1
00:00:00,600 --> 00:00:05,440
Speaker 1: Picture this. You're staring at chat GPT, wondering if you're

2
00:00:05,480 --> 00:00:10,840
Speaker 1: barely scratching the surface of what's possible. Maybe you've dabbled

3
00:00:10,840 --> 00:00:14,239
Speaker 1: with generating a few emails, or perhaps you're already using

4
00:00:14,240 --> 00:00:17,760
Speaker 1: it daily, but like most of us, there's that nagging

5
00:00:17,960 --> 00:00:22,680
Speaker 1: feeling you could be doing so much more. At south

6
00:00:22,720 --> 00:00:27,000
Speaker 1: By Southwest, Sydney last year, I remember futurist Brian Johnson

7
00:00:27,080 --> 00:00:30,520
Speaker 1: dropped a truth bomb that stuck with me. He said,

8
00:00:30,600 --> 00:00:33,560
Speaker 1: AI is not taking your job. It's a human who

9
00:00:33,640 --> 00:00:38,320
Speaker 1: is using AI who's taking your job. That hit home

10
00:00:39,360 --> 00:00:41,960
Speaker 1: and for the past two years I have been living

11
00:00:42,000 --> 00:00:45,720
Speaker 1: and breathing jen Ai, making it an essential part of

12
00:00:45,760 --> 00:00:50,720
Speaker 1: my toolkit, and through my consultancy Inventium, I've seen firsthand

13
00:00:50,800 --> 00:00:54,400
Speaker 1: how proper AI training that we've been doing with many

14
00:00:54,520 --> 00:01:00,040
Speaker 1: of our clients can transform the way people work. That

15
00:01:00,240 --> 00:01:04,680
Speaker 1: is why I'm thrilled to bring you something special over

16
00:01:04,720 --> 00:01:07,560
Speaker 1: the next four tuesdays, I am going to be joined

17
00:01:07,560 --> 00:01:13,279
Speaker 1: by Inventium's resident at Genai Guru Neo Applin, and together

18
00:01:13,560 --> 00:01:16,920
Speaker 1: we'll share some game changing tips that will help you

19
00:01:17,080 --> 00:01:22,319
Speaker 1: level up how you use Genai. Just quickly, before we

20
00:01:22,319 --> 00:01:25,440
Speaker 1: get into today's episode if you want to go defilt

21
00:01:25,480 --> 00:01:28,800
Speaker 1: and learn how to work with Ai instead of worrying

22
00:01:28,840 --> 00:01:31,920
Speaker 1: about being replaced by it. I have something that I

23
00:01:31,959 --> 00:01:36,120
Speaker 1: think you will love. But my consultancy Inventium, We've created

24
00:01:36,160 --> 00:01:40,240
Speaker 1: the Genai Productivity System. It's a self paced online course

25
00:01:40,280 --> 00:01:44,160
Speaker 1: designed to help you boost productivity and save literally hours

26
00:01:44,240 --> 00:01:47,960
Speaker 1: every week by using Jenai a whole lot smarter. In

27
00:01:48,120 --> 00:01:51,240
Speaker 1: just a few hours, you'll go from Jenai dabbler to

28
00:01:51,400 --> 00:01:55,600
Speaker 1: Jeni expert with practical tools you can start using immediately.

29
00:01:56,240 --> 00:02:00,000
Speaker 1: Head to inventium dot com, dot au, forward slash Jenai

30
00:02:00,440 --> 00:02:07,920
Speaker 1: to check it out and get started today. Welcome to

31
00:02:08,000 --> 00:02:11,840
Speaker 1: How I Work, a show about habits, rituals, and strategies

32
00:02:11,960 --> 00:02:15,800
Speaker 1: for optimizing your day. I'm your host, doctor Amantha Imber.

33
00:02:16,480 --> 00:02:19,600
Speaker 1: So many people assume that Jenai should just get it

34
00:02:19,680 --> 00:02:23,919
Speaker 1: right from the start. What is the biggest misconception about

35
00:02:23,960 --> 00:02:26,200
Speaker 1: how Ai understands prompts.

36
00:02:27,000 --> 00:02:30,440
Speaker 2: The biggest misconception is actually that Jenai will get it

37
00:02:30,520 --> 00:02:32,880
Speaker 2: right from the start and it will understand exactly what

38
00:02:32,919 --> 00:02:35,760
Speaker 2: you really want to achieve from the words you've actually

39
00:02:35,800 --> 00:02:39,600
Speaker 2: written within that prompt, so it will make an assumption

40
00:02:40,200 --> 00:02:42,919
Speaker 2: trying to make it easier for you. So you need

41
00:02:42,960 --> 00:02:45,320
Speaker 2: to give the context behind what you're trying to achieve

42
00:02:45,680 --> 00:02:47,919
Speaker 2: and all the backgrounds so it can do a great job.

43
00:02:48,400 --> 00:02:52,079
Speaker 2: And so the misconception is unless you that it will

44
00:02:52,120 --> 00:02:56,120
Speaker 2: give that great answer without you giving all that background.

45
00:02:56,760 --> 00:03:00,680
Speaker 1: So there's quite a counterintuitive approach that you said in

46
00:03:00,760 --> 00:03:04,560
Speaker 1: terms of getting Ai to interview us instead of just

47
00:03:04,760 --> 00:03:07,560
Speaker 1: giving it instructions. So can you tell me a bit

48
00:03:07,600 --> 00:03:10,720
Speaker 1: about how this works and why it is so effective.

49
00:03:11,080 --> 00:03:14,200
Speaker 2: Yeah, it's like using a Jenai like a person. So

50
00:03:14,400 --> 00:03:17,160
Speaker 2: if you've got a work colleague or you've got an

51
00:03:17,240 --> 00:03:20,040
Speaker 2: expert you've brought in to do something, you don't just say, hey,

52
00:03:20,120 --> 00:03:21,679
Speaker 2: look i'd like you to do this and then just

53
00:03:21,760 --> 00:03:25,239
Speaker 2: let them go. There's always a conversation involved, so they're

54
00:03:25,280 --> 00:03:28,440
Speaker 2: asking what you wanted to achieve, what the background is.

55
00:03:28,600 --> 00:03:32,480
Speaker 2: They'll ask you clarifying questions about it. Treating Jenai like

56
00:03:32,560 --> 00:03:35,960
Speaker 2: that expert or like someone within your team changes the

57
00:03:36,080 --> 00:03:39,880
Speaker 2: mindset of you, the person working with Jenai, so that

58
00:03:40,280 --> 00:03:42,440
Speaker 2: it'll then have a conversation with you so you can

59
00:03:42,480 --> 00:03:45,240
Speaker 2: give the information to it. That way, you don't have

60
00:03:45,280 --> 00:03:47,960
Speaker 2: to feel like you're writing a massive tome before you

61
00:03:48,160 --> 00:03:50,840
Speaker 2: put a prompt in so you can say, hey, look,

62
00:03:51,080 --> 00:03:53,040
Speaker 2: ask me as many questions as you need so you

63
00:03:53,040 --> 00:03:55,280
Speaker 2: can do a great job. And then we've got those

64
00:03:55,360 --> 00:03:58,520
Speaker 2: questions going with a conversation going, then it understands the

65
00:03:58,520 --> 00:04:00,720
Speaker 2: background and it'll do a much better because of it.

66
00:04:01,120 --> 00:04:04,200
Speaker 1: So let's take a real world example. Say I'm writing

67
00:04:04,440 --> 00:04:09,000
Speaker 1: a sales email to a prospect. How would you suggest

68
00:04:09,040 --> 00:04:13,400
Speaker 1: that I prompt AI in order to get the best

69
00:04:13,440 --> 00:04:15,400
Speaker 1: results to get this amazing sales email.

70
00:04:16,160 --> 00:04:18,520
Speaker 2: Yeah, okay, So the first thing I do is say,

71
00:04:19,240 --> 00:04:20,960
Speaker 2: these are the goals that I want to achieve with

72
00:04:21,000 --> 00:04:23,839
Speaker 2: this sales email. Don't go straight to the sales email

73
00:04:23,920 --> 00:04:27,120
Speaker 2: content itself. The reason for that is that GENAI then

74
00:04:27,200 --> 00:04:29,919
Speaker 2: knows loosely what the outcome needs to fulfill, what the

75
00:04:29,960 --> 00:04:32,800
Speaker 2: goals that it needs to fulfill are. Then tell it

76
00:04:32,839 --> 00:04:34,520
Speaker 2: what you're wanting to achieve. Look, I want a sales

77
00:04:34,520 --> 00:04:36,920
Speaker 2: email with these and these and these information in it.

78
00:04:37,360 --> 00:04:40,440
Speaker 2: And then you'll say, hey, before you do any work,

79
00:04:40,800 --> 00:04:43,440
Speaker 2: don't do any work, But before you do any work,

80
00:04:43,480 --> 00:04:45,000
Speaker 2: what I want you to do is ask me is

81
00:04:45,080 --> 00:04:48,520
Speaker 2: many clarifying questions that you need until you're ninety five

82
00:04:48,520 --> 00:04:51,040
Speaker 2: percent confident that you can do a job right now,

83
00:04:51,040 --> 00:04:54,520
Speaker 2: people ask me why ninety five percent, And it's about

84
00:04:54,560 --> 00:04:57,039
Speaker 2: a problem with English language, Like if you say I

85
00:04:57,080 --> 00:04:59,760
Speaker 2: want you to be kind of confident or mostly confident,

86
00:05:00,240 --> 00:05:03,400
Speaker 2: how confident is that? Whereas if you're really specific and

87
00:05:03,440 --> 00:05:06,039
Speaker 2: you're saying ninety five percent confident, it means don't ask

88
00:05:06,080 --> 00:05:09,360
Speaker 2: me a thousand questions, ask me enough questions until you've

89
00:05:09,400 --> 00:05:12,120
Speaker 2: got a good background and you feel very confident you

90
00:05:12,160 --> 00:05:14,520
Speaker 2: can do the job right, So ask me as many

91
00:05:14,600 --> 00:05:17,560
Speaker 2: questions as you as you need so you're ninety five

92
00:05:17,600 --> 00:05:19,720
Speaker 2: percent confident in getting the job done well.

93
00:05:20,320 --> 00:05:23,760
Speaker 1: So what kind of questions does the AI tend to

94
00:05:23,920 --> 00:05:27,000
Speaker 1: ask when we use this technique.

95
00:05:26,720 --> 00:05:30,359
Speaker 2: It's a blend of kind of one oh one questions

96
00:05:30,520 --> 00:05:34,520
Speaker 2: and really detailed ones, so it's clarifying on the goals

97
00:05:34,560 --> 00:05:37,400
Speaker 2: and things like audience and things like that. It'll ask,

98
00:05:37,720 --> 00:05:41,000
Speaker 2: but it will also ask particular specific things about the

99
00:05:41,040 --> 00:05:45,200
Speaker 2: topic you've got. So if it's a sales style example

100
00:05:45,240 --> 00:05:48,720
Speaker 2: we're going through, for example, it will ask questions on

101
00:05:48,760 --> 00:05:51,839
Speaker 2: the client, the client's problem. It'll may ask questions on

102
00:05:51,880 --> 00:05:55,080
Speaker 2: the budget, even those kind of things, because it wants

103
00:05:55,120 --> 00:05:58,039
Speaker 2: to have a full basis of all the background so

104
00:05:58,040 --> 00:05:59,120
Speaker 2: it can give you the best result.

105
00:06:00,120 --> 00:06:03,680
Speaker 1: What happens if the AI asks me a question that

106
00:06:03,760 --> 00:06:06,040
Speaker 1: I just don't know how to answer, what should I do?

107
00:06:06,520 --> 00:06:09,120
Speaker 2: Yeah, you got two choices there. One is you can

108
00:06:09,160 --> 00:06:12,039
Speaker 2: say I'm not sure and that's enough and that's fine,

109
00:06:12,080 --> 00:06:14,920
Speaker 2: there's no problems with that. The other one I use

110
00:06:14,960 --> 00:06:17,440
Speaker 2: a lot is I'm not sure, but I would like

111
00:06:17,520 --> 00:06:19,919
Speaker 2: some suggestions on different avenues. We can go here, and

112
00:06:19,960 --> 00:06:22,320
Speaker 2: I want you to give me those suggestions, and so

113
00:06:22,360 --> 00:06:24,680
Speaker 2: then it'll know that it has another conversation and other

114
00:06:25,040 --> 00:06:26,560
Speaker 2: round of questions and answers for you.

115
00:06:27,240 --> 00:06:30,520
Speaker 1: So is it important to add specific areas for the

116
00:06:30,560 --> 00:06:35,360
Speaker 1: AI to probe, like, you know, around what's my brand

117
00:06:35,440 --> 00:06:39,599
Speaker 1: voice or industry terminology for example. That's up to you.

118
00:06:40,560 --> 00:06:43,760
Speaker 2: I'd say that for probably fifty percent of the time,

119
00:06:43,960 --> 00:06:47,240
Speaker 2: I just say the ninety five percent confidence and that's enough.

120
00:06:47,839 --> 00:06:50,680
Speaker 2: But other times, when I'm working on something very specific,

121
00:06:51,160 --> 00:06:53,480
Speaker 2: I will get it to do questions in a very

122
00:06:53,480 --> 00:06:57,600
Speaker 2: specific area, particularly if i'm so if it's a deep analysis,

123
00:06:57,680 --> 00:06:59,920
Speaker 2: then it might be I want you to ask questions

124
00:07:00,000 --> 00:07:02,240
Speaker 2: in this particular direction so I know that I've got

125
00:07:02,279 --> 00:07:05,520
Speaker 2: the analysis right. So or it might be for writing

126
00:07:05,640 --> 00:07:08,200
Speaker 2: it will be very much on the style or tone,

127
00:07:08,480 --> 00:07:11,240
Speaker 2: making sure that that matches my audience, those kind of things.

128
00:07:11,600 --> 00:07:14,000
Speaker 1: And is there a limit to how many rounds of

129
00:07:14,080 --> 00:07:17,320
Speaker 1: Q and A we should go through or like, you know,

130
00:07:17,600 --> 00:07:19,640
Speaker 1: because I feel like I could just keep refining and

131
00:07:19,680 --> 00:07:22,560
Speaker 1: refining until we hit that ninety five percent confidence level.

132
00:07:22,600 --> 00:07:24,280
Speaker 1: But how many rounds am I going through?

133
00:07:25,240 --> 00:07:27,360
Speaker 2: Up to you, And it really depends on the work

134
00:07:27,400 --> 00:07:30,160
Speaker 2: you're doing. What I've found is that a lot of

135
00:07:30,160 --> 00:07:33,120
Speaker 2: the Jenai will just ask that one round of questions,

136
00:07:33,240 --> 00:07:35,240
Speaker 2: you answer it, and then it'll just go great and done.

137
00:07:36,120 --> 00:07:39,760
Speaker 2: What I've found is if I get the responses back

138
00:07:39,760 --> 00:07:41,480
Speaker 2: from the Jenai and I've got a bit of a

139
00:07:41,520 --> 00:07:45,400
Speaker 2: gut feeling that potentially it may not know everything, or

140
00:07:45,440 --> 00:07:48,200
Speaker 2: potentially it might be going in a different direction, what

141
00:07:48,320 --> 00:07:51,320
Speaker 2: I'll do then I'll say, do you have any more questions?

142
00:07:51,640 --> 00:07:54,320
Speaker 2: So it's actually you need to prompt it more to

143
00:07:54,360 --> 00:07:58,000
Speaker 2: ask you questions that it's unlikely that it'll just want

144
00:07:58,040 --> 00:08:00,360
Speaker 2: to ask questions forever and ever. So it's more of

145
00:08:00,400 --> 00:08:03,680
Speaker 2: a work hard to get more questions than get fewer.

146
00:08:03,960 --> 00:08:06,760
Speaker 1: Okay, So for someone listening today who wants to try

147
00:08:06,800 --> 00:08:10,280
Speaker 1: this technique, what is one simple way to just get started.

148
00:08:10,920 --> 00:08:14,040
Speaker 2: Yeah, ask your prompt, and then at the bottom of

149
00:08:14,040 --> 00:08:16,200
Speaker 2: you prompt, just the normal prompt you'd put in, So

150
00:08:16,240 --> 00:08:18,320
Speaker 2: don't worry about the prompt, just do your normal work there,

151
00:08:18,600 --> 00:08:21,440
Speaker 2: and then at the bottom you say, before you respond,

152
00:08:21,720 --> 00:08:24,120
Speaker 2: ask me any questions you've got until you're ninety five

153
00:08:24,120 --> 00:08:26,520
Speaker 2: percent confident in doing a great job. That's it.

154
00:08:26,920 --> 00:08:31,120
Speaker 1: I love that and I personally use this technique multiple

155
00:08:31,160 --> 00:08:35,120
Speaker 1: times a day. It's completely transformed how I use Jenai.

156
00:08:35,360 --> 00:08:38,880
Speaker 1: So I hope that for people listening that you have

157
00:08:39,000 --> 00:08:43,679
Speaker 1: the confidence now to try and apply this today. If

158
00:08:43,720 --> 00:08:47,120
Speaker 1: you're keen to learn even more Genai tricks and hacks,

159
00:08:47,320 --> 00:08:51,680
Speaker 1: check out Inventium's Genai Productivity System. It's a self faced

160
00:08:51,679 --> 00:08:56,199
Speaker 1: course to transform you from a Jenai dabbler to jenaipro

161
00:08:56,520 --> 00:09:00,160
Speaker 1: in just a few hours. Head to inventium dot com

162
00:09:00,360 --> 00:09:03,840
Speaker 1: a forward slash jnai to check it out and get

163
00:09:03,840 --> 00:09:07,320
Speaker 1: started today. If you like today's show, make sure you

164
00:09:07,400 --> 00:09:10,440
Speaker 1: hit follow on your podcast app to be alerted when

165
00:09:10,520 --> 00:09:13,800
Speaker 1: new episodes drop. How I Work was recorded on the

166
00:09:13,840 --> 00:09:16,600
Speaker 1: traditional land of the Warrangery people, part of the kul

167
00:09:16,640 --> 00:09:19,160
Speaker 1: And Nation. A big thank you to Martin Nimber for

168
00:09:19,240 --> 00:09:20,280
Speaker 1: doing the sound mix,