This document discusses how artificial intelligence (AI) is being harnessed by the pharmaceutical industry. It provides examples of how AI is enhancing drug development, clinical trials, customer engagement through chatbots, and compliance monitoring for patients. While AI shows great potential, the document also cautions that cybersecurity and managing expectations are important challenges to consider when implementing AI technologies in healthcare.

1. Artificial Intelligence:
Friend
or
in Innovation and Digitalwww.eyeforpharma.com
Disruptive
influence
How technology is changing
the doctor-patient relationship
Digital dinosaur
Pharma enters a brave new
digital world

2. 2
Contents
8Disruptive influence
How technology is enhancing the
doctor-patient relationship
4Artificial intelligence:
Friend or foe?
AI is transforming how
pharma engages stakeholders
11An unwelcome guest
Helpful tips to engage HCPs online
14Digital dinosaur
Can pharma shake off its prehistoric
image when it comes to digital?

3. 3
www..eyeforpharma.com
“Success consists of stumbling from failure to
failure with no loss of enthusiasm” – a phrase
apparently coined by Winston Churchill. If indeed
he did, this was a remarkably good vision of the
age of the machine. The phrase neatly sums up why
an artificial intelligence can in many ways become
superior to a human
Humans tend to give up, computers do not. Combine
that with the ability to learn from each failure and you
have a very progressive mindset indeed.
Healthcare – an information-rich industry that affects
every one of us and is increasingly being streamlined
by electronic patient records and digitized medical
literature – provides a very fertile ground for artificial
insights. You can see why tech giants like IBM and
Google’s DeepMind have chosen healthcare as their first
target. You can see why chatbots and natural language
processing are expected to have their first applications
in helping patients. You can also see why pharma has
taken a keen interest in multiple areas, from research to
customer education.
This issue of Trends Magazine begins by exploring
how we can find the positive – rather than the feared
negatives – in AI, a technology that, by definition, we
do not fully control. This is followed by three pieces
exploring the illustrious possibilities that data and
technology can afford us – as long as we maintain
perspective and open-mindedness. Some definite
lessons for the future!
Paul Simms
Chairman
eyeforpharma
Welcome

4. Artificial intelligence:
How pharma is harnessing
the power of AI
t is safe to say that popular fiction has not been too kind
to artificial intelligence. From Mary Shelley’s time-honored
parable about a creator destroyed by his creation through
countless sci-fi depictions (think Skynet’s killing machines). AI
has become a metaphor for mankind’s self-destructive power.
Yet, while these depictions capture our imagination, do they
match reality?
The Oxford Dictionary defines artificial intelligence as: “The theory and
development of computer systems able to perform tasks normally requiring
human intelligence, such as visual perception, speech recognition, decision-
making, and translation between languages.”
AI has arrived in pharma, says Chris Keenan, Head of Worldwide Medical
Customer Engagement, Medical Capabilities at Bristol-Myers Squibb, who
points to the manifold possibilities within healthcare.
“We have AI imaging tools that can see what the human eye cannot; even
the best doctors in the world are starting to augment their practice by
leveraging these imaging tools to look beyond what they could see before.
For example, you can predict cancer earlier if you can process larger amounts
of patient data,” he says.
4
Friend
or

5. 5
www.eyeforpharma.com
“Where you would typically
have a pharmacist or a
highly skilled individual
within a customer service
department, you now have
chatbots.”
AI is also impacting the development of medicines,
offering the ability to accelerate clinical trials. “It truly
is about the patients,” says Keenan. “The quicker we
can bring safe products to market, the more options
physicians will have to treat their patients.”
Chatbots that utilize speech recognition and virtual
assistance are a revolutionary step in the customer
engagement space, he says. “Where you would typically
have a pharmacist or a highly skilled individual in a
customer service department, you now have chatbots,
where people can engage a pharmaceutical company
with an FAQ type of question. And remember the old
days when your chat would only be available from eight
to five? Why? Because there’s a human. Today, we can
send information 24/7, 365 days a year.”

6. 6
AI’s unique ability to process and make sense of vast data sets
is strengthening touchpoints right across the spectrum, says
Keenan. “Cognitive reasoning affords us the ability to look at
massive amounts of data – real-world data, social data, internal
data – and truly understand what the computer talking about. If
you are an oncologist and neuroscience is new to you, we have
the ability, or we will have the ability, to decipher that data,
so when we have a conversation face to face, we are better
prepared with the data that person is looking for. It affords us
greater flexibility, responsiveness and speed,” he says.
Keenan has high hopes for the future of speech recognition
to “disseminate information in an appropriate, balanced, safe
and accurate way” to HCPs, through further developing the FAQ
model. Natural language processing is the next evolutionary step,
using advanced algorithms to respond to the complex nuances of
human conversation.
“AI will automate
administrative
activities so
doctors can focus
on practicing
medicine.”
“It will empower the healthcare provider; getting data quicker,
more accurately and with greater speed will help them become
better at what they are doing,” he says. “Doctors throughout
the world are getting burnt out. Why? It is not solely because
they are seeing too many patients, it is because the amount of
information to be absorbed from electronic medical records is
like drinking from a firehose. The data they have to capture is an
administrative task and its killing them. AI will automate some of
these administrative activities so doctors can focus on practicing
medicine with their patients.
“The more that we can give them an effortless customer
experience, the better they can perform for patients and the
better we can serve them and ultimately patients, he says.
The emergence of virtual systems is also playing a pivotal role
in supporting one of medical’s core functions, says Keenan. “We
want to make sure our products are used safely and appropriately.
Many of our products are based on compliance. Now, virtual
assistance can be available 24/7, instead of keeping a call center
open for 24 hours a day, which is challenging to do these days,
especially with high-skilled talent.
“These virtual systems can help compliance with meds. Diabetes
is a prime example, where people have to manage diet, they
have to manage adherence; virtual systems help you do that
and that’s AI engagement, not engaging a person.”
Keenan may be an AI advocate, but he is also cognizant
of the potential risks. He warns against diving head
first into an emerging technology without a clear and
comprehensive business objective. “If you do not have a
practical understanding of the problem you are trying to ‘solve’,
it is doomed to fail,” he says.

7. www.eyeforpharma.com
Whereas, “if you have a clear business objective of how you
are trying to improve the customer experience, delivering value
for your organization by automating routine administrative tasks,
so that you can invest more into RD than administrative work,
there are real business problems to be solved.”
Keenan has a clear goal he is working towards. “I want to be
able to deliver medical information to HCPs 24/7, 365 days a
week, making sure that it is relevant and balanced. With machine
cognitive reasoning, the technology is available for basic FAQs,
but it’s not yet there to engage in deeper dialogue from a scientific
perspective. But I guarantee it will be.”
This feeds into another area of concern; managing AI
expectations. “With these tools are emerging technologies, do
not think that they are going to solve all your problems, they
are not. You have to be realistic because much of the industry’s
overpromising the capabilities of what many of these tools can
achieve.”
There is also the serious issue of cyber security to consider,
says Keenan. While the threat is not exclusive to AI, digitizing
more and more of our business operations increases the chance
of sensitive information being compromised.
“There is nothing more personal than electronic medical
records. It is one thing to steal our social security number, it is
another thing to steal our medical records, who I am as a person
and what health challenges I am confronted with.”
So, looking forwards, will AI supersede everything else as the
main driver of innovation? “It is not going to replace one-to-one
relationships, it is just another step forward in how we engage
patients and the healthcare community,” says Keenan.
AI’s “autochannels” such as chatbots simply supplement
the current multichannel experience. “You will still have chat
functions, the phone and face-to-face engagement that will
always be prominent for us in pharma,. Doctors want to talk and
have peer-to-peer conversations.”
Chris Keenan
Head of Worldwide Medical
Customer Engagement, Medical
Capabilities
Bristol-Myers Squibb
Meet our
contributor
“It is one thing to steal our social
security number, it is another thing to
steal our electronic medical records.”

8. n his book The Tipping Point, Malcolm Gladwell writes
about that magic moment when “an idea, trend, or social
behavior crosses a threshold, tips, and spreads like wildfire”.
Those in the know say that the use of technology in
medicine is currently at that tipping point, and that the
practice of medicine will fundamentally change over
coming decades.
Healthcare is currently “low resolution”, says Buchan, who
is also Professor of Public Health Informatics at The University
of Manchester. “We know a lot about patients when things go
wrong, when they come into the clinic and details are taken, we
get a snapshot in time. At higher resolution, the view extends
across the patient journey, tapping into the rhythms of life. A very
important rhythm is medication behavior, about which very little
is truly known.”
In a more connected and predictive world, we will look
at therapeutics within the context of biology, behaviors and
environments. “We are on the verge of a really interesting journey
for clinical therapeutics and patient experience, which will allow
us to have high-definition healthcare. This will treat the patient
journey and not just the stops along the way.”
For example, stratification will become more accurate, with
Technology is about to fundamentally change the
doctor-patient relationship, says Microsoft Director of
Healthcare Research, Iain Buchan
8

9. www.eyeforpharma.com
advanced biological knowledge leading to
more precise care, when coupled with
technological advances, such as AI and
personalization.
With such pervasive technology,
transparency will come to the fore.
“If I go to my GP or look online for
treatment options today, it is hard to
see how certain those options are. It may
be that current evidence can’t determine
whether treatment A or B is best for me. In
this situation, the most transparent answer
from the doctor (or the algorithm consulted
before making the appointment), is, ‘Don’t know:
Would you like to enter a study to discover the
best treatment in this situation, helping patients
like you?’ The very nature of the doctor-patient
consultation needs to change, and the right AI could
underpin this evolution,” says Buchan.
9
“The very nature of the doctor-
patient consultation needs to
change, and the right AI could
underpin this evolution.”

10. Such connectivity will also expose the “invisible” individual
pharmacology of each patient, including the timing of medicines,
and the “meta-medication” of lifestyle factors such as diet,
alcohol, physical activity and sleep patterns.
“This will lead to a more purposeful connection between
physician and patient. We are moving towards a world where
personal health data and algorithms provide a view of the road
ahead for your health, whereas old-style medical records were
rear-view mirrors of healthcare,” he says
“If patients donate data to be learnt from, healthcare systems
have a duty to harness those data for better therapeutic
understanding and development. Separate systems for post-
marketing surveillance, clinical trials and prescribing quality
management, etc. are wasteful – they need to merge into ‘bigger
brains’ at the core of health systems.”
Such algorithms will become an inherent part of any new
therapy, meaning the success of a drug will not simply be down
to the molecule, he asserts. “It is wasteful to support medication
Iain Buchan
Director of Healthcare Research
Microsoft UK
adherence in diabetes, for example, without using the same
patient contacts to influence diet and physical activity.”
These algorithm-drug combinations could potentially show
up different subgroups of people who need a different type
of therapy, so the boundaries between drug development and
discovery may blur.
Unsurprisingly, this type of intelligent medicine means the
physician’s role will change irrevocably, with a collaborative
effort needed across tech, pharma and medicine. “As patients get
more insights into what is happening in the body, feedback loops
are created that can lead to better outcomes. However, this kind
of disruptive observation will also pose challenges in trial design
for researchers and regulators.”
Physicians and patients will interact differently, and across all
channels, says Buchan. “Trust is central; trust not only in preserving
patient privacy but also in making fullest use of personal data for
the healthcare choices that can and should be computed.”
Meet our
contributor
10
“Trust is central; trust not only
in preserving patient privacy
but also in making fullest use
of personal data.”

11. 11
www.eyeforpharma.com
Four reasons why pharma struggles to engage
HCPs online (and how to change it)
harma needs to find new ways to promote its medicines and engage
physicians, says Bozidar Jovicevic, VP, Global Head Digital Medicines,
Sanofi, who sees four key issues plaguing pharma.

12. 12
The Reasons
Holding HCP attention
“The number one commercial channel for decades has been sales reps, visiting
physicians and influencing them to prescribe drugs. That channel is slowly
dying – for two reasons. Firstly, doctors are either closing their doors to sales
reps or limiting the visit to two minutes, so you have sales reps driving around
with a price per visit of US$300-500, maybe making five sales calls per day.
That’s ten minutes with physicians each day.”
“Secondly, we know that the doctors are increasingly spending time online
to stay up to date. In Europe, 63% of physicians spend more than six hours
online each week. They are spending time online and closing their doors to
pharma offline.”
Uninspired content
Pharma is “notoriously bad” at engaging physician’s online in a meaningful and
sustained way, says Jovicevic. “In the online world, it is easy for physicians to
stop engaging at any point if it is not meaningful or relevant to them. So, if
pharma is not relevant to them in the first 10 seconds, all they need to do is
close the tab/window. It takes one click.
“For physicians, they face an overwhelming amount of available content.
They find it difficult, if not impossible, to stay up to date. This is one issue
where pharma can be helpful.”
Inexperience
Another problem is pharma’s lack of expertise of engaging physicians in a
multichannel world, he says. “In pharma, there is a very small number of solid
digital marketeers, people who understand online engagement, analytics,
traffic, conversions, psychology, etc. There is also a lack of digital leaders who
have skills and can lead the digital transformation.”
Outdated processes
Pharma’s lack of digital leadership is further compounded by a largely legacy-
driven approach to product development in IT. “Most IT platforms need to be
common (one platform customized for each country but still one platform)
and cloud-based. Pharma needs to adopt agile processes and leave the
waterfall behind.
“On top of this, the top 20 countries in a big pharma organization may use
five different CRM systems, so it is very difficult to harmonize and create
a global, digital multichannel strategy. Add that to the many non-common
online platforms that each country is building its own apps, developing its own
engagement platforms for physicians,” says Jovicevic.

13. www.eyeforpharma.com
The Solutions
Bozidar Jovicevic
VP, Global Head Digital Medicines
Sanofi
Meet our
contributor
1. Build rapport
”Pharma has to shift its online approach from an “unwelcome
pest to a welcome guest” by providing value to physicians and
building good will, says Jovicevic.
“This could be providing education – it means talking about
something not necessarily to do with their drugs, but this
will then build up goodwill and form a relationship and get
permission to talk about their products, what I call ‘permission-
based marketing’.
3. Centralize digital content
If people need an example of a company that has centralized its
digital content offerings, they should consider “a little site called
Facebook”, says Jovicevic. “When you log into Facebook everything is
there, not separate apps for each country – it is all accessible with the
one password and it adapts to your own needs.”
A centralized legal and compliance team devoted to digital issues
should also be established.
4. Aggregate what’s already
out there
With content, it is about finding the “white space”, says Jovicevic, so
pharma companies should analyze the online content and platforms
that physicians are already using to find opportunities.
One clear opportunity lies in aggregating content – collecting
and curating vast swaths of online content to make it easier for the
individual physicians to digest.
Jovicevic knows about aggregating content – he was central
in pioneering the first global eCongress platform for medical
e-conferences while at Novartis. This gathered the best content from
the huge number of medical conferences held each year and hosted
it on a platform called VivindaTV, which he says looks like “the Netflix
of medical conferences”.
It included the latest breaking abstracts, videos in real-time, chat
forums (in development), all personalized to the doctor’s interests
and/specialty. “This was of huge value to physicians and is a good
example of finding the white space (confirmed by NPS score of 71
from physicians using the platform).”
2. Rethink innovation
Another key element is not to “over-engineer” the innovation.
For Jovicevic, in large pharma companies, or any heavily
regulated industry, innovation “breaks” at middle management,
and also in legal and compliance. A long process of de-risking as
the innovative project moves through a large corporation means
the resulting success might be so minor that it is ultimately
inconsequential.
Organizations should create innovation teams or groups, with
a relatively high level of autonomy. They should be separate
from the rest of the organization – he even recommends a separate
building. Executive sponsorship is key – without this, the innovation
will continue to fall foul of mid-management overthinking.
“Large companies are designed to execute proven business models
– they know their stuff. They know how to sell, how to produce and
how to talk to customers. Innovation works very differently – you are
searching for a business model, you are testing. Companies need to
build a prototype, test, tweak, and try again.”

14. 14
e have all heard it said that pharma is a digital dinosaur,
with nimble Silicon Valley start-ups whizzing ahead to
explore exciting new frontiers, leaving pharma to lumber along
in their wake, barely breaking a sweat.
While Sanofi’s Chief Data Officer, Milind Kamkolkar, agrees that there’s rarely
smoke without fire, pharma may not be as prehistoric as some observers think.
“I’ve done some consultant work with other ‘modern industries’ and we beat
ourselves up too much. The big issue for all industries is that, today, business is
technology, and tech companies are eating our business.”
Tech companies have the edge over pharma because digital is built into the
base of their business models and woven into the fabric of their operations, he
says. In pharma, that job falls to the CDO. “Right now, we treat the CDO as an
evangelist and glorified cheerleader, but the larger role is around industrializing
digital as an operating model.”
Can pharma shake off its
prehistoric image when it
comes to digital?

15. 15
www.eyeforpharma.com
To truly embed digital in the business, companies need
to better manage their investments, he says, with the
drive for that coming from the top, he says. “The CEO
must be committed to industrializing digital; it can’t be an
afterthought. It can’t be, ‘Let’s hire a CDO and let them do all
this wonderful work’. That person can facilitate and lead the
change, but you need to build the organization around it.”
Digital centers of excellence can provide support but
they are often given “a small pot of money from which they
have to beg, borrow and steal from different brand teams,
franchises or groups,” says Kamkolkar – an approach that
“doesn’t work if you want to industrialize digital”.
SHIFTING GEARS
There are two paradigm shifts in pharma that are helping
to legitimize digital operations as a revenue generator, he
says. The first is beyond the pill initiatives. “The shift is from
brand into therapeutic area support, care pathways and truly
understanding the science – in understanding why certain
effects happen across different patient populations and how
they manifest themselves into customer experiences with
their physician or healthcare systems. Digital can facilitate
this level of engagement.”
The second shift is that pharma is claiming a stake in the
healthcare apparatus, resulting in a “horizontalization of the
industry”, says Kamkolkar. “We used to think of healthcare as
a vertical industry but that is no longer the case. There are
numerous industries claiming a healthcare position, engaging
with patients, physicians or regulators. They are not coming
from within the classical boundaries, and they are forcing us
to rethink, in a very positive way, our current business model.”
What about the multichannel customer experience space?
“Has anyone really cracked the shell on that? I am not entirely
sure. There are some unique use cases – let’s call them niche
winds – that help cement belief, but what gets in the way
is the need to understand the levers for a digital operating
model in order to make it relevant for your customer base in
a particular market. I am not sure that is fully understood.”
Marketing teams could also better reflect the current
multichannel climate, he says. “Most members of our
marketing teams today were once reps, but reps are just one
of many channels now. Personal relationships still make a
difference but rep-led engagement is already eroding.”
“We treat the CDO
as [a] glorified
cheerleader, but the
larger role is around
industrializing
digital.”

16. 16
“The one asset
we don’t
treat with the
enough respect
and diligence is
information.”
Pharma also needs to improve how it manages data pools. “The
one asset we don’t treat with the enough respect and diligence
is information, even though it is a highly monetizable entity,” he
says. “We need to start putting greater financial discipline on
what we do with data. At Sanofi, we have created a framework
called ‘Fair’, humbly borrowing from the RD world of ‘fair’ data
standards – findable, accessible, interoperable and reusable. The
number one issues for most of our business users, internal or
external, is that we simply don’t know if we have the information
or not, so the ability to find information is the biggest issue.”
AUTOMATE TO INNOVATE
With the analysis of data, is the future automated? “Yes and
no. You still need scientific context,” says Kamkolkar. “One of
the things we are implementing is patient ontology based on
knowledge graph (Google’s advanced search engine capability)
and machine learning capabilities because the way medical
people speak about a disease is not the way patients speak about
it. For example, a patient might say, ‘Oh man, my head is about to
explode,’ and it’s not a physical explosion, rather a headache. But
how do you capture that effectively?”
Automation will move us towards seamless synchronicity, he
predicts. “We spend countless hours defining
‘master-data’ entities; when it
comes to the customer, for
example, one country might
say ‘HCP’ while another
says ‘physician’, and in the
olden days of master-data
management, you had to code
that in as two separate entities
then create a joining table that
allowed you to understand it all.
Today, you can use knowledge graph
capabilities to remove the barriers –
or politics- of dictionary management.
This is where I see automation having a massive role to play.”
Another area where sophisticated digital tools can help
eliminate language barriers is with literature reviews. “Many of
us produce publications and have a very different ontology set
under which that publication or disease recommendation was
made. Even if it comes from industry standards, by the time it gets
implemented in companies, it’s heavily configured. As a result,
when you go back and do a literature search – the definition of a
gene in one company may not necessarily be the same in another.
The beauty of knowledge graph technology is that it can see how
different companies define it.”

17. www.eyeforpharma.com
“Apps are a “grossly
unexplored area.”
Milind Kamkolkar
Chief Data Officer
Sanofi
Meet our
contributor
17
“WE’VE GOT AN APP FOR THAT”
Looking forwards, where does Kamkolkar think pharma should be
heading in the digital sphere?
Apps are a “grossly unexplored area”, he says, particularly
because the apps that have gained the most traction have been
in therapeutic areas, which has proven problematic. “Companies
have often been too afraid to accept data on other medications
from other companies. Part of it has to do with the regulatory
binding that we are in today, and the risk-reward mechanism also
plays into that.”
Another issue that’s impeded the development of apps is that
the “first iteration of apps weren’t developed with a commercial
model in mind. The easiest way to check is to see when the app
was last updated – if there wasn’t a product mindset behind the
app, it was a complete waste of money.”
Key questions need to be asked, he adds. “Do you have a
product management group that is looking at the app as an entity
that contributes to PL? Is it something that is going to improve
either the customer experience or the patient journey?”
Failing to ask these questions during the development process
– regardless of the initial buzz an app generates at launch – may
see the app falter afterwards. So, how do you sustain momentum?
“The brand manager may have changed and the app is floating
around in the app store with no one saying, ‘Hang on a second, do
users really want to use it this way? What do we need to change
to make it a more compelling experience?’ That rarely happens,”
laments Kamkolkar.
An app’s shelf-life is often linked to employee incentives, he
adds. “Often, these apps are created to fulfill an objective and
advance a person along a career path. So, the objective is met
and they’ve got their bonus, but then what? CDOs need to work
closely with HR to create a plan that incentives people to persist
in the world of digital health.”
Kamkolkar offers a promising yet pragmatic vision for the future
of digital. “I hope that 10 years from now, it is just part of business
and is a new revenue stream. That should be the goal. To achieve
that, you need people who really understand digital economics,
who know how to consolidate initial successes so there are not
just niche successes coming through.”
It is also high time for a culture change, Kamkolkar says,
rebuking the tech industry’s lack of diversity in general. Diversity
will drive better outcomes, he stresses.
“Digital and data, in many ways, are the operating levers of how
we work in the future,” Kamkolkar affirms. Yet, it’s not just down
to pharma companies. “Payers can’t do it alone, regulators can’t
do it alone and we certainly can’t do it alone.”