time again, as you would do with the junior analyst. And

423
00:26:53.500 --> 00:26:57.150
I think the more organizations start to think of

424
00:26:57.150 --> 00:27:00.830
agentic systems of colleagues, of basically just another part of

425
00:27:00.830 --> 00:27:04.390
the workforce, the faster they will be the ones that actually

426
00:27:04.390 --> 00:27:06.710
gain margin benefits from it.

427
00:27:09.910 --> 00:27:13.750
Your pricing agents benchmarks

428
00:27:13.750 --> 00:27:16.630
internal logic against the market reality.

429
00:27:17.350 --> 00:27:21.030
What's the structural pricing error that retailers report?

430
00:27:23.050 --> 00:27:26.490
Okay, so I haven't seen a single retailer yet who

431
00:27:27.050 --> 00:27:30.250
has been able to be very strong

432
00:27:30.810 --> 00:27:34.090
in utilizing internal relationships

433
00:27:34.490 --> 00:27:38.250
in their pricing. So what happens very often is that many retailers

434
00:27:38.250 --> 00:27:41.770
are very much focused on hey, how is this one single product

435
00:27:43.130 --> 00:27:46.770
priced compared to the market to other other

436
00:27:46.770 --> 00:27:50.270
retailers selling this one product? What every

437
00:27:50.270 --> 00:27:53.950
retailer does is they compare their own assortment

438
00:27:53.950 --> 00:27:57.710
very much with the assortment of other retailers.

439
00:27:57.710 --> 00:28:01.310
And what they neglect is that a lot of the pricing,

440
00:28:01.310 --> 00:28:04.350
a lot of the pricing power that they have is not

441
00:28:04.510 --> 00:28:08.270
comparing themselves to competitors,

442
00:28:08.350 --> 00:28:12.190
but to the internal assortment. Because very often when a consumer is

443
00:28:12.190 --> 00:28:15.790
already in the stores, they will make a decision

444
00:28:15.790 --> 00:28:18.710
not based on what is the price of the competitor, but they will make a

445
00:28:18.710 --> 00:28:22.120
price based on what are the pro, the other products, the

446
00:28:22.120 --> 00:28:25.440
alternatives, the substitutes that are already in the shelf.

447
00:28:26.560 --> 00:28:30.400
And especially with regards to decisions that impede the margins,

448
00:28:30.480 --> 00:28:34.240
customers just retailers typically don't have a, let's

449
00:28:34.240 --> 00:28:37.880
say, sophisticated enough system to

450
00:28:37.880 --> 00:28:40.920
understand the relationships. As you mentioned before, the

451
00:28:40.920 --> 00:28:44.680
cannibalization impact of not only delisting a product, but also

452
00:28:44.680 --> 00:28:48.460
changing the price of one product or what does it do to all the other

453
00:28:48.460 --> 00:28:52.060
products in the assortment. And this relationship model is something that

454
00:28:52.460 --> 00:28:56.220
retailers should think about. However, they're doing it

455
00:28:56.300 --> 00:28:58.780
way too little because it's very complex technically.

456
00:29:01.100 --> 00:29:02.939
So the classical substitute products

457
00:29:04.460 --> 00:29:08.100
substitute variant. Pricing as soon as you get to

458
00:29:08.100 --> 00:29:11.780
10,000, let's say 10,000 products. It's

459
00:29:11.780 --> 00:29:15.450
very hard to manage a very clean relationship mapping of your

460
00:29:15.450 --> 00:29:19.010
products so that for example, when you change one product, you also change

461
00:29:19.570 --> 00:29:23.090
either in, let's say proportionally

462
00:29:23.410 --> 00:29:27.250
or at the exact same level. The price of variance can also

463
00:29:27.250 --> 00:29:31.010
be the same for substitutes to just understand what does the change

464
00:29:31.010 --> 00:29:34.770
of one product's price due to the rest of the, let's say,

465
00:29:34.770 --> 00:29:36.450
very similar, very related products.

466
00:29:39.340 --> 00:29:42.940
We've been talking about margins a lot and leaders often

467
00:29:43.020 --> 00:29:46.620
freeze under margin pressure. What differentiates

468
00:29:46.780 --> 00:29:50.300
those who accelerate versus those who stall?

469
00:29:52.380 --> 00:29:55.980
I think a lot is in the tools. Some

470
00:29:56.140 --> 00:29:59.980
retailers still consider power bi as state of the art

471
00:29:59.980 --> 00:30:03.700
and A lot of them still stick with Excels, which is

472
00:30:03.700 --> 00:30:07.070
very convenient because they don't have to learn new tools. But the

473
00:30:07.070 --> 00:30:10.670
spreadsheet logic is failing. It's failing as soon as you have

474
00:30:10.670 --> 00:30:14.110
hundreds of products, not even mentioning

475
00:30:14.510 --> 00:30:18.270
thousands or millions of products. And I think retailers

476
00:30:18.270 --> 00:30:21.870
just need to accept the fact that let's say human

477
00:30:21.950 --> 00:30:23.710
driven and spreadsheets driven

478
00:30:25.710 --> 00:30:29.470
working modes are not optimized

479
00:30:29.470 --> 00:30:32.910
to support retailers where decisions are

480
00:30:32.910 --> 00:30:36.590
incredibly complex. And yeah, in the end,

481
00:30:36.590 --> 00:30:40.150
competition will not wait. Competition from other markets,

482
00:30:40.870 --> 00:30:43.910
new more digital versions of retailers popping up. And

483
00:30:44.470 --> 00:30:48.230
the margin winners will be the ones that rewire how

484
00:30:48.310 --> 00:30:51.750
fast product decisions can be made because they have the data

485
00:30:51.750 --> 00:30:55.430
ready, readily available, and then they have the systems that

486
00:30:55.430 --> 00:30:59.150
output recommendations and then review the

487
00:30:59.150 --> 00:31:02.870
recommendations. So basically the shift from going I need to do

488
00:31:02.870 --> 00:31:06.670
the spreadsheet analysis myself to I

489
00:31:06.670 --> 00:31:10.470
only review the output of systems and I can then

490
00:31:10.470 --> 00:31:14.310
fine tune for a few hours instead of being the one who has

491
00:31:14.310 --> 00:31:15.630
done the analysis for weeks.

492
00:31:18.750 --> 00:31:20.590
I know how this feels.

493
00:31:23.630 --> 00:31:27.470
You gotta be creative here. You guys

494
00:31:27.470 --> 00:31:30.960
scaled at a pace even few teams

495
00:31:30.960 --> 00:31:34.000
attempt which moment

496
00:31:34.240 --> 00:31:37.920
demanded your highest level of personal, personal resilience.

497
00:31:39.280 --> 00:31:42.880
Honestly, there, there are many. As a founder, you constantly

498
00:31:42.880 --> 00:31:46.400
are on the edge, but when you have a strong team,

499
00:31:46.400 --> 00:31:50.240
you, you can go through very, very, very long hours

500
00:31:50.240 --> 00:31:54.040
together. And I think that's the happiest and the

501
00:31:54.040 --> 00:31:56.480
best parts. But for me it's a combination of several

502
00:31:57.530 --> 00:32:01.050
countless sleepless nights and pressure to keep customers happy.

503
00:32:03.130 --> 00:32:06.330
Yeah, but I would say whenever

504
00:32:07.210 --> 00:32:11.010
the team is willing to go the extra mile, is willing to

505
00:32:11.010 --> 00:32:13.370
go beyond, to basically be able to deliver

506
00:32:15.530 --> 00:32:19.330
the results that we are promising with the agents, that's the

507
00:32:19.330 --> 00:32:23.170
moments where let's say, well, that's our

508
00:32:23.170 --> 00:32:23.930
source of energy.

509
00:32:27.370 --> 00:32:31.130
Your source of energy. Okay, I get the

510
00:32:31.130 --> 00:32:34.730
feeling that Asian driven operations are

511
00:32:34.810 --> 00:32:38.650
here compressing time, which is just talked about quarterly to

512
00:32:38.650 --> 00:32:42.010
weekly in the interview. Which

513
00:32:44.890 --> 00:32:48.490
legacy workflow becomes obsolete here?

514
00:32:49.210 --> 00:32:49.610
First.

515
00:32:53.270 --> 00:32:54.150
White collar

516
00:32:56.550 --> 00:32:59.350
tasks that require desk research.

517
00:33:00.230 --> 00:33:04.070
All the research work will be automated in the

518
00:33:04.070 --> 00:33:07.870
future. And you can see this already in consulting,

519
00:33:07.870 --> 00:33:11.670
which they have huge pressures. It's the same for

520
00:33:11.670 --> 00:33:15.350
lawyers, it's the same for any kind of research assistants. Because

521
00:33:16.670 --> 00:33:19.870
I don't know if you have used deep research from

522
00:33:20.190 --> 00:33:23.390
Gemini and ChatGPT, but it's incredible how

523
00:33:23.710 --> 00:33:27.470
much depth and how much data can be processed with these

524
00:33:27.470 --> 00:33:30.990
tools. And they can do it significantly better than humans can.

525
00:33:31.230 --> 00:33:35.070
So everything related to research and analysis will be something

526
00:33:35.230 --> 00:33:38.910
that will be obsolete in the future. And humans will

527
00:33:38.910 --> 00:33:42.630
switch from doing the analysis to making the decisions. Basically exactly what we

528
00:33:42.630 --> 00:33:46.210
discussed before with the leaders of margin drivers

529
00:33:47.810 --> 00:33:50.610
and compressing months of analysis

530
00:33:52.050 --> 00:33:55.410
into days, if not sometimes even hours. Making

531
00:33:55.570 --> 00:33:59.330
fast decisions, making counterintuitive decisions

532
00:34:00.130 --> 00:34:03.970
makes explainability become something strategic.

533
00:34:04.370 --> 00:34:07.570
How do you maintain trust when high impact

534
00:34:07.730 --> 00:34:11.500
recommendation contradicts their intuition?

535
00:34:13.500 --> 00:34:17.180
I like that question, especially in

536
00:34:17.180 --> 00:34:20.820
an In a machine learning driven world that we had

537
00:34:20.820 --> 00:34:24.380
for the past, let's say between 2016

538
00:34:24.460 --> 00:34:28.220
to 2022, explainability was

539
00:34:28.220 --> 00:34:31.820
really hard because you had a result.

540
00:34:32.860 --> 00:34:36.450
Your machine learning model told you this should be the new price,

541
00:34:37.250 --> 00:34:40.610
but there was very little explainability behind it.

542
00:34:41.250 --> 00:34:44.730
What is now interesting with generative AI is that the

543
00:34:44.730 --> 00:34:48.130
models have reasoning tokens and they can

544
00:34:48.290 --> 00:34:51.890
output and they can explain why they came to a

545
00:34:51.890 --> 00:34:55.330
decision. And this reasoning is getting closer and

546
00:34:55.330 --> 00:34:59.050
closer to the reality of why the model decided to do what it

547
00:34:59.050 --> 00:35:02.850
decided to do. And this is actually one of the things that we have realized

548
00:35:02.850 --> 00:35:06.640
with customers is they love to see the thinking

549
00:35:06.640 --> 00:35:10.160
of the model. So not only the final output of why you did it, but

550
00:35:10.160 --> 00:35:13.840
even the steps, the thinking steps that happened, the tool calls that

551
00:35:13.840 --> 00:35:17.680
happened, the web search, which websites did it browse, all of

552
00:35:17.680 --> 00:35:20.520
that to make it very very very transparent.

553
00:35:21.800 --> 00:35:25.600
So I think in the last 10 years there was a bit

554
00:35:25.600 --> 00:35:29.400
of the dogma to

555
00:35:29.640 --> 00:35:32.920
hide, a bit of the complexity behind it, to not confuse

556
00:35:33.160 --> 00:35:36.640
especially the managers, but also the category

557
00:35:36.640 --> 00:35:40.280
managers and also the executive teams. But now with LLMs

558
00:35:40.280 --> 00:35:43.600
you can be very very very transparent about why the

559
00:35:43.680 --> 00:35:47.360
model did what it did. And very often it's you can find

560
00:35:47.680 --> 00:35:51.200
immediately in the prompt why it did the

561
00:35:51.200 --> 00:35:54.840
decisions the way it did and then you can together with the customers

562
00:35:54.840 --> 00:35:58.440
define the prompt to make it basically just be more

563
00:35:58.440 --> 00:36:01.280
effective for their categories and their company.

564
00:36:02.410 --> 00:36:05.970
At the time of recording and admittedly we were recording this in early

565
00:36:05.970 --> 00:36:09.770
December, Europe has a 180 day window before global

566
00:36:10.490 --> 00:36:14.170
platforms widened the gap. What is the one move

567
00:36:14.490 --> 00:36:17.050
retailers must make now?

568
00:36:19.850 --> 00:36:21.690
I might be a bit biased but

569
00:36:24.490 --> 00:36:28.140
it's the super obvious one. Skip

570
00:36:28.140 --> 00:36:30.500
all this digitalization

571
00:36:32.180 --> 00:36:35.740
work. Just everything related to pure IT

572
00:36:35.740 --> 00:36:39.340
transformation needs to be done in

573
00:36:39.340 --> 00:36:42.900
parallel to all the agentic more business related

574
00:36:43.060 --> 00:36:46.820
work that needs to happen immediately. No retailer has

575
00:36:46.820 --> 00:36:50.420
the time now to wait for months and months and months if not

576
00:36:50.420 --> 00:36:53.860
years to do IT transformation SAP for

577
00:36:53.860 --> 00:36:57.640
Hana Transformations. But at the same time they should be already

578
00:36:57.640 --> 00:37:01.360
working in the business. Not specifically with it,

579
00:37:01.360 --> 00:37:05.160
but the businesses need to start working on agent based systems and see

580
00:37:05.160 --> 00:37:08.200
what's already possible with them. Play around with it, find

581
00:37:08.760 --> 00:37:12.280
very very highly valuable use cases that

582
00:37:12.840 --> 00:37:16.120
will make an impact on the margin within four to eight weeks.

583
00:37:17.400 --> 00:37:21.160
And I would say if retailers, especially in Europe and

584
00:37:21.160 --> 00:37:24.990
the US just feel more confident to do these things, they will

585
00:37:24.990 --> 00:37:28.670
see how advanced retailers are. And that's one of the things that we hear

586
00:37:28.670 --> 00:37:32.310
very often. Typically the response is after

587
00:37:32.790 --> 00:37:36.510
a month. They say, I would have expected AI to be three

588
00:37:36.510 --> 00:37:40.350
to four years away from what you showed us. And I think this

589
00:37:40.350 --> 00:37:44.030
is where you need to get the retailers to want to work more with

590
00:37:44.030 --> 00:37:47.830
it. I've heard over and over that focus

591
00:37:48.470 --> 00:37:52.170
determines survivability, especially for you guys. What,

592
00:37:52.400 --> 00:37:55.760
which idea did you kill that unlocked your current

593
00:37:55.760 --> 00:37:59.520
trajectory? In the beginning of Zinlin, we were focusing

594
00:37:59.840 --> 00:38:03.120
on the. On a very similar problem, on the assortment problem,

595
00:38:03.440 --> 00:38:06.840
but for manufacturers which have very

596
00:38:06.840 --> 00:38:10.640
complex portfolios, sometimes hundreds of thousands

597
00:38:10.640 --> 00:38:14.360
of SKUs. And as soon as

598
00:38:14.360 --> 00:38:18.160
we started talking to retailers who are the buyers of those, let's say, complex

599
00:38:18.160 --> 00:38:21.880
assortment, complex portfolios, we realized that the interest and the

600
00:38:21.880 --> 00:38:25.640
traction that we received was. It was instantly

601
00:38:25.640 --> 00:38:29.360
obvious. And we basically pivoted our entire company within three

602
00:38:29.360 --> 00:38:33.120
hours to this new idea because of the traction

603
00:38:33.120 --> 00:38:35.440
that we received there. And I think

604
00:38:36.880 --> 00:38:40.000
being able to take very, very strong decisions and

605
00:38:41.040 --> 00:38:44.760
let's say, reshifting the organization's focus, not only individual

606
00:38:44.760 --> 00:38:48.330
focus, but the focus of the entire organization on a new topic, new type of

607
00:38:48.330 --> 00:38:51.930
customers, and then very narrow scope of,

608
00:38:51.930 --> 00:38:55.570
let's say, features. This is what makes the startup incredibly

609
00:38:55.570 --> 00:38:59.130
powerful. And to be very honest, it's not about killing

610
00:38:59.130 --> 00:39:02.970
ideas very often, it's just about delaying them. You know that customers

611
00:39:02.970 --> 00:39:06.650
ask for something, you know that customers are willing to pay

612
00:39:06.650 --> 00:39:10.410
for something, but you just don't have the capacity to serve them

613
00:39:10.410 --> 00:39:14.110
in the quality that you want to show to

614
00:39:14.110 --> 00:39:17.670
them. So basically you just, you have a roadmap and

615
00:39:17.670 --> 00:39:21.190
you put it off until the features that you're building

616
00:39:21.190 --> 00:39:24.870
are brilliant and you can scale them first before adding the new

617
00:39:24.870 --> 00:39:28.710
ones. I just want to ask you this

618
00:39:28.710 --> 00:39:31.990
again because I'm not sure I heard it correctly. You

619
00:39:31.990 --> 00:39:35.270
pivoted once with a decision within three hours.

620
00:39:35.990 --> 00:39:39.670
Yes. I think that's the future of business.

621
00:39:41.580 --> 00:39:45.180
That actually ties in into my last official question. Your product

622
00:39:45.180 --> 00:39:48.940
velocity is unusually high. What principle

623
00:39:49.260 --> 00:39:53.020
anchors precision without slowing execution?

624
00:39:53.500 --> 00:39:57.020
The two things that we do brilliantly, and it's also related

625
00:39:57.020 --> 00:40:00.660
to the one before that is very,

626
00:40:00.660 --> 00:40:03.900
very, very high focus on the most important features,

627
00:40:04.540 --> 00:40:08.230
only the ones that are really meaningful to the customer. And then

628
00:40:09.030 --> 00:40:12.390
we don't obsess with

629
00:40:12.470 --> 00:40:16.190
technical details. We don't obsess with the beauty

630
00:40:16.190 --> 00:40:18.870
of the code. We obsess with the customer value

631
00:40:19.910 --> 00:40:23.589
and we Every single day we question ourselves, is this

632
00:40:23.589 --> 00:40:27.030
the most impactful thing we can do that really saves

633
00:40:27.350 --> 00:40:30.870
margin for the user? And we have it on our boards,

634
00:40:30.950 --> 00:40:34.670
we have it in our notions. The one key metric that we track

635
00:40:34.670 --> 00:40:37.640
is did we improve the margin of our customers?

636
00:40:38.280 --> 00:40:40.920
And in the end, if we don't do that,

637
00:40:42.440 --> 00:40:46.160
we're building a bad product. So this is what has helped us

638
00:40:46.160 --> 00:40:49.320
to keep the product velocity very, very high.

639
00:40:51.240 --> 00:40:55.000
I see. That was the official

640
00:40:55.080 --> 00:40:57.400
questions and we usually end with

641
00:40:58.920 --> 00:41:02.610
our two standard questions. Are you open to talk to new investors?

642
00:41:03.010 --> 00:41:06.130
We are currently well funded, so we are not talking to investors,

643
00:41:06.690 --> 00:41:09.970
but we are aiming to

644
00:41:10.770 --> 00:41:13.650
speak to investors again beginning summer 2026.

645
00:41:14.690 --> 00:41:18.530
Are you currently looking for talented employees?

646
00:41:19.650 --> 00:41:23.090
Yes, we are. We are looking for the best and most

647
00:41:23.090 --> 00:41:26.450
ambitious and boldest engineers. We have two,

648
00:41:27.250 --> 00:41:30.610
we're thinking in two teams. At Zenline, we have product

649
00:41:30.610 --> 00:41:34.410
engineers which are close to the customer, but in the end

650
00:41:34.410 --> 00:41:37.730
they're responsible for building features and building the product. And then we have

651
00:41:37.970 --> 00:41:41.810
forward deployed engineers, which is a term coined by Palantir, which

652
00:41:41.810 --> 00:41:45.370
is engineers who are managing the

653
00:41:45.370 --> 00:41:49.170
customer relationship. So in the end,

654
00:41:49.410 --> 00:41:52.890
Zenline is built purely by engineers. However,

655
00:41:53.610 --> 00:41:56.970
we then distinguish between the ones that own the product and the ones that own

656
00:41:56.970 --> 00:41:58.090
the customer relationship.

657
00:42:00.250 --> 00:42:03.690
Albert, you've shown that the future of retail belongs to

658
00:42:03.690 --> 00:42:07.130
operators who move faster, see deeper

659
00:42:07.450 --> 00:42:10.970
and trust systems built for action rather than

660
00:42:10.970 --> 00:42:14.250
observation. You clarified where margin

661
00:42:14.410 --> 00:42:18.130
heights, where complexity wastes energy, and

662
00:42:18.130 --> 00:42:21.880
why agentic decision redefines competitiveness.

663
00:42:22.120 --> 00:42:25.080
For those who want to follow your work or understand

664
00:42:25.160 --> 00:42:28.760
Zenline's impact more directly, Visit try Zenline

665
00:42:28.760 --> 00:42:32.600
AI or connect with you on LinkedIn. Thank you for bringing

666
00:42:32.600 --> 00:42:35.480
structural clarity to sector that really needs it.

667
00:42:37.080 --> 00:42:38.040
Thank you so much.

668
00:42:43.560 --> 00:42:47.030
That's all for Find more news, streams,

669
00:42:47.270 --> 00:42:48.310
events and

670
00:42:48.310 --> 00:42:52.630
interviews@www.startuprad.IO.

671
00:42:53.110 --> 00:42:55.030
remember, sharing is car.