Listen to Audio
Listen Anywhere
September 2026
In this episode, Jürgen Wiedemann, Head of Pharma Technology Systems, and Christoph Ebbecke, Venture Manager, join the podcast to discuss Bayer AG's SOLIDA-1 facility. Learn how the project team developed an innovative ballroom concept with integrated end-to-end Autonomous Operations, and more highlights about the 2026 ISPE Facility of the Year Awards (FOYA) Operational Excellence and Pharma 4.0™ category winners.
1
00:00:00,000 --> 00:00:09,840
Welcome to the ISPE podcast, shaping the future of pharma, where ISPE supports you on your journey, fueling innovation, sharing insights,
2
00:00:09,840 --> 00:00:14,160
thought leadership, and empowering a global community to reimagine what's possible.
3
00:00:15,275 --> 00:00:20,555
Hello, and welcome to the ISPE podcast, shaping the future of pharma.
4
00:00:21,035 --> 00:00:22,795
I'm Bob Chew, your host.
5
00:00:22,955 --> 00:00:32,880
And today, we have another episode where we'll be sharing the latest insights and thought leadership on manufacturing technology, supply chains,
6
00:00:32,880 --> 00:00:36,880
and regulatory trends impacting the pharmaceutical industry.
7
00:00:37,280 --> 00:00:43,680
You will hear directly from the innovators, experts, and professionals driving progress and shaping the future.
8
00:00:44,405 --> 00:00:48,885
Thank you again for joining us, and now let's dive into this episode.
9
00:00:49,525 --> 00:00:58,165
Our topic today is Bayer AG's SOLIDA-1 facility of the year award category winning facility.
10
00:00:59,040 --> 00:01:08,240
To share more about this topic, I would like to welcome Juergen Wiedemann, Head of Pharma Technology Systems and Christophe
11
00:01:08,240 --> 00:01:12,635
Ebbecke, Venture Manager, both of Bayer AG.
12
00:01:12,955 --> 00:01:14,715
Welcome to the podcast, gentlemen.
13
00:01:14,715 --> 00:01:15,595
Hi, Bob.
14
00:01:15,595 --> 00:01:16,155
Nice to be
15
00:01:16,155 --> 00:01:16,715
here.
16
00:01:17,515 --> 00:01:17,995
Yeah.
17
00:01:17,995 --> 00:01:18,875
Thank you very much.
18
00:01:18,875 --> 00:01:21,115
Appreciate the chance, to be here today.
19
00:01:21,355 --> 00:01:31,310
Well, it's great to have you on, and let's start off by giving our audience an overview of this facility, the project, the why and
20
00:01:31,310 --> 00:01:32,030
the what.
21
00:01:32,190 --> 00:01:40,110
Bayer SOLIDA-1, and you can see there isn't much fantasy in the name because SOLIDA comes from the SOLIDA production.
22
00:01:40,825 --> 00:01:50,585
It is a new launch and supply facility for all nonhormonal SOLIDA products Bayer will have in the next
23
00:01:50,905 --> 00:01:52,585
ten, twenty, thirty years.
24
00:01:53,510 --> 00:02:03,110
And basically, it, over the next years will, in this role, replace existing facility, which is
25
00:02:03,110 --> 00:02:05,750
from the beginning of the eighties, got old.
26
00:02:06,225 --> 00:02:16,065
And so a few years ago, the decision was taken instead of trying to, make the old facility new, we go for a new facility and takes the
27
00:02:16,065 --> 00:02:25,910
opportunity, yeah, to, build a new facility and to also take the opportunity to think about what
28
00:02:25,910 --> 00:02:28,710
might be the right facility for the future.
29
00:02:29,910 --> 00:02:31,670
Why is it in Leverkusen?
30
00:02:31,670 --> 00:02:41,585
It is because we have development and production on the same location, and that gives you and we will have talk about this probably later that gives you
31
00:02:41,585 --> 00:02:46,065
a big advantage in the timeline you need to bring product to the market.
32
00:02:46,225 --> 00:02:46,465
Okay.
33
00:02:48,280 --> 00:02:56,920
The SOLIDA-1 received recognition in both operational excellence and Pharma 4.0.
34
00:02:57,560 --> 00:03:03,720
What makes this facility unique, and what do you believe most impressed the judges?
35
00:03:04,175 --> 00:03:09,135
On the second part of the question, to be honest, I can only speculate.
36
00:03:10,415 --> 00:03:16,095
But what makes SOLIDA-1 unique is not, I think, one individual technology.
37
00:03:16,175 --> 00:03:23,980
It is probably that SOLIDA in pharmaceutical production were completely reengineered throughout the project.
38
00:03:25,180 --> 00:03:28,140
Nothing was just copied to go the easy way.
39
00:03:28,380 --> 00:03:33,855
Everything was challenged and discussed, like, do you have vertical flow or horizontal flow?
40
00:03:33,855 --> 00:03:37,135
Is a recipe in the machine or in the MES system?
41
00:03:37,455 --> 00:03:40,815
Do we need to improve equipment for better cleanability?
42
00:03:41,375 --> 00:03:44,015
How is the maintenance concept in the future?
43
00:03:44,175 --> 00:03:48,810
Where do we need what availability of equipment or even redundancy?
44
00:03:48,970 --> 00:03:50,410
How to automate?
45
00:03:50,410 --> 00:03:52,810
How to get data out of the machine?
46
00:03:53,130 --> 00:04:02,985
So there's a lot of things where we, basically very often did Pioneer work if we could not find something, could
47
00:04:02,985 --> 00:04:05,065
not find something that is really good.
48
00:04:05,625 --> 00:04:14,825
Basically, good existing solutions were optimized, but a lot of new solutions were developed, resulting in consequent design principles, modularity,
49
00:04:16,830 --> 00:04:26,590
ballroom concept, the full automation concept, full access to data and information sounds easy, but if you really want to do it, it gets horrible and
50
00:04:26,590 --> 00:04:27,150
complicated.
51
00:04:28,085 --> 00:04:33,685
And all that as a major condition for future proof, factory.
52
00:04:34,165 --> 00:04:43,770
Did this, analysis, this sort of requirements definition and concept design effort, Was that longer
53
00:04:43,770 --> 00:04:46,090
than a normal project?
54
00:04:46,490 --> 00:04:51,130
And do you think you got the benefits of that extra effort?
55
00:04:51,370 --> 00:04:51,850
Yeah.
56
00:04:51,850 --> 00:05:00,985
We, what we did is, before we started with the investment project itself, upfront, we took our time to make
57
00:05:00,985 --> 00:05:06,905
feasibility studies to investigate the machines of the different suppliers.
58
00:05:06,905 --> 00:05:08,985
We did really test runs with material.
59
00:05:09,650 --> 00:05:19,650
For example, we looked into all the continuous production equipment that was at that time available, checked it, and did test runs,
60
00:05:19,650 --> 00:05:22,610
and figured out what is advantage and disadvantage.
61
00:05:23,650 --> 00:05:32,065
Because when you build a new facility that is maybe a once in a decade's opportunity to change sits.
62
00:05:32,305 --> 00:05:32,785
Yeah?
63
00:05:32,785 --> 00:05:37,265
So, I just said that Leverkusen was built in the beginning of the eighties.
64
00:05:37,720 --> 00:05:47,240
We are now in 2026, forty years down the road at least, which means that maybe the next SOLIDA-1 comes in forty years
65
00:05:47,240 --> 00:05:47,880
again.
66
00:05:48,280 --> 00:05:54,445
And if we would only copy something, we would have in forty years something that is nearly a 100 years old.
67
00:05:54,685 --> 00:05:55,245
Yeah?
68
00:05:55,245 --> 00:05:58,605
So and I think that is what really the major driver.
69
00:05:58,605 --> 00:06:06,765
And we got, especially the order from the management not to just copy but rethink SOLIDA production.
70
00:06:07,690 --> 00:06:09,450
Well, that's really fantastic.
71
00:06:10,330 --> 00:06:17,290
Many pharmaceutical manufacturers struggle to balance speed, flexibility, and quality.
72
00:06:17,770 --> 00:06:27,335
What challenge was Bayer trying to solve with SOLIDA-1, and how did that vision shape the design of the facility?
73
00:06:27,655 --> 00:06:28,775
Great question.
74
00:06:29,335 --> 00:06:36,720
And the conventional wisdom is that you have to choose between speed, flexibility, and cost.
75
00:06:36,720 --> 00:06:44,800
And most pharmaceutical manufacturing plants that we see today follow that logic, and they're optimized for a specific purpose.
76
00:06:45,905 --> 00:06:50,865
A fixed product, a specific technology, or a range of capacity.
77
00:06:51,745 --> 00:07:01,660
With SOLIDA-1, we have decided to take a different approach, believing that, technology allows us
78
00:07:01,660 --> 00:07:07,580
to reach all these schools without being forced to accept the trade off.
79
00:07:09,660 --> 00:07:14,620
And so we engineered SOLIDA-1 for three core goals.
80
00:07:15,935 --> 00:07:25,295
First, bring drugs to patients as fast as possible, meaning we needed rapid product launches and flexible, scalable
81
00:07:25,295 --> 00:07:26,175
supply.
82
00:07:27,375 --> 00:07:37,270
We wanted to deliver the highest quality and also achieve competitive cost, and not only for one specific product, but
83
00:07:37,270 --> 00:07:42,710
as a manufacturing platform for our evolving internal and external pipeline.
84
00:07:43,830 --> 00:07:53,725
And to move beyond this traditional trade off and achieve the goal of you know basically achieving all three, we took two
85
00:07:53,725 --> 00:07:54,605
main levers.
86
00:07:54,605 --> 00:08:04,240
Number one, multilevel modularity, which allows SOLIDA-1 to evolve in line with pipeline and
87
00:08:04,240 --> 00:08:05,600
market requirements.
88
00:08:06,640 --> 00:08:13,600
And we can adapt SOLIDA-1 for different products, technologies, and also capacity requirements.
89
00:08:15,065 --> 00:08:24,105
And number two, second lever, by utilizing the potential of automation and digital integration in order to,
90
00:08:24,745 --> 00:08:33,860
manage the complexity of a modular, multiproduct facility and still achieve a high level of efficiency.
91
00:08:35,060 --> 00:08:44,745
And the result was honestly not your standard production facility, but it was rather a manufacturing platform which
92
00:08:44,745 --> 00:08:51,945
will and is constantly evolving and will probably look very different in ten years compared to today.
93
00:08:52,105 --> 00:08:58,345
I know a lot of projects get very ambitious around automation and digital integration.
94
00:08:59,190 --> 00:09:03,750
And in some cases it really brings the project to its knees.
95
00:09:04,630 --> 00:09:13,430
How did you overcome the complexity of automation and digital integration in order to achieve
96
00:09:13,430 --> 00:09:17,225
your speed to project delivery?
97
00:09:17,545 --> 00:09:27,545
Bob, to be honest, before we started SOLIDA-1 and we did all these feasibility studies, we did not only do a feasibility study when it comes
98
00:09:27,545 --> 00:09:31,740
to how does a machine look like or how does a factory look like.
99
00:09:31,980 --> 00:09:41,820
We also interviewed about, 12 engineering companies, and we visited, I would say, at least 10 sites, new construction sites
100
00:09:41,820 --> 00:09:43,500
of, big peer companies.
101
00:09:44,325 --> 00:09:54,005
And our target was not only to see what do they do technically, but also how do you set up a project that you do not die in the
102
00:09:54,005 --> 00:09:54,725
project.
103
00:09:55,285 --> 00:09:55,845
Yeah.
104
00:09:56,005 --> 00:10:05,520
And because experience was that it is more common that big projects have big problems and not the other way
105
00:10:05,520 --> 00:10:06,160
around.
106
00:10:06,320 --> 00:10:16,135
For example, what we figured out is to have a water damage in the last months of the project before CNQ
107
00:10:16,215 --> 00:10:17,975
seems to be the normal case.
108
00:10:18,535 --> 00:10:18,855
Yeah?
109
00:10:18,855 --> 00:10:20,135
It happened to me.
110
00:10:20,375 --> 00:10:20,855
Yeah?
111
00:10:20,855 --> 00:10:26,615
Some idiot leaves some valve open, and the water goes into your clean room walls.
112
00:10:26,870 --> 00:10:27,430
Yeah?
113
00:10:27,590 --> 00:10:28,950
It seems to be normal.
114
00:10:28,950 --> 00:10:29,430
Yeah?
115
00:10:29,590 --> 00:10:37,510
And that was something, for example, where we thought about if this happens to everyone, you need to think about it.
116
00:10:37,510 --> 00:10:45,285
And that, for example, was the reason that we try to have no water lines on top of the cleanroom walls.
117
00:10:45,765 --> 00:10:46,325
Yeah?
118
00:10:46,725 --> 00:10:56,090
And there design things with which you can increase the probability of success of a
119
00:10:56,090 --> 00:10:59,850
project from the design, but also from the organization.
120
00:11:00,170 --> 00:11:00,650
Yeah?
121
00:11:01,050 --> 00:11:10,895
And so that was just an example where we saw how others work we also learned what should we do different in
122
00:11:10,895 --> 00:11:15,455
order to increase the probability of success.
123
00:11:15,535 --> 00:11:17,375
We'll never say it on a project.
124
00:11:17,375 --> 00:11:18,015
That's sure.
125
00:11:18,015 --> 00:11:19,775
Because things happen.
126
00:11:19,775 --> 00:11:20,255
Yeah?
127
00:11:20,495 --> 00:11:22,575
But he can increase the probability.
128
00:11:22,920 --> 00:11:28,680
And the other thing is, probably later on we talk about how you organize your project.
129
00:11:29,080 --> 00:11:32,920
Also, there we did go a little bit different approach.
130
00:11:32,920 --> 00:11:42,615
For example, we invited the peer companies at the beginning of FEL two, the partner companies for the construction.
131
00:11:43,015 --> 00:11:50,295
And we used already the execution companies starting FEL two to help us to plan the facility.
132
00:11:50,970 --> 00:11:51,530
Yeah?
133
00:11:51,530 --> 00:11:56,810
Normally, do a planning, then you go through a bidding process, and then you find your partners.
134
00:11:56,810 --> 00:11:57,050
Yeah?
135
00:11:57,050 --> 00:12:03,370
We did this at least months earlier when, if not a year earlier.
136
00:12:03,450 --> 00:12:03,610
Yeah?
137
00:12:04,275 --> 00:12:05,875
And so that was something.
138
00:12:05,875 --> 00:12:09,715
So there were a lot of lessons learned out of this investigation.
139
00:12:10,355 --> 00:12:14,115
What can increase your probability, also in automation.
140
00:12:14,755 --> 00:12:23,440
And, of course, sometimes it's good you don't know how complicated something is because maybe otherwise you would not have started it.
141
00:12:23,440 --> 00:12:24,000
Yeah?
142
00:12:24,240 --> 00:12:30,080
And so we did not face the problems we expected, but, of course, we faced other problems.
143
00:12:30,355 --> 00:12:30,995
Yeah?
144
00:12:31,475 --> 00:12:37,475
But I I think, of course, you can go the easy way.
145
00:12:37,475 --> 00:12:46,850
But as I said, if you build a new facility maybe every twenty, thirty, forty years, I think whatever problems you face in a
146
00:12:46,850 --> 00:12:52,290
project, not to change something is probably the biggest fault.
147
00:12:52,770 --> 00:12:54,050
That's, yeah, that's great.
148
00:12:54,050 --> 00:12:57,375
And maybe I can add one one, aspect.
149
00:12:57,375 --> 00:13:02,335
I think it's a healthy level of, pragmatism in the project team as well.
150
00:13:02,975 --> 00:13:08,335
That, yes, as Jorgen said, you don't shy away from the difficult, topics.
151
00:13:08,335 --> 00:13:09,535
You know, try out new things.
152
00:13:10,590 --> 00:13:16,590
But, I guess sometimes it's important to not get stuck in the details and try to force solutions.
153
00:13:17,070 --> 00:13:26,465
But, go for pragmatic solutions, try to apply technologies that are proven, bring them them to your problem, and, achieve something maybe
154
00:13:26,465 --> 00:13:33,585
that's good enough that gets you there, but that allows you to move forward anyways.
155
00:13:33,585 --> 00:13:38,225
And I think that's something that the team definitely brought to the table from the beginning.
156
00:13:38,625 --> 00:13:39,265
That's great.
157
00:13:40,050 --> 00:13:47,170
You describe SLETA-one as a modular and adaptable manufacturing platform.
158
00:13:47,810 --> 00:13:55,915
Can you explain what that means in practice and why modularity is becoming so important for the future of pharmaceutical manufacturing.
159
00:13:55,915 --> 00:14:05,915
Well, we designed SOLIDA-1 from the ground up to be able to constantly evolve in line with pipeline, product, and market
160
00:14:05,915 --> 00:14:06,475
demands.
161
00:14:07,520 --> 00:14:14,960
Because in recent years, it has become increasingly more volatile and difficult to predict these factors.
162
00:14:15,680 --> 00:14:25,455
And what this required is a holistic modularity strategy, which in SOLIDA-1 extends from the total facility level down to the
163
00:14:25,455 --> 00:14:29,535
individual process unit on the manufacturing level.
164
00:14:30,415 --> 00:14:40,070
The facility itself is designed as a series of independent modules, which each serve a specific function, production, logistics,
165
00:14:40,070 --> 00:14:41,030
utilities.
166
00:14:41,670 --> 00:14:50,310
And the modules are connected by a central spine that serves as the backbone for material and, personnel flow.
167
00:14:50,895 --> 00:15:00,015
The spine is extendable in both directions, which allows us to add modules stepwise as they are needed, which
168
00:15:00,015 --> 00:15:06,360
significantly minimizes the upfront investment risk, for the company.
169
00:15:07,240 --> 00:15:17,000
Going down to the modules themselves, they are also designed with a standardized repeatable layout, which concentrates all support functions and
170
00:15:17,000 --> 00:15:26,585
utilities at the front of the module, and creates an open production area in the back area of the module facing away from the spine
171
00:15:27,065 --> 00:15:31,945
that can be used to design, process, and manufacturing environment freely.
172
00:15:32,540 --> 00:15:42,140
And this creates a standardized blueprint that, you know, we can reuse to copy and paste, expand new modules,
173
00:15:43,100 --> 00:15:48,985
which helps us to be significantly faster and also more cost efficient.
174
00:15:51,145 --> 00:16:00,985
A new module can be commercially ready in roughly two years, including qualification, and that's about a full year quicker than, at a
175
00:16:00,985 --> 00:16:02,905
traditional, facility.
176
00:16:03,680 --> 00:16:09,120
And it also allows us about 25% in cost savings compared to traditional plants.
177
00:16:09,680 --> 00:16:14,480
And then moving down to the, you know, manufacturing, the process level.
178
00:16:15,760 --> 00:16:25,145
Within the production area, we are creating modularity by using standalone process units, that are connected
179
00:16:25,145 --> 00:16:35,110
flexibly using robots and AGVs, and that allows us to adapt and change manufacturing processes
180
00:16:35,990 --> 00:16:38,630
without having to reconfigure the site.
181
00:16:38,950 --> 00:16:48,575
And it also allows us to replace single pieces of equipment and, manufacturing equipment without really adapting
182
00:16:48,575 --> 00:16:51,375
the building or the module itself.
183
00:16:52,095 --> 00:17:01,055
The facility is designed to reduce time to market by up to a year while supporting everything from clinical production to commercial scale manufacturing.
184
00:17:01,890 --> 00:17:08,050
How does SOLIDA-1 enable that faster path from development to patient access?
185
00:17:08,370 --> 00:17:12,370
Yeah, SOLIDA-1, there's no magic behind this.
186
00:17:14,985 --> 00:17:16,185
That's the truth.
187
00:17:16,425 --> 00:17:26,265
SOLIDA-1 accelerates the path from development to the patient by ensuring, continuous transition, especially on the interface from clinical
188
00:17:26,265 --> 00:17:29,305
development to, commercial production.
189
00:17:29,940 --> 00:17:39,700
And that starts, for example, with the colocation of clinical supply and commercial manufacturing in one location where you, for example, do not have to repeat
190
00:17:39,700 --> 00:17:44,115
stability studies if you do both at the same location.
191
00:17:44,755 --> 00:17:45,075
Yeah?
192
00:17:45,075 --> 00:17:47,795
And that saves a lot of time and, of course, money.
193
00:17:48,275 --> 00:17:58,100
Furthermore, the equipment in the different scale, yeah, so in the commercial scale or in the clinical supply, we have developed a concept that they
194
00:17:58,100 --> 00:17:59,540
match to each other.
195
00:18:01,540 --> 00:18:08,820
So from the equipment type, so that, the tech transfer, you can reduce the risk of problems there.
196
00:18:09,275 --> 00:18:13,675
And bad surprises you may have if the equipment does not fit to each other.
197
00:18:13,835 --> 00:18:14,395
Yeah?
198
00:18:14,795 --> 00:18:24,670
And, again, the colocation not only pays off with the regulatory authorities, but also, of course, since the information needs to flow
199
00:18:24,670 --> 00:18:33,870
over a few meters from one colleague to the other in those directions, from production to clinical supply or the other way around,
200
00:18:34,605 --> 00:18:44,285
The probability that they talk about something and they understand what is going on in the other plant is
201
00:18:44,285 --> 00:18:47,005
very high because we talk about meters.
202
00:18:47,005 --> 00:18:51,590
And they probably meet during, they even meet for the next lunch anyhow.
203
00:18:51,990 --> 00:19:01,985
So I think that is something you can hard measure, but information flow between people who meet during lunch is, of course, better than people who might be hundreds
204
00:19:01,985 --> 00:19:09,505
of thousands miles away from each other, yeah, where theoretically the information flows, but in reality it does not work.
205
00:19:10,385 --> 00:19:20,100
And the high level of data availability starting from the first batch, that of course supports scale up modeling,
206
00:19:20,100 --> 00:19:22,820
process optimization and tech transfer.
207
00:19:23,220 --> 00:19:33,155
For example, during the tech transfer of the first product in SOLIDA-1, the modeling of the process parameters in the fluid bed granulator started already
208
00:19:33,155 --> 00:19:40,675
during the last minutes of the first batch when the material not even has left the fluid bed granulator.
209
00:19:40,915 --> 00:19:41,475
Yeah?
210
00:19:41,475 --> 00:19:51,060
Because the data available and in the same moment somebody was sitting in Berlin and starting his modeling machine with the data he got from the machine.
211
00:19:51,460 --> 00:19:52,180
Yeah.
212
00:19:52,180 --> 00:20:02,155
So it's, I think, close proximity and regulatory advantages together with, very
213
00:20:02,155 --> 00:20:05,035
uncomplicated information flow.
214
00:20:05,035 --> 00:20:10,315
That are the main levers, which bring the product, faster through the patient.
215
00:20:10,395 --> 00:20:20,130
Did I just, hear that maybe there was digital twins, either of the equipment, the process, or perhaps both in play here?
216
00:20:20,290 --> 00:20:20,690
Yeah.
217
00:20:20,690 --> 00:20:27,330
The word, digital twin is something when you ask a lot of people, you get different answers.
218
00:20:27,330 --> 00:20:28,530
What is a digital twin?
219
00:20:29,445 --> 00:20:39,365
When it comes to the single machine, of course, you have models of, for example, especially the fluid bed granulator of how does a fluid
220
00:20:39,365 --> 00:20:40,725
bed granulator work.
221
00:20:40,725 --> 00:20:43,445
That is a digital twin of the machine itself.
222
00:20:44,220 --> 00:20:53,900
But also during the project, we, use something you can call a digital twin, where, for example, we had a
223
00:20:53,900 --> 00:20:56,860
three-d model of the facility.
224
00:20:57,665 --> 00:21:01,345
But not only for the sake of having a three-d model.
225
00:21:01,665 --> 00:21:06,865
And then we, transferred this three-d model into gaming software.
226
00:21:07,905 --> 00:21:17,800
So you could run through, you could really run through SOLIDA-1 Play Operator, And that we used, during the design phase and when
227
00:21:17,800 --> 00:21:27,640
the facility was not finished, to, optimize, the construction and the planning, together with the
228
00:21:27,640 --> 00:21:28,200
operators.
229
00:21:29,075 --> 00:21:29,475
Yeah.
230
00:21:29,475 --> 00:21:32,355
Because it's a different thing when you look on the layout.
231
00:21:32,515 --> 00:21:39,155
A lot of people, when you show them the layout, it's hard for them to see what does it really mean once it's built.
232
00:21:39,395 --> 00:21:45,410
But, it helped a lot, to, yeah, to play operator in your process.
233
00:21:45,970 --> 00:21:52,850
And then you figure out, there is a little door missing for the samples or whatever.
234
00:21:53,410 --> 00:21:53,970
Yeah.
235
00:21:53,970 --> 00:21:56,370
So it was very much used.
236
00:21:56,370 --> 00:22:02,945
And so also during the project, digitalization played a big role.
237
00:22:03,345 --> 00:22:11,745
We tested more things than we at the end used because we also had a lot of things which we tried and we figured out.
238
00:22:12,280 --> 00:22:15,400
We are probably a few years too early for that.
239
00:22:15,640 --> 00:22:25,240
For example, we used virtual reality or augmented reality glasses because one idea was,
240
00:22:25,685 --> 00:22:35,205
what if an operator has no monitor and only the glasses on his on his nose and, runs through the batch record with gesture?
241
00:22:35,765 --> 00:22:36,085
Yeah?
242
00:22:38,070 --> 00:22:39,030
Did not work.
243
00:22:39,030 --> 00:22:40,870
The connection was not sufficient.
244
00:22:40,870 --> 00:22:45,190
And they are simply too heavy to have some eight hours on your nose.
245
00:22:45,190 --> 00:22:45,590
Yeah?
246
00:22:45,590 --> 00:22:47,510
So that is something we tried.
247
00:22:47,590 --> 00:22:52,150
And I can give you at least twenty twenty examples where we tried new technologies.
248
00:22:52,865 --> 00:22:55,185
And finally we came to the result.
249
00:22:55,265 --> 00:23:00,385
It may work maybe in five years, but today it does not work.
250
00:23:00,705 --> 00:23:01,105
Yeah.
251
00:23:01,105 --> 00:23:08,305
The industry talks a lot about Farmer four point o, but Zelyta demonstrates it in action.
252
00:23:09,040 --> 00:23:18,640
And what are some of the technologies or combinations of technologies that have been most transformative in creating an autonomous, connected
253
00:23:18,640 --> 00:23:20,160
manufacturing environment?
254
00:23:20,720 --> 00:23:26,135
Well, I would say that, the most transformative element is probably not one technology.
255
00:23:26,135 --> 00:23:28,855
I'd put it like you just said it.
256
00:23:29,415 --> 00:23:38,935
It's probably the effective combination of technologies, devices, and software systems into one connected ecosystem.
257
00:23:40,670 --> 00:23:50,030
Because a fully automized pharmaceutical production facility is only possible if you manage to integrate ERP, MES
258
00:23:50,030 --> 00:23:57,215
systems, production equipment, robotics, and logistics equipment into one ecosystem.
259
00:23:58,175 --> 00:24:07,690
And creating one functional ecosystem really is not that easy, and, it required a couple novel solutions that,
260
00:24:08,010 --> 00:24:13,530
we noticed are not available or were not available, on the market anywhere.
261
00:24:13,530 --> 00:24:20,410
And one example which stands out is the production flow control system, the PFC system.
262
00:24:22,115 --> 00:24:32,115
The PFC system was custom developed by the SOLIDA-1 team, and it serves as the central orchestration layer for
263
00:24:32,595 --> 00:24:34,435
process driven material flow.
264
00:24:35,400 --> 00:24:42,920
What it does is it tracks continuously the real time location of all logistics devices.
265
00:24:43,560 --> 00:24:50,760
It generates transport orders and communicates with the AGV fleet, the automated guided vehicles.
266
00:24:51,475 --> 00:25:01,235
And it thus creates the interface between the material flow on the shop floor in real life and the manufacturing execution ecosystem, the
267
00:25:01,235 --> 00:25:02,915
MES system on the factory level.
268
00:25:04,360 --> 00:25:13,160
And this interface so far typically relied on an operator, you know, that was responsible for oversight,
269
00:25:13,640 --> 00:25:14,840
prioritization.
270
00:25:15,480 --> 00:25:25,395
And what we achieved is that the PFC provides this level of intelligence that previously was provided by an operator and allows us
271
00:25:25,395 --> 00:25:27,395
to reach a new level of automation.
272
00:25:27,395 --> 00:25:36,780
And I think that's what Pharma four point o in practice is, is it's integrating equipment, software, tools, and data into
273
00:25:36,780 --> 00:25:42,220
one connected and, I'm gonna say, increasingly intelligent operating system.
274
00:25:42,700 --> 00:25:51,255
So following along with that, one particular interesting point is that automation isn't replacing people necessarily.
275
00:25:51,495 --> 00:25:52,935
It's changing how they work.
276
00:25:53,655 --> 00:26:03,360
So how have employee roles evolved, within this facility, and what new skills are becoming most important in
277
00:26:03,360 --> 00:26:06,160
such a highly automated facility?
278
00:26:06,560 --> 00:26:13,680
Yeah, as you said, a SOLIDA-1 is highly automated, but it is for sure not a people-free facility.
279
00:26:15,040 --> 00:26:25,005
The role of people moves in general away from repetitive and physical demanding activities, what we have in current, in, let's
280
00:26:25,005 --> 00:26:34,950
say, more older facilities right now, towards, more the role of supervising the process, root cause analysis, and troubleshooting, and
281
00:26:34,950 --> 00:26:36,390
process optimization.
282
00:26:37,030 --> 00:26:46,550
In the first half of career, I was always working in operations and spent a lot of years as a plant manager in different plants.
283
00:26:47,015 --> 00:26:56,775
And usually, I needed 90% of if I had a problem, I needed 90% of the available time to get some data, and then I had a few minutes left
284
00:26:56,775 --> 00:26:59,655
to figure out what do the data mean.
285
00:26:59,815 --> 00:27:00,055
Yeah?
286
00:27:00,590 --> 00:27:06,750
And that, I think, is now completely the other way around, at least.
287
00:27:06,830 --> 00:27:07,390
Yeah?
288
00:27:07,390 --> 00:27:14,910
You get the data, and you can spend your time in what does it mean, where is the really problem, and how to solve it.
289
00:27:14,910 --> 00:27:15,150
Yeah?
290
00:27:15,315 --> 00:27:25,075
And that was, in the past, very exhausting to get data because sometimes you need to sit down and simply, with a piece of paper, write things down from batch, from
291
00:27:25,075 --> 00:27:27,235
batch, from batch, and each batch.
292
00:27:27,235 --> 00:27:27,315
Yeah?
293
00:27:29,040 --> 00:27:35,360
The skill set of people is moving towards a better understanding of the overall process.
294
00:27:35,440 --> 00:27:35,680
Yeah?
295
00:27:35,680 --> 00:27:39,600
You need people who more understand what is the overall process.
296
00:27:39,600 --> 00:27:44,975
Not somebody who's only an expert in a tabletting machine is running a bigger machine.
297
00:27:44,975 --> 00:27:45,375
Yeah?
298
00:27:45,375 --> 00:27:52,255
The understanding of the way of working of machines and systems and the meaning and interpretation of technical data.
299
00:27:52,895 --> 00:28:02,500
And automation and data, of course, make inefficiency and process problems much more transparent in typical manually factories and
300
00:28:02,500 --> 00:28:08,340
develop more push to solve problems instead of developing creative workarounds.
301
00:28:08,340 --> 00:28:08,820
Yeah?
302
00:28:08,980 --> 00:28:18,555
So, I think, to be honest, I think very often people try find a way to work around the problem instead of solving
303
00:28:18,555 --> 00:28:19,035
it.
304
00:28:19,035 --> 00:28:19,515
Yeah?
305
00:28:19,915 --> 00:28:29,250
Automation does not remove expertise from manufacturing, but it allows people more, especially in operation, to
306
00:28:29,250 --> 00:28:38,610
better develop and use the expertise, really do the things that a person should do and not do things which do not create value.
307
00:28:39,535 --> 00:28:49,375
But it also means that during the design and implementation of a facility, you need to be more precise in the process understanding
308
00:28:49,615 --> 00:28:50,655
and planning.
309
00:28:50,975 --> 00:28:51,215
Yeah.
310
00:28:51,215 --> 00:29:00,040
In the past, maybe you you have a room, you put a tablet press in the middle, and you say the operator will figure out how to run it.
311
00:29:00,200 --> 00:29:00,680
Yeah?
312
00:29:00,840 --> 00:29:04,440
But if you automate it, you need to make up your mind.
313
00:29:04,440 --> 00:29:05,960
Where is the pallet coming in?
314
00:29:06,195 --> 00:29:09,075
How to lift it to the machine?
315
00:29:09,075 --> 00:29:09,635
Yeah?
316
00:29:09,635 --> 00:29:11,555
How to get the machine running?
317
00:29:11,555 --> 00:29:13,075
Who starts the recipe?
318
00:29:13,075 --> 00:29:22,830
So you you need basically, I would say, you need you are more forced to do what you should do in other facilities also, more precise and good
319
00:29:22,830 --> 00:29:28,430
planning, and not, puts a problem later on the shoulders of the operator.
320
00:29:30,110 --> 00:29:30,590
Yeah.
321
00:29:30,590 --> 00:29:40,285
So this facility introduces a ballroom manufacturing concept for oral solid dosage production, which allows robots, your automated
322
00:29:40,285 --> 00:29:45,085
guided vehicles, and materials to flow freely throughout the facility.
323
00:29:45,725 --> 00:29:52,130
What advantages has this approach delivered compared to traditional manufacturing layouts?
324
00:29:52,770 --> 00:29:53,170
Yeah.
325
00:29:53,170 --> 00:30:01,090
So traditional layouts plays every process step typically in a separate room, which, creates an interface
326
00:30:02,715 --> 00:30:05,355
interrupting the manufacturing process flow.
327
00:30:05,675 --> 00:30:14,955
And, what we decided is to put nearly all major process units into one open multifunctional space,
328
00:30:15,700 --> 00:30:17,780
which we call the ballroom.
329
00:30:18,340 --> 00:30:27,940
And, this ballroom gives us a completely new level of production flexibility as it enables free movement
330
00:30:28,165 --> 00:30:37,205
of robots and AGVs between these process units that are not interrupted by doors, material locks, or other pieces of equipment.
331
00:30:37,765 --> 00:30:47,750
And what this does is it allows us to change the production process very flexibly and very quickly without requiring any
332
00:30:47,750 --> 00:30:48,950
reconfiguration.
333
00:30:49,670 --> 00:30:55,670
And it also means that, equipment access is much easier for maintenance, for example.
334
00:30:56,235 --> 00:31:05,995
And looking into future requirements, it gives us a a much higher level of adaptability, in terms of replacing or
335
00:31:05,995 --> 00:31:07,115
upgrading equipment.
336
00:31:09,480 --> 00:31:18,360
Zooming out a little bit, what it does on the facility level is that, using the ballroom concept enables a roughly
337
00:31:18,760 --> 00:31:26,485
25% reduction in building footprint because we're eliminating, duplicate logistics and traffic space.
338
00:31:27,365 --> 00:31:34,725
And, what I wanna point out, the the ballroom doesn't limit productivity or efficiency.
339
00:31:35,690 --> 00:31:42,890
Because what we do is, we produce up to five simultaneous batches of the same product within the ballroom.
340
00:31:44,330 --> 00:31:49,930
So it's nothing that would bring our, our efficiency down.
341
00:31:51,665 --> 00:31:53,905
It basically gives us flexibility.
342
00:31:53,905 --> 00:32:03,720
And through using technology, automation tools, we manage to keep the efficiency, at a good level
343
00:32:03,720 --> 00:32:06,440
for for us from a manufacturing standpoint.
344
00:32:06,440 --> 00:32:12,760
So you've described SLETA-one as a major step towards lights out manufacturing.
345
00:32:12,920 --> 00:32:21,945
How close is the industry to truly autonomous pharmaceutical production, and what key capabilities still need to mature?
346
00:32:22,905 --> 00:32:29,305
I would say ZOLIDA1 demonstrates that a high level of automation is already possible in pharmaceutical production.
347
00:32:29,650 --> 00:32:37,970
And if you look in other industries, like car industry, SOLIDA-1 is maybe even more than normal than the exception,
348
00:32:39,330 --> 00:32:39,890
yeah?
349
00:32:41,170 --> 00:32:49,775
Major activities like material handling, weighing, the different formulation production steps, mostly cleaning, not completely.
350
00:32:50,095 --> 00:32:54,415
Documentation are automated and digitally connected.
351
00:32:54,895 --> 00:33:04,230
However, the next step from my perspective would be to use the data and artificial intelligence to let the factory optimize
352
00:33:04,230 --> 00:33:05,110
itself.
353
00:33:05,750 --> 00:33:06,310
Yeah?
354
00:33:06,870 --> 00:33:10,950
And we had this in mind from the beginning on when we started.
355
00:33:10,950 --> 00:33:14,045
Artificial intelligence was freak stuff.
356
00:33:14,045 --> 00:33:17,645
We heard about it, but we were not sure what comes out later on.
357
00:33:17,725 --> 00:33:18,285
Yeah?
358
00:33:19,885 --> 00:33:29,360
But so I could now say it was all planned, but I would in reality, there's more good luck that we builded a
359
00:33:29,360 --> 00:33:33,840
facility with this data backbone, which you need to use AI.
360
00:33:33,920 --> 00:33:34,480
Yeah?
361
00:33:35,120 --> 00:33:45,115
By luck, we we did it exactly in the right moment because AI just in the last two years came to birth really and is now more and more
362
00:33:45,115 --> 00:33:45,915
used.
363
00:33:46,955 --> 00:33:56,510
I would say if he goes this way that you say one day, yeah, it it does take care of himself, The biggest hurdle probably
364
00:33:56,670 --> 00:34:04,910
is in the heads of the people to accept that process parameters may be changed one day without the approval of a person.
365
00:34:05,230 --> 00:34:15,185
And to be honest, we already could do this, but I think that is something which the
366
00:34:15,265 --> 00:34:16,705
same like with the ballroom.
367
00:34:16,705 --> 00:34:21,905
We we asked authorities to come in before we and showed them the concept.
368
00:34:21,905 --> 00:34:24,330
Because the question is, you can have a great idea.
369
00:34:24,330 --> 00:34:27,370
If it does not get accepted, you have a problem.
370
00:34:27,450 --> 00:34:27,930
Yeah?
371
00:34:27,930 --> 00:34:37,595
And I I think also that is something where we need to maybe to give the people some time to accept it, that process parameters
372
00:34:37,595 --> 00:34:42,795
get optimized by artificial intelligence.
373
00:34:43,915 --> 00:34:53,610
Furthermore, what I think is a deeper understanding of the physical properties of the materials will make operation more predictive and foster automation.
374
00:34:53,770 --> 00:34:57,690
I think their pharmaceutical industry has some way to go.
375
00:34:58,170 --> 00:35:08,125
And finally, it will be very important when you go for automated production to develop processes with the later automation already in mind.
376
00:35:08,445 --> 00:35:18,270
For example, you should avoid manually sampling and IPCs and use much more automated online analytics Because it
377
00:35:18,270 --> 00:35:19,310
does not make sense.
378
00:35:19,310 --> 00:35:28,990
You you save the operator, runs the machine, but then somebody comes with a sample glass and takes a sample, goes back to the IPC lab, and then the machine needs to wait for the result.
379
00:35:29,310 --> 00:35:39,115
You know, I spoke with an executive in the data center AI, industry, and, he said that, really, you know, AI is
380
00:35:39,115 --> 00:35:47,870
based on sophisticated statistics, which is what pharmaceutical manufacturing, process control,
381
00:35:49,070 --> 00:35:58,910
identifying the process parameters is really based on, and that we should be using, the the AI
382
00:35:59,445 --> 00:36:06,005
for more statistical applications and less so for the sort of ChatGPT kind of stuff.
383
00:36:06,725 --> 00:36:16,070
But looking looking ahead, AI enabled decision making, digital twins, predictive maintenance,
384
00:36:16,630 --> 00:36:21,270
and real time quality monitoring have been noted as the next phase of Pharma four point o.
385
00:36:22,310 --> 00:36:26,870
Which of these innovations do you believe will have the greatest impact over the next five years?
386
00:36:27,695 --> 00:36:37,535
So I'd say that the the the real transformation will probably come from a combination of of these technologies, and a
387
00:36:37,535 --> 00:36:45,470
combination of these technologies around, trusted, contextualized, and basically high quality manufacturing and process data.
388
00:36:47,710 --> 00:36:55,310
Because in order to leverage the potential of data, I would honestly even put these technologies in somewhat of an order.
389
00:36:55,935 --> 00:36:56,335
Right?
390
00:36:56,335 --> 00:37:03,615
I mean, real time monitoring enables a new level of transparency of the state of the manufacturing process.
391
00:37:04,655 --> 00:37:12,160
This transparency, again, enables the application of, for example, digital twins to simulate what might happen.
392
00:37:12,720 --> 00:37:20,400
And, again, as the next step, AI helps to identify patterns with the goal to create predictive tools.
393
00:37:21,485 --> 00:37:28,765
And I'd say, therefore, these technologies kind of mesh and work together to, create real life value.
394
00:37:29,405 --> 00:37:39,210
And what we did at SOLIDA-1, we made sure to get the foundation for all of this right, meaning we're collecting,
395
00:37:39,370 --> 00:37:48,810
I'm gonna say, near real time data holistically throughout the entire manufacturing process, and we're merging it into one integrated database.
396
00:37:49,755 --> 00:37:59,115
And this allows us to continuously integrate, further develop, and combine technologies to drive the progression that we currently see
397
00:37:59,115 --> 00:38:09,040
from, I would say, connected operations towards intelligent, predictive, and still well governed manufacturing.
398
00:38:10,000 --> 00:38:19,200
And maybe allow me to allow me to say one thing, regarding the, you know, ChatGPT, chatbot, comment you made earlier, and I completely agree with you.
399
00:38:19,625 --> 00:38:24,265
The statistical analysis is probably where the real value lies.
400
00:38:24,665 --> 00:38:34,460
But we see one example at SOLIDA-1 that goes a little bit in the direction of the chatbot functionality, which is that now that we
401
00:38:34,460 --> 00:38:44,395
collect all the data, we can give operators the, application and the chance to use GenAI tools to kind
402
00:38:44,395 --> 00:38:53,435
of search in the data and make it easier to access the information that they need, which doesn't replace the analytics and, you know, it doesn't help all the way.
403
00:38:53,915 --> 00:38:58,715
But I'm sure you've done it before if you try to find information in a large document.
404
00:38:58,715 --> 00:39:02,920
It's it's a lot easier to just ask ChatGPT, give me a little update.
405
00:39:02,920 --> 00:39:05,000
That's what we do at SOLIDA-1 as well.
406
00:39:06,200 --> 00:39:16,115
Helping the operators, you know, look in batch records, try to find information on on deviations, equipment status, and, you know, help them
407
00:39:16,115 --> 00:39:18,115
get the information as fast as possible.
408
00:39:18,435 --> 00:39:20,035
Great, great.
409
00:39:20,195 --> 00:39:29,570
So for organizations considering their own digital transformation journey, what are the biggest lessons learned from this project that they can apply to
410
00:39:29,570 --> 00:39:35,810
today regardless of the size or maturity of their manufacturing operations?
411
00:39:36,450 --> 00:39:37,010
Yeah.
412
00:39:37,650 --> 00:39:47,185
Let's start with a simple truth - Pharma four point o or digitalization or however you call it is, not just a technology.
413
00:39:47,745 --> 00:39:52,385
It is an operational transformation that changes the way how we work.
414
00:39:52,740 --> 00:40:02,740
Because things it's it's not just an aids that you, instead of typing yourself or writing on a piece of paper, suddenly you have
415
00:40:02,740 --> 00:40:06,515
a computer and you print something, but basically do the same.
416
00:40:06,675 --> 00:40:07,875
Here it is different.
417
00:40:07,955 --> 00:40:10,035
It changes the way how you work.
418
00:40:10,835 --> 00:40:12,595
You need to better understand.
419
00:40:12,595 --> 00:40:22,420
In order to do so, you need to better understand your operational processes in detail and the purpose of them not hesitating to change them
420
00:40:22,420 --> 00:40:23,460
for the better.
421
00:40:23,940 --> 00:40:29,780
That is not only valid for the project team, but as well for the operations team later on.
422
00:40:30,660 --> 00:40:40,505
Second, any new facility, new production line or even piece of equipment if you have smaller changes because you do not everybody builds a
423
00:40:40,505 --> 00:40:42,985
new facility every second day.
424
00:40:43,625 --> 00:40:47,865
All of that is the opportunity to change the way of working.
425
00:40:48,400 --> 00:40:52,480
I would say never just copy old stuff.
426
00:40:52,640 --> 00:40:57,120
Take the opportunity to think about if you may change something.
427
00:40:57,680 --> 00:41:02,720
Third, in any project, start your digitalization subproject on day one.
428
00:41:03,545 --> 00:41:09,705
In the the history, very often the focus on investment project was on steel and concrete.
429
00:41:10,105 --> 00:41:19,950
The digital factory is at least as important as the steel and concrete factory, even if you don't see
430
00:41:19,950 --> 00:41:22,670
that much and if the CapEx is not so high.
431
00:41:22,910 --> 00:41:27,630
But in order that it works, it is very important.
432
00:41:28,110 --> 00:41:33,345
We identified and involved the execution companies, for example, already.
433
00:41:34,225 --> 00:41:34,785
Yeah.
434
00:41:34,945 --> 00:41:41,825
Fourth point, if you have a complex project, involves the partner companies in an early phase, as I already said.
435
00:41:41,825 --> 00:41:41,985
Yeah?
436
00:41:42,400 --> 00:41:48,000
But not only them, also the users, the QA, etcetera, early on.
437
00:41:48,000 --> 00:41:49,040
Very important.
438
00:41:49,600 --> 00:41:58,585
For example, we identified and involved the execution companies already in the very early phase of the project starting FEO two.
439
00:41:58,585 --> 00:42:04,025
Was additional work because you do something usually you do for the bidding process.
440
00:42:04,025 --> 00:42:08,185
We did this in early phase, but it paid off a lot.
441
00:42:09,225 --> 00:42:17,400
Topic number five, develop design principles, standards, and stick to them without being not flexible to improve them.
442
00:42:17,480 --> 00:42:22,760
That reduces workload and helps you to focus on the really complicated stuff.
443
00:42:23,415 --> 00:42:29,095
Because in the digital factory, like I said, you need to be more precise in planning.
444
00:42:29,095 --> 00:42:31,255
You need to be more professional.
445
00:42:31,575 --> 00:42:41,140
So try to get get the easy stuff away, work with standards so you focus with your people on the really important things.
446
00:42:41,700 --> 00:42:47,620
The majority of the money you spend on is on brain and not on the square meter cost of concrete.
447
00:42:48,260 --> 00:42:56,055
So avoid tight custom made designs, which create a lot of complexity later on in the engineering.
448
00:42:56,535 --> 00:43:01,095
Six, and that comes to a word we used already, modularity.
449
00:43:01,095 --> 00:43:04,695
Use modularity and clear interfaces between the modules.
450
00:43:05,320 --> 00:43:11,800
And that's not because, for the sake of modularity, that helps you to reduce complexity.
451
00:43:11,800 --> 00:43:17,160
You have clear interfaces, and each Lego block has a role and plays its role.
452
00:43:17,320 --> 00:43:22,635
And if something changes, this block changes, but, yeah, your whole building changes.
453
00:43:22,715 --> 00:43:24,635
It makes the thing much easier.
454
00:43:24,955 --> 00:43:31,595
And that does not only work for building architecture, but that is something where I can see it easily with the modules and so on.
455
00:43:31,930 --> 00:43:41,690
That also works for equipment setup and IT systems, yeah, in which, for example, we try to separate
456
00:43:41,690 --> 00:43:50,465
the data from the software because your software may change in the next years, but you want to keep your data.
457
00:43:50,545 --> 00:43:51,025
Yeah?
458
00:43:51,025 --> 00:43:57,905
So a factory today is more like an iPhone, where you can change the software.
459
00:43:58,465 --> 00:44:07,060
And even if in five years your program to show your photos is different, you still have your data for the photos.
460
00:44:07,300 --> 00:44:07,860
Yeah?
461
00:44:08,020 --> 00:44:17,815
And that is something flexibility in the design helps you a lot when you keep things separated, for example, not mixing
462
00:44:17,815 --> 00:44:19,095
data with software.
463
00:44:19,255 --> 00:44:22,135
Well, this has been a great conversation.
464
00:44:23,735 --> 00:44:26,375
This project is really cutting edge.
465
00:44:26,900 --> 00:44:36,100
The innovation of the overall process design, the open ballroom, the use of integrated, digital,
466
00:44:36,580 --> 00:44:46,355
technologies, how you're changing how people, what their roles are, and basically upskilling, what they
467
00:44:46,355 --> 00:44:47,315
focus on.
468
00:44:47,875 --> 00:44:49,875
This is all super fantastic.
469
00:44:49,875 --> 00:44:54,355
And so I'd like to thank Juergen and Christophe for their time with us today.
470
00:44:55,075 --> 00:44:59,940
It is great to hear more about Bayer's SOLIDA-1 facility.
471
00:45:00,900 --> 00:45:07,620
That brings us to the end of another episode of the ISPE podcast, shaping the future of pharma.
472
00:45:08,580 --> 00:45:17,415
Please be sure to subscribe so that you don't miss future conversations with the innovators, experts, and changemakers driving our industry forward.
473
00:45:18,535 --> 00:45:28,270
On behalf of all of us at ISPE, thank you for listening, and we'll see you next time as we continue to explore the ideas, trends, and
474
00:45:28,270 --> 00:45:30,910
people shaping the future of pharma.
