WEBVTT

1
00:00:00.083 --> 00:00:02.210
What challenges are
organizations running into

2
00:00:02.210 --> 00:00:04.254
as they try to adopt AI

3
00:00:04.254 --> 00:00:06.006
inside their engineering teams?

4
00:00:06.006 --> 00:00:07.424
One of the key things that we're seeing

5
00:00:07.424 --> 00:00:09.884
is organizations are rolling out AI tools

6
00:00:09.884 --> 00:00:11.011
for their software engineers

7
00:00:11.011 --> 00:00:13.221
and focusing on "adoption."

8
00:00:13.221 --> 00:00:15.640
So are people actually using the tools?

9
00:00:15.640 --> 00:00:17.308
And what we find is that's not enough

10
00:00:17.308 --> 00:00:19.811
because if you're just using the tool

11
00:00:19.811 --> 00:00:21.104
to answer a question,

12
00:00:21.104 --> 00:00:23.064
it's not actually changing
how you're working.

13
00:00:23.064 --> 00:00:25.191
So, Matt, you heard him explain

14
00:00:25.191 --> 00:00:26.568
some of these huge challenges.

15
00:00:26.568 --> 00:00:28.403
You're not just using
the tool to use the tool.

16
00:00:28.403 --> 00:00:31.740
So how did BCG and AWS
come together to partner

17
00:00:31.740 --> 00:00:35.368
to help teams at AWS
overcome these challenges?

18
00:00:35.368 --> 00:00:36.953
Yeah, well, first of all,
I couldn't agree more.

19
00:00:36.953 --> 00:00:38.413
It's not just about the adoption,

20
00:00:38.413 --> 00:00:39.414
it's meaningful usage.

21
00:00:39.414 --> 00:00:40.498
And in our case, that means

22
00:00:40.498 --> 00:00:43.084
are you actually deploying
code to production,

23
00:00:43.084 --> 00:00:44.335
and is it actually making it

24
00:00:44.335 --> 00:00:45.754
to customers in the form of features?

25
00:00:45.754 --> 00:00:46.963
So the first thing we did

26
00:00:46.963 --> 00:00:48.590
is spend a lot of time upfront:

27
00:00:48.590 --> 00:00:49.716
How are we gonna measure this?

28
00:00:49.716 --> 00:00:51.760
Like, how are we gonna measure
adoption in the first place,

29
00:00:51.760 --> 00:00:52.469
but how are we gonna measure

30
00:00:52.469 --> 00:00:56.347
that translation from
adoption to POC to production

31
00:00:56.347 --> 00:00:57.557
to actually that meaningful usage?

32
00:00:57.557 --> 00:01:00.185
So we established some
topline metrics together,

33
00:01:00.185 --> 00:01:01.394
and then, from there, we could set

34
00:01:01.394 --> 00:01:02.604
some goals for the project.

35
00:01:02.604 --> 00:01:05.774
What is BCG offering

36
00:01:05.774 --> 00:01:08.526
to help clients improve productivity?

37
00:01:08.526 --> 00:01:10.028
So first, you can talk about those tools,

38
00:01:10.028 --> 00:01:11.237
and then I want get back to you

39
00:01:11.237 --> 00:01:13.448
about what BCG is offering.

40
00:01:13.448 --> 00:01:14.491
Yeah, well, you know, first of all,

41
00:01:14.491 --> 00:01:16.701
it's actually not as much about the tools.

42
00:01:16.701 --> 00:01:18.244
Like yes, you need to
be accessing your tools,

43
00:01:18.244 --> 00:01:19.871
but it's more about the expertise.

44
00:01:19.871 --> 00:01:21.915
And so what we found was hugely beneficial

45
00:01:21.915 --> 00:01:23.541
after we established what the metrics are,

46
00:01:23.541 --> 00:01:25.543
like how are we gonna
measure meaningful adoption,

47
00:01:25.543 --> 00:01:28.129
is embedding people
directly into the teams

48
00:01:28.129 --> 00:01:30.840
who can credibly speak
to using these tools

49
00:01:30.840 --> 00:01:32.425
and actually getting to the outcomes

50
00:01:32.425 --> 00:01:33.259
that we both wanted.

51
00:01:33.259 --> 00:01:35.553
So you heard him talk about
embedding your experts,

52
00:01:35.553 --> 00:01:37.764
but what is BCG doing specifically

53
00:01:37.764 --> 00:01:39.599
to help this process?

54
00:01:39.599 --> 00:01:42.060
Yeah, so, it's actually quite challenging

55
00:01:42.060 --> 00:01:44.896
to get the behavior change to happen.

56
00:01:44.896 --> 00:01:47.023
What we found is there
are actually three reasons

57
00:01:47.023 --> 00:01:49.359
why software developers really resist

58
00:01:49.359 --> 00:01:50.902
using the tools effectively.

59
00:01:50.902 --> 00:01:51.903
One is they don't know

60
00:01:51.903 --> 00:01:53.696
that it can do what it can do.

61
00:01:53.696 --> 00:01:55.073
The second is they have habits.

62
00:01:55.073 --> 00:01:55.907
You know, we're all human,

63
00:01:55.907 --> 00:01:57.408
we do things the way that we do them.

64
00:01:57.408 --> 00:01:59.744
And then the third is there's
this actual real sense

65
00:01:59.744 --> 00:02:01.162
of sort of identity threat.

66
00:02:01.162 --> 00:02:04.165
People worry that if the
AI tool can do what they do

67
00:02:04.165 --> 00:02:06.584
that it makes them sort of
irrelevant to the world.

68
00:02:06.584 --> 00:02:09.671
And so there's a lot of
sort of fear and resistance.

69
00:02:09.671 --> 00:02:11.256
So what we found is you have

70
00:02:11.256 --> 00:02:12.757
to get over all three of these things.

71
00:02:12.757 --> 00:02:15.176
And so what we do: One,
is you do some training.

72
00:02:15.176 --> 00:02:17.262
So you have to tell them
how to use the tools

73
00:02:17.262 --> 00:02:18.096
and have to show them

74
00:02:18.096 --> 00:02:19.347
how to use the tools effectively.

75
00:02:19.347 --> 00:02:20.807
So, you know, how do you actually

76
00:02:20.807 --> 00:02:22.392
explicitly use the tool?

77
00:02:22.392 --> 00:02:24.435
But the second is you have
to give them coaching.

78
00:02:24.435 --> 00:02:27.272
And so we need to have peer engineers,

79
00:02:27.272 --> 00:02:31.276
and so we will bring in peer
engineers from our team.

80
00:02:31.276 --> 00:02:33.444
And then we'll actually, with our clients,

81
00:02:33.444 --> 00:02:35.780
help them develop their own coaches

82
00:02:35.780 --> 00:02:37.115
that are peer engineers

83
00:02:37.115 --> 00:02:39.075
that are experts in using AI.

84
00:02:39.075 --> 00:02:40.118
That's such a key point

85
00:02:40.118 --> 00:02:42.620
because everyone seems threatened by AI.

86
00:02:42.620 --> 00:02:45.498
But what you're saying is
it's not gonna threaten you,

87
00:02:45.498 --> 00:02:47.542
it's going to enhance you.

88
00:02:47.542 --> 00:02:49.502
What about your key goals
for next year, Matt?

89
00:02:49.502 --> 00:02:51.087
What are the key goals?

90
00:02:51.087 --> 00:02:51.921
Yeah, well the key goals

91
00:02:51.921 --> 00:02:53.506
are really doubling
down on what we learned.

92
00:02:53.506 --> 00:02:55.466
Like year to date, and thanks,

93
00:02:55.466 --> 00:02:56.718
you know, to the work we've done with BCG,

94
00:02:56.718 --> 00:02:59.053
we have a 31% productivity improvement.

95
00:02:59.053 --> 00:03:02.182
And we've actually shipped
27% more features this year.

96
00:03:02.182 --> 00:03:03.516
And sort of getting back
to what you were saying,

97
00:03:03.516 --> 00:03:05.018
like, that's empowering for developers.

98
00:03:05.018 --> 00:03:06.769
They went from maybe some of them,

99
00:03:06.769 --> 00:03:08.980
some pockets of resistance
to feeling empowered.

100
00:03:08.980 --> 00:03:10.982
‘Cause they're shipping more features,

101
00:03:10.982 --> 00:03:12.984
they're shipping more coded production.

102
00:03:12.984 --> 00:03:14.319
And we want to, you know,
double down on that.

103
00:03:14.319 --> 00:03:16.070
End of the day, what
does a developer want?

104
00:03:16.070 --> 00:03:18.531
They want their code in
the hands of customers

105
00:03:18.531 --> 00:03:19.449
and customers using that.

106
00:03:19.449 --> 00:03:22.452
And they're able to do
that faster and do more.

107
00:03:22.452 --> 00:03:24.704
So we're gonna do more.

108
00:03:24.704 --> 00:03:27.248
(upbeat music)