WEBVTT

1
00:00:02.978 --> 00:00:05.160
(soft music)

2
00:00:05.160 --> 00:00:08.830
We're in an era of amazing medical breakthroughs.

3
00:00:09.870 --> 00:00:12.930
Some people have referred to it as a golden era of medicine,

4
00:00:12.930 --> 00:00:14.700
but the paradox of it is

5
00:00:14.700 --> 00:00:17.160
that the clinical development process

6
00:00:17.160 --> 00:00:19.410
still remains challenging.

7
00:00:19.410 --> 00:00:22.980
It takes a very long time, it's very costly,

8
00:00:22.980 --> 00:00:27.540
and it has enormous potential to be improved

9
00:00:27.540 --> 00:00:30.870
through further efficiencies and speed.

10
00:00:30.870 --> 00:00:33.540
Patients are waiting for these therapies,

11
00:00:33.540 --> 00:00:36.300
and the sooner that we can get them to patients,

12
00:00:36.300 --> 00:00:37.680
the sooner that those patients

13
00:00:37.680 --> 00:00:39.330
can benefit from those therapies.

14
00:00:41.190 --> 00:00:45.870
AI has an enormous potential to help streamline

15
00:00:45.870 --> 00:00:47.910
and accelerate clinical development.

16
00:00:47.910 --> 00:00:49.920
There are strategic decisions

17
00:00:49.920 --> 00:00:51.900
that need to be made at the beginning of a trial,

18
00:00:51.900 --> 00:00:56.100
around the study design, which countries, which sites,

19
00:00:56.100 --> 00:01:00.150
all of those decisions can be optimized.

20
00:01:00.150 --> 00:01:04.110
This undertaking to create an AI-based clinical trial

21
00:01:04.110 --> 00:01:06.090
is one in which you have to simultaneously

22
00:01:06.090 --> 00:01:09.030
solve multiple problems.

23
00:01:09.030 --> 00:01:12.180
And working with BCG,

24
00:01:12.180 --> 00:01:15.990
we were able to identify the key workstreams

25
00:01:15.990 --> 00:01:17.580
that were critical.

26
00:01:17.580 --> 00:01:19.200
You had to get the data right,

27
00:01:19.200 --> 00:01:21.480
you had to build a modern data platform

28
00:01:21.480 --> 00:01:23.970
that was machine learning optimized,

29
00:01:23.970 --> 00:01:27.420
a data science team capable of building predictive models,

30
00:01:27.420 --> 00:01:29.160
intelligent applications

31
00:01:29.160 --> 00:01:32.280
that our teams could actually benefit from,

32
00:01:32.280 --> 00:01:34.260
and if any of those did not work,

33
00:01:34.260 --> 00:01:36.033
the entire thing does not work.

34
00:01:36.960 --> 00:01:39.270
Responsible AI was critical

35
00:01:39.270 --> 00:01:42.720
in how we thought about bringing AI to Syneos Health.

36
00:01:42.720 --> 00:01:45.690
In many cases the way we think about it is

37
00:01:45.690 --> 00:01:48.180
AI produces the first recommendation,

38
00:01:48.180 --> 00:01:49.350
but it's always filtered

39
00:01:49.350 --> 00:01:53.043
through the expertise and judgment of an expert human.

40
00:01:54.540 --> 00:01:58.620
We've seen enormous early indications of progress

41
00:01:58.620 --> 00:02:00.990
in terms of acceleration.

42
00:02:00.990 --> 00:02:04.470
We have been able to ultimately make better decisions

43
00:02:04.470 --> 00:02:06.120
at the strategy level,

44
00:02:06.120 --> 00:02:09.630
and share with sponsors the rationale, the evidence,

45
00:02:09.630 --> 00:02:12.780
the data behind our recommendations.

46
00:02:12.780 --> 00:02:14.190
I don't think we would've made

47
00:02:14.190 --> 00:02:15.630
anywhere close to that progress

48
00:02:15.630 --> 00:02:18.780
without the expertise of BCGX.

49
00:02:18.780 --> 00:02:19.950
It used to be the case,

50
00:02:19.950 --> 00:02:21.930
where the vast majority of the time

51
00:02:21.930 --> 00:02:23.880
that data scientists spent,

52
00:02:23.880 --> 00:02:27.990
was on the preparation, cleaning, mastering of data,

53
00:02:27.990 --> 00:02:30.810
so it could be useful for analysis.

54
00:02:30.810 --> 00:02:32.670
Now it's the inverse.

55
00:02:32.670 --> 00:02:35.040
We have created through our data platform,

56
00:02:35.040 --> 00:02:37.050
our data hygiene efforts,

57
00:02:37.050 --> 00:02:38.430
the ability for data scientists

58
00:02:38.430 --> 00:02:39.840
to spend only 20% of their time

59
00:02:39.840 --> 00:02:42.420
with the basic requirements around data,

60
00:02:42.420 --> 00:02:46.950
and 80% of their time delivering value through analysis.

61
00:02:46.950 --> 00:02:51.950
The potential for AI to dramatically help biotechs

62
00:02:52.890 --> 00:02:56.070
realize the value of their asset more quickly,

63
00:02:56.070 --> 00:02:57.750
with greater certainty

64
00:02:57.750 --> 00:03:01.683
and greater ultimate commercial impact is massive.

65
00:03:02.730 --> 00:03:06.150
This whole idea of accelerating and maximizing

66
00:03:06.150 --> 00:03:08.760
the impact of life-enhancing therapies

67
00:03:08.760 --> 00:03:11.643
really drives the teams at Syneos Health.