1
00:00:05,960 --> 00:00:07,920
Speaker 1: Welcome to Fear and Greed Q and A where we

2
00:00:07,960 --> 00:00:11,760
Speaker 1: ask and answer questions about business, investing, economics, politics, and

3
00:00:11,800 --> 00:00:14,440
Speaker 1: plenty more. I'm suan Alma. We all know the value

4
00:00:14,480 --> 00:00:18,319
Speaker 1: of AI in automating tasks, beating up processes, and even

5
00:00:18,360 --> 00:00:21,520
Speaker 1: improving customer services through chat pots and the like. But

6
00:00:21,600 --> 00:00:23,799
Speaker 1: where to from here? AI can do a lot more

7
00:00:23,880 --> 00:00:26,720
Speaker 1: than just automation, and for small businesses it could be

8
00:00:26,840 --> 00:00:28,960
Speaker 1: a game changer. In fact, it will be a game

9
00:00:29,080 --> 00:00:31,280
Speaker 1: changer as long as you know how to use it.

10
00:00:31,400 --> 00:00:34,800
Speaker 1: James Bergen is executive general manager of Technology, Research and

11
00:00:34,880 --> 00:00:38,960
Speaker 1: Advocacy at Zero, which is a terrific supporter of this podcast. James,

12
00:00:39,000 --> 00:00:40,240
Speaker 1: welcome back to Fear and Greed.

13
00:00:40,560 --> 00:00:41,920
Speaker 2: Sure, and it's a pleasure to be back. Thanks so

14
00:00:42,040 --> 00:00:42,640
Speaker 2: much for having me.

15
00:00:43,040 --> 00:00:46,840
Speaker 1: AI is evolving from sort of repetitive tasks to more

16
00:00:46,880 --> 00:00:51,960
Speaker 1: strategic functions. It's a big step. Where are we up

17
00:00:52,000 --> 00:00:56,040
Speaker 1: to in that process? And I'm sure we're underutilizing it

18
00:00:56,040 --> 00:00:58,840
Speaker 1: at the moment, but how do we get there? Yeah?

19
00:00:58,840 --> 00:01:01,160
Speaker 2: I mean, I think it's a definitely an evolution that

20
00:01:01,200 --> 00:01:03,720
Speaker 2: we're seeing happen with the technology and the capabilities that

21
00:01:03,760 --> 00:01:06,880
Speaker 2: the technology is unlocking. As all of us are figuring

22
00:01:06,880 --> 00:01:09,920
Speaker 2: out the best ways to incorporate these new tools and

23
00:01:10,560 --> 00:01:13,320
Speaker 2: techniques into how we do our work. And I think

24
00:01:13,319 --> 00:01:15,160
Speaker 2: that we've been talking about Jena of AI for a

25
00:01:15,160 --> 00:01:17,319
Speaker 2: while and now obviously we're talking about a gent KI

26
00:01:17,520 --> 00:01:20,080
Speaker 2: and we're talking about this increased capability, and really, what

27
00:01:20,120 --> 00:01:23,440
Speaker 2: I think it's starting to do is surface more opportunities

28
00:01:23,520 --> 00:01:26,360
Speaker 2: for insight to be brought to us, which then helps

29
00:01:26,400 --> 00:01:30,000
Speaker 2: identify opportunities themselves. So it becomes like this nice self

30
00:01:30,040 --> 00:01:32,399
Speaker 2: fulfilling loop, if you like. And then it helps with

31
00:01:32,400 --> 00:01:35,880
Speaker 2: strategic decisioning. And when I think strategic decisioning, I think

32
00:01:35,920 --> 00:01:38,640
Speaker 2: strategy is about a course of action, and so anything

33
00:01:38,680 --> 00:01:41,160
Speaker 2: that can help advise on a course of action is

34
00:01:41,560 --> 00:01:42,120
Speaker 2: a good thing.

35
00:01:42,400 --> 00:01:44,479
Speaker 1: How do you get used to using AI? So I mean,

36
00:01:44,640 --> 00:01:47,440
Speaker 1: I'll get to the strategy part. We had Scott far

37
00:01:47,520 --> 00:01:50,160
Speaker 1: quite from a lassion on and off air. He said, Hey,

38
00:01:50,200 --> 00:01:53,360
Speaker 1: if you want to learn how to use AI, just

39
00:01:53,560 --> 00:01:56,400
Speaker 1: use it in everyday life. So that night when Haym's

40
00:01:56,400 --> 00:01:58,640
Speaker 1: blake for Jackie, we actually did meal planning for the

41
00:01:58,680 --> 00:02:01,480
Speaker 1: next three nights we've been on holidays. I was saying, hey,

42
00:02:01,480 --> 00:02:03,720
Speaker 1: what's the best place things to do so rather than

43
00:02:03,760 --> 00:02:07,040
Speaker 1: google any of that stuff, we our chat GPT in

44
00:02:07,120 --> 00:02:10,160
Speaker 1: our daily lives, which I think is a fantastic way

45
00:02:10,720 --> 00:02:12,480
Speaker 1: then to morph it into. But do you have it?

46
00:02:12,520 --> 00:02:14,800
Speaker 1: I mean, presumably you're not until you agree with that,

47
00:02:14,919 --> 00:02:17,320
Speaker 1: But you know, is that how we learn AI?

48
00:02:17,800 --> 00:02:20,880
Speaker 2: I think everyone learns within a new tool or new technology.

49
00:02:20,919 --> 00:02:23,880
Speaker 2: Everyone probably learns in a slightly different way. But I

50
00:02:24,040 --> 00:02:27,200
Speaker 2: have been I think my high school chemistry teacher would

51
00:02:27,200 --> 00:02:29,440
Speaker 2: find it hilarious how often I use this as an example,

52
00:02:29,480 --> 00:02:31,840
Speaker 2: but I kind of hark back to what we learned

53
00:02:31,840 --> 00:02:34,720
Speaker 2: when we were doing science in school at whatever level

54
00:02:34,720 --> 00:02:36,400
Speaker 2: we took that to. You learn how to do an

55
00:02:36,400 --> 00:02:39,520
Speaker 2: experiment right, and you frame up a hypothesis and then

56
00:02:39,760 --> 00:02:42,840
Speaker 2: you go and conduct an experiment. You test that hypothesis,

57
00:02:42,880 --> 00:02:45,359
Speaker 2: and you try and gather data along the way to

58
00:02:45,400 --> 00:02:48,640
Speaker 2: see if you can confirm it or prove that it's false.

59
00:02:49,160 --> 00:02:51,480
Speaker 2: And that's the mentality that I take, and I try

60
00:02:51,520 --> 00:02:54,080
Speaker 2: and encourage, you know, with my colleagues and with my

61
00:02:54,080 --> 00:02:57,440
Speaker 2: family as well as to experiment safely, but start by

62
00:02:57,440 --> 00:02:59,400
Speaker 2: forming a bit of a hypothesis and then testing the

63
00:02:59,400 --> 00:03:02,000
Speaker 2: capability of these tools in a safe way and going, oh,

64
00:03:02,120 --> 00:03:05,080
Speaker 2: actually I've got some new data points. Maybe my hypothesis

65
00:03:05,120 --> 00:03:07,960
Speaker 2: was right, maybe it was wrong. Okay, let's evolve from there,

66
00:03:07,960 --> 00:03:11,000
Speaker 2: because it's continuous learning. Anytime anyone says they've got this

67
00:03:11,120 --> 00:03:13,120
Speaker 2: all figured out, I don't know if they've been paying

68
00:03:13,120 --> 00:03:15,960
Speaker 2: attention because it moves so quickly. So it has to

69
00:03:16,000 --> 00:03:18,320
Speaker 2: be constantly something that we're looking to search and to

70
00:03:18,400 --> 00:03:20,960
Speaker 2: learn and to grow in our knowledge about. Okay, so

71
00:03:21,040 --> 00:03:23,880
Speaker 2: let's bring that to SME small and medium sized businesses.

72
00:03:24,720 --> 00:03:28,920
Speaker 2: How can they use AI to find opportunities to flag

73
00:03:29,200 --> 00:03:31,000
Speaker 2: financial risks in what's going on?

74
00:03:31,120 --> 00:03:31,800
Speaker 1: How do they do that?

75
00:03:32,080 --> 00:03:34,519
Speaker 2: Yeah? I mean I think a lot of the capability

76
00:03:34,520 --> 00:03:37,280
Speaker 2: that's exciting in this space is the ability to look

77
00:03:37,360 --> 00:03:41,120
Speaker 2: through and identify patterns and look at anomalies and things

78
00:03:41,160 --> 00:03:44,040
Speaker 2: like that. I mean, in our case, we have JAX,

79
00:03:44,040 --> 00:03:48,360
Speaker 2: which is just ask zero, our superagent, and what we're

80
00:03:48,360 --> 00:03:50,080
Speaker 2: looking at is how that will soon be able to

81
00:03:50,120 --> 00:03:54,120
Speaker 2: improve cash flow or compare loan interest rates across banks.

82
00:03:54,440 --> 00:03:57,200
Speaker 2: So when a small business is looking at evaluating loan

83
00:03:57,240 --> 00:04:01,320
Speaker 2: options or navigating tax changes, or you know, looking against

84
00:04:01,360 --> 00:04:04,240
Speaker 2: the market itself, how can that be brought to you,

85
00:04:04,280 --> 00:04:08,160
Speaker 2: that external data be brought to support smarter and more

86
00:04:08,200 --> 00:04:10,840
Speaker 2: sort of tailored decisions, if you like. And I think

87
00:04:11,120 --> 00:04:14,840
Speaker 2: the ability for these tools to really help with analyzing

88
00:04:14,960 --> 00:04:19,000
Speaker 2: large amounts of data from different areas is quite quite

89
00:04:19,040 --> 00:04:20,760
Speaker 2: a strength that's exciting to explore.

90
00:04:21,839 --> 00:04:24,039
Speaker 1: So obviously we're talking about Zero here, but you know,

91
00:04:24,080 --> 00:04:27,080
Speaker 1: if you're using a vendor such as Zero, should you

92
00:04:27,120 --> 00:04:31,480
Speaker 1: now expect those that sort of AI help in your products.

93
00:04:32,680 --> 00:04:37,359
Speaker 2: I think people are expecting that organizations like Zero are

94
00:04:37,560 --> 00:04:41,800
Speaker 2: looking to embrace and to utilize technology in smart ways

95
00:04:41,920 --> 00:04:43,640
Speaker 2: in order to be able to help add value in

96
00:04:43,640 --> 00:04:46,560
Speaker 2: the same way that Zero has a pedigree and history

97
00:04:46,600 --> 00:04:49,680
Speaker 2: of leveraging the cloud and leveraging capability that can be

98
00:04:49,720 --> 00:04:53,240
Speaker 2: brought to help small businesses and their advisors, their accountants

99
00:04:53,279 --> 00:04:56,040
Speaker 2: and bookkeepers. I think that expectation has been there for

100
00:04:56,080 --> 00:04:57,960
Speaker 2: a while. And if I think about, you know, the

101
00:04:58,000 --> 00:05:02,039
Speaker 2: technology platforms and comeanies that I use, my expectation is

102
00:05:02,040 --> 00:05:05,240
Speaker 2: not so much are you using AIS? Are you using

103
00:05:05,320 --> 00:05:10,080
Speaker 2: the capabilities that are being brought on board by technology, Okay,

104
00:05:10,240 --> 00:05:12,599
Speaker 2: like AI maybe, but there are others as well, and

105
00:05:12,640 --> 00:05:14,240
Speaker 2: are you leveraging those in a way that can really

106
00:05:14,279 --> 00:05:17,320
Speaker 2: help me? And I think that's a fair expectation because

107
00:05:17,320 --> 00:05:20,960
Speaker 2: it's really about organizations providing value with the tools that

108
00:05:21,000 --> 00:05:21,479
Speaker 2: are out there.

109
00:05:21,880 --> 00:05:23,920
Speaker 1: So let's look at the other side. Why the risks? Then,

110
00:05:24,120 --> 00:05:25,720
Speaker 1: what should SMEs.

111
00:05:25,360 --> 00:05:27,960
Speaker 2: Be worried about. I don't know about worried, but I think,

112
00:05:28,000 --> 00:05:30,480
Speaker 2: as I was saying with the experimentation mentality, it's trying

113
00:05:30,520 --> 00:05:32,599
Speaker 2: to look at these things you safely, and look at

114
00:05:32,640 --> 00:05:34,800
Speaker 2: where the boundaries are, and look at how you can

115
00:05:34,880 --> 00:05:38,000
Speaker 2: work on some of the boundaries of trust and what

116
00:05:38,360 --> 00:05:41,320
Speaker 2: lets you trust something versus maybe be a little bit

117
00:05:41,360 --> 00:05:43,800
Speaker 2: wary of it. You know, in our case, we take

118
00:05:43,839 --> 00:05:46,160
Speaker 2: that really seriously, obviously, because we're always about trying to

119
00:05:46,200 --> 00:05:49,560
Speaker 2: focus on security and privacy and rering the right expertise

120
00:05:49,600 --> 00:05:52,080
Speaker 2: and a personalized approach. We talk about that as well.

121
00:05:52,640 --> 00:05:54,480
Speaker 2: So I think looking at these tools through the lens

122
00:05:54,520 --> 00:05:56,920
Speaker 2: of the capabilities that they provide and then seeing where

123
00:05:56,920 --> 00:06:00,000
Speaker 2: the limitations of those capabilities are. I mean, a classic

124
00:06:00,080 --> 00:06:03,560
Speaker 2: example with generative AI is oftentimes we're used to asking

125
00:06:03,920 --> 00:06:07,880
Speaker 2: our technology quite deterministic questions. You know, one plus one

126
00:06:08,120 --> 00:06:11,760
Speaker 2: always equals to and with these models, we're often dealing

127
00:06:11,760 --> 00:06:16,000
Speaker 2: with probabilities rather than these exact answers. So one plus

128
00:06:16,040 --> 00:06:19,640
Speaker 2: one you know, probably equals to is a very different question.

129
00:06:20,080 --> 00:06:25,360
Speaker 2: And so asking for advice or input into discussions where

130
00:06:25,720 --> 00:06:28,800
Speaker 2: the stakes can open up for opportunity and the way,

131
00:06:28,960 --> 00:06:30,839
Speaker 2: you know, give me four or five different ways about

132
00:06:30,839 --> 00:06:33,479
Speaker 2: thinking about this, it's probably a better way than in

133
00:06:33,520 --> 00:06:35,919
Speaker 2: a specific I've got a really targeted question that I

134
00:06:35,920 --> 00:06:39,080
Speaker 2: want to ask, unless you're using a tool which has

135
00:06:39,120 --> 00:06:42,839
Speaker 2: been constrained and designed specifically for that. So again, in

136
00:06:42,880 --> 00:06:45,560
Speaker 2: our case, accuracy is are non negotiable in accounting. So

137
00:06:45,600 --> 00:06:48,480
Speaker 2: that's a really core driver behind Jack's and we have

138
00:06:48,560 --> 00:06:50,880
Speaker 2: a thing called Jack's a Sure, which is really about

139
00:06:50,920 --> 00:06:54,200
Speaker 2: taking a proprietary control system and managing and validating the

140
00:06:54,279 --> 00:06:57,279
Speaker 2: data that's being processed by the AI. So that's enhancing

141
00:06:57,279 --> 00:07:02,400
Speaker 2: accuracy and reducing hallucinations. The term that's often become used

142
00:07:02,440 --> 00:07:04,560
Speaker 2: for some of these tools, particularly when it comes to

143
00:07:05,160 --> 00:07:08,400
Speaker 2: comparing that to maybe solely large language models which don't

144
00:07:08,400 --> 00:07:10,800
Speaker 2: have those same kind of constraints placed on them.

145
00:07:11,200 --> 00:07:13,880
Speaker 1: Okay, I don't know that I'm on the right track here,

146
00:07:13,880 --> 00:07:16,360
Speaker 1: but I'm going to talk about augmented intelligence exactly what

147
00:07:16,440 --> 00:07:18,600
Speaker 1: that means and how that fits into it.

148
00:07:19,000 --> 00:07:21,960
Speaker 2: Yeah, I've talked about augmented intelligence it feels like for

149
00:07:22,000 --> 00:07:24,800
Speaker 2: many many years, because I think that's I like to say,

150
00:07:24,840 --> 00:07:27,280
Speaker 2: that's what I think AI really means. It's not so

151
00:07:27,400 --> 00:07:30,840
Speaker 2: much the artificial intelligence I mean, strictly speaking, it is,

152
00:07:30,880 --> 00:07:34,360
Speaker 2: but I prefer to think about AI as augmented intelligence.

153
00:07:34,440 --> 00:07:38,360
Speaker 2: How can we augment human intelligence, human capability, How can

154
00:07:38,360 --> 00:07:40,480
Speaker 2: we help in the data day of what we're doing,

155
00:07:40,680 --> 00:07:43,240
Speaker 2: and in the case of zero, how can we help

156
00:07:43,280 --> 00:07:45,120
Speaker 2: small businesses in the day to day of running of

157
00:07:45,160 --> 00:07:48,400
Speaker 2: their operations. And so when we think about augmented intelligence,

158
00:07:48,440 --> 00:07:51,280
Speaker 2: we think about how can we learn more about how

159
00:07:51,280 --> 00:07:53,760
Speaker 2: your business runs and how can we help you automate

160
00:07:53,960 --> 00:07:57,400
Speaker 2: those routine tasks in those workflows. And that's something which

161
00:07:57,640 --> 00:08:01,040
Speaker 2: in our case Jacks does orchestrating multiple AI agents behind

162
00:08:01,040 --> 00:08:03,080
Speaker 2: the scenes to help cut busy work out of the

163
00:08:03,080 --> 00:08:05,440
Speaker 2: equation or to reduce toil as a phrase I use

164
00:08:05,480 --> 00:08:08,720
Speaker 2: as well. But the goal is really to augment the capability,

165
00:08:08,760 --> 00:08:13,240
Speaker 2: to amplify the capability of the small business themselves. There's

166
00:08:13,280 --> 00:08:15,560
Speaker 2: nothing artificial about that at all. It's just that the

167
00:08:15,600 --> 00:08:18,680
Speaker 2: capability that we're using to leverage that is traditional AI.

168
00:08:19,120 --> 00:08:21,040
Speaker 1: So what's the bottom line here for SMS? You know,

169
00:08:21,080 --> 00:08:27,560
Speaker 1: in five years a well operated small business, it should

170
00:08:27,560 --> 00:08:32,360
Speaker 1: be fully augmented or artificial intelligence reliant for one of

171
00:08:32,400 --> 00:08:35,400
Speaker 1: a better term. But I mean, the goal presumably is

172
00:08:35,400 --> 00:08:37,040
Speaker 1: for a small business to make more money, I mean,

173
00:08:37,720 --> 00:08:41,520
Speaker 1: and presumably this is where AI comes into it.

174
00:08:41,520 --> 00:08:43,800
Speaker 2: It helps with the scaling part. Right, If you're a

175
00:08:43,840 --> 00:08:47,040
Speaker 2: small business and you're trying to reach more customers, you're

176
00:08:47,040 --> 00:08:50,440
Speaker 2: trying to meet more needs. Then to do that at scale,

177
00:08:50,960 --> 00:08:53,560
Speaker 2: what are the capabilities the tools that can help you

178
00:08:53,600 --> 00:08:56,640
Speaker 2: do that? And AI is part of that toolkit increasingly,

179
00:08:57,240 --> 00:08:59,800
Speaker 2: and sometimes it's not AI itself, it's more that the

180
00:08:59,800 --> 00:09:03,280
Speaker 2: tools that you are using are becoming more artificially intelligent,

181
00:09:03,360 --> 00:09:06,120
Speaker 2: you know, and additional capabilities of being brought online by

182
00:09:06,160 --> 00:09:09,240
Speaker 2: those So I don't know if it's every five years

183
00:09:09,280 --> 00:09:14,719
Speaker 2: every small businesses there's AI enabled or dependent or I

184
00:09:14,760 --> 00:09:17,160
Speaker 2: think it's more a matter of as time goes on,

185
00:09:17,640 --> 00:09:20,960
Speaker 2: how do small businesses best leverage the capabilities that are

186
00:09:21,000 --> 00:09:23,000
Speaker 2: out there in a way that helps them achieve whatever

187
00:09:23,040 --> 00:09:24,920
Speaker 2: it is that they're trying to achieve. And if that

188
00:09:25,040 --> 00:09:27,920
Speaker 2: enables them to grow and to scale and really importantly

189
00:09:28,120 --> 00:09:31,280
Speaker 2: give them time back, that's a great thing. And really

190
00:09:31,360 --> 00:09:32,680
Speaker 2: that's what we try and do with the work that

191
00:09:32,720 --> 00:09:34,320
Speaker 2: we're doing, and I know others do as well.

192
00:09:34,720 --> 00:09:36,360
Speaker 1: James, thank you for talking to Fear and Greed.

193
00:09:36,800 --> 00:09:38,640
Speaker 2: My pleasure to Sewan, thanks very much for having me.

194
00:09:38,559 --> 00:09:42,120
Speaker 1: As James Bergen, executive general manager of Technology Research and

195
00:09:42,200 --> 00:09:44,959
Speaker 1: Advocacy at Zero, a great supporter of this podcast. I'm

196
00:09:45,000 --> 00:09:47,559
Speaker 1: Sean Aylmer and this is Dare and Reed Q and

197
00:09:47,640 --> 00:09:48,840
Speaker 1: Daity