- Exploring
the Impacts
of Agentic AI on
the Relationship
with HCPs
Who We Are: Authorship & Team
The Authorship team is made up of multidisciplinary professionals with expertise in data, technology, business, and innovation. The entire Artef act team works collaboratively to deliver high-impact solutions, always a ligned with our clientsˇ needs.
Alexandra Mangeard
Managing Partner Brasil
Ilton Chaves
Data Consultant
Marina Rodrigues
Data Engineer
Ricardo Kim
Senior Data Consulting Manager
Larissa Costa
Data Engineer
Paolo Gozdzink
Marketing Specialist
Aline Martins
Analytics Engineer
Lucas Milhorato
Data Consultant
Lay d’Arc
Graphic Designer
About Autorship Artefact
Artefact accelerates the adoption of data and Artificial Intelligence to positively impact people and organizations. We offer a wide range of services, from strategy to operations, implementing AI solutions across industries to help companies capture the competitive advantage of data and AI transformation.
linkedin-in globe youtube
Summary
Summary
Introduction 1
New Scenario 3
The New Scenario in the Healthcare Sector
3
Digitization, data, and the new patient behavior 4
What is Agentic AI?
10
Technological advancement and potential for increased business value 11
Challenges and Solutions 15
The Four Critical Challenges in the Relationship with HCPs And How Agentic AI Can Overcome Them
16
Volumetry: How to reach and connect with all HCPs? 16
Quality: Are we delivering the right information to the right HCP? 17
Frequency: How to maintain ˝top of mind˛ a among so many competing interactions? 18
Trust: How to build a long-term relationship and real partnership?
19
Sales Rep Agentic Hub 20
Building the Sales Rep Agentic Hub The Future of Medical Representation is Autonomous, Multichannel, and Scientifically Relevant
21
From Sales Rep to Intelligent Ecosystem 22
The Scientific-Contextual Engagement Machine 24
Strategic Gains for the Industry 25
Recommended Technical Architecture 25
Summary
Real Application Cases 28
Real Application Cases How Companies Are Implementing Agentic AI in Practice
29
Use case 1: summary of HCP interactions 29
Use case 2: medical writing assistant 31
Use case 3: AI persona generator 33
Ethical & Compliance 35
Ethical and Compliance Considerations in Agentic AI for Healthcare 36
Privacy and Continuous Consent 36
Accountability: Who is responsible for automated decisions? 37
Bias and Representativeness 37
Transparency and Explainability 37
Recommended Governance Frameworks 38
Adoption Roadmap 39
How to Implement Agentic AI 40
Deep Dive: Individual PoV 42
Deep Dive: Use Cases with real impact 42
Deep Dive: Scale transformation through a transversal program 43
Artefact 44
Why Artefact is the Right Partner 45
Conclusion 49
Links & References 52
Glossary 54
Summary
1
Introduction Unraveling Agentic AI in the Context of Healthcare
The healthcare ecosystem is vast and complex, involving multiple actors such as patients, pharmaceutical companies, hospitals, health plan operators, regulators, and, of course, healthcare professionals HCPs. Among all these stakeholders, HCPs play a unique and strategic role: they are the trusted intermediaries between the industry and the patient. They recommend treatments, prescribe medications, monitor patient progress, and often directly influence treatment adherence.
For this reason, over the last few decades, the pharmaceutical and healthcare industry has been investing heavily in building and maintaining relationships with these professionals. According to data from Evaluate Pharma and McKinsey, globally around 30% of pharmaceutical companies’ commercial budgets are allocated to HCP engagement initiatives, representing over US$ 90 billion per year these initiatives include actions by medical representative teams, scientific congresses, in-person events, continuing medical education, and distribution of technical and scientific materials. In Brazil, large groups allocate between 15% and 25% of their marketing and relationship budgets to actions directly aimed at HCPs.
The return on these investments, although difficult to measure accurately, is of significant strategic importance: studies indicate that up to 60% of prescription and medical recommendation decisions can be directly influenced by interactions between companies and healthcare professionals. In other words, the impact of these relationships is profound yet still underutilized in terms of scale, personalization, and intelligence.
It is in this context that Agentic AI emerges as a game-changer. By combining LLMs (Large Language Models) with autonomous and interactive decision flows, it opens up the possibility of scaling relationships with HCPs in a more personalized, efficient, and data-driven way.
Introduction
2
It is not just about replacing in-person visits, but about reimagining the medical-industrial relationship as an intelligent digital ecosystem, where AI agents can:
Answer scientific questions in real-time, based on the latest publications and research;
Help HCPs develop through personalized learning paths, with varied content based on detected knowledge gaps or the demonstrated interests of each professional;
Anticipate the most relevant topics for each HCP, analyzing behavior patterns, preferences, and contextual data;
Assist in clinical decision-making, suggesting evidence and options based on similar cases;
Personalize content through dynamic persona generation, based on interactions and content consumption, enabling real-time micro- segmentations;
Collect systematic feedback naturally (including sentiment analysis), measuring the depth and effectiveness of each interaction to generate richer and more actionable insights for optimizing strategies;
Execute complex engagement journeys without overwhelming human teams (omnichannel orchestration).
Agentic AI is more than just a tool itˇs a catalyst for the healthcare industry. In
addition to optimizing the current ROI of investments with HCPs, Agentic AI offers a
unique opportunity for the healthcare sector to fundamentally redefine its role in the
patient journey. By empowering HCPs with predictive insights and personalized
support, we are building a smarter, more proactive, and patient-centered ecosystem
one where the relationship between the industry, the professional, and the patient
becomes more collaborative and impactful.
Alexandra Mangeard, Managing Partner Brasil
Introduction
3
Chapter 1
The New Scenario
in the Healthcare Sector
- T
4
Digitization, data, and the new patient behavior
The healthcare sector is experiencing a turning point. The convergence of digital transformation, changing patient behavior, and expanded access to data about HCPs (healthcare professionals) is creating a new competitive landscape where maintaining relevance requires more than good campaigns: it demands offering real, continuous, and personalized value.
The digitalization of interactions with healthcare professionals has created a highly fragmented ecosystem, but also one full of opportunities. The number of new digital services aimed at HCPs has skyrocketed in recent years, ranging from clinical platforms (such as Memed, Mevo, iClinic) to AI-based medical decision support systems. Some of these solutions add value and strengthen relationships with doctors, but many others contribute to a scenario of noise, dispersion, and information overload.
At a relatively low cost, it is possible to build extremely rich and dynamic profiles of these professionals. Sources such as LinkedIn (professional history), Memed (prescription patterns), Google Maps (patient feedback), medical portals (browsing behavior and scientific interest), social networks (networking structure), and clinic ERPs have begun to offer valuable inputs to personalize interactions, predict needs, and accurately measure engagement. What was once built based on inference can now be managed with concrete and actionable data.
New Scenario
5
At the same time, we now have access to an
unprecedented volume of data about HCPs.
At a relatively low cost, it is possible to build extremely rich and dynamic profiles of these professionals. Sources such as LinkedIn (professional history), Memed (prescription patterns), Google Maps (patient feedback), medical portals (browsing behavior and scientific interest), social networks (networking structure), and clinic ERPs have begun to offer valuable inputs to personalize interactions, predict needs, and accurately measure engagement. What was once built based on inference can now be managed with concrete and actionable data.
This ability to better understand HCPs
becomes even more critical in the face of
a structural change: doctors are losing some
of their influence over patient decisions.
The modern patient is digital, autonomous, and informed or at least believes to be. Before, the doctor’s word was indisputable. Today, patients resort to tools such as ChatGPT, YouTube, and Google before consulting a healthcare provider or purchasing medication. And this is not just an impression there is concrete data that demonstrates this change in behavior.
New Scenario
6
In the United States, a survey of 2,000 adults revealed that 52% have already used AI systems, such as ChatGPT, to describe their symptoms. Of these, 84% stated that the diagnosis provided was correct, reinforcing patients’ growing trust in AI-based technologies for initial healthcare.
In Australia, a study conducted in June 2024 with over 2,000 participants indicated that 9.9% used ChatGPT for health-related questions in the first half of the year alone. Additionally, 38.8% of those who have not yet used it are considering doing so in the next six months, which demonstrates the potential for expansion of this type of use.
When it comes to social networks, a recent global report pointed out that over 60% of people use these platforms to seek health information, with prominence for channels such as YouTube, WhatsApp, and Facebook. This search for content in alternative sources shows that the contemporary patient is more active, curious, and digitally empowered factors that challenge and simultaneously create opportunities for healthcare brands.
New Scenario
7
Global Virtual Healthcare Services Market Size, by Mode of Consultation, 2022 2032 (Billion USD)
Vídeo
Áudio Mensagens Quiosque
60
50
40
30
20
10
7.1
50.9
8.7
10.4
13.3
16.7
20.0
22.7
27.8
33.2
40.6
0
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
22.4% $50.9 B
The Market Will Grow at a CAGR of
Projected Market Size by 2032 in USD
New Scenario
8
Use of Social Media in Healthcare Marketing
Social Media Usage (March 2022)
Snapchat 13%
Tik-Tok 17%
Pinterest 32%
Vimeo 40%
Blogs 74%
LinkedIn 90%
Twitter 90%
Instagram 92%
YouTube 94%
Facebook 98%
0
20% 40% 60% 80% 100%
Media Usage in %
Source: Market.us Media
This transformation is accentuated by a new service delivery logic. Many HCPs have started to act as “personal brands” and have diversified their patient care approaches. Full check-ups, prevention programs, telemedicine consultations, and continuous digital support via apps or messages have become part of the new value offering of these professionals. And with the rising cost of health plans, patients of different profiles are beginning to seek alternative solutions including the free use of artificial intelligences for basic guidance. The HCP, therefore, needs to reinvent themselves: they cease to be merely a prescriber and also become a relationship manager with the patient.
New Scenario
9
In this new scenario, brands operating
in healthcare need to do more than
communicate: they need to facilitate.
Supporting the HCP with reliable scientific content, market data, technologies that integrate their care, and tools that sustain their reputation can be the path to establishing a new partnership relationship.
But this facilitation needs to go beyond personalization based solely on the profile:
it is essential to also consider the specific moment of each professional their career stage, their current clinical context, their daily priorities, and even their level of prior exposure to the brand. This dynamic approach allows delivering the right content, on the right channel, at the most relevant moment.
It is a moment of transition where scientific marketing, CRM, and artificial intelligence meet to offer smarter, more responsive, and truly physician-centric journeys in their real context of action.
New Scenario
10
Chapter 2
What is
Agentic AI?
11
Agents represent a recent technological advance, with great potential to increase value for companies
From RPA to Agentic AI, the business value of intelligent systems has increased tenfold through:
High adaptability to the environment
High learning capacity
High contextual interaction
Business Value
2000 2012 2022 2025
RPA
supremacy
Automation of back office operations with simple and static logic orchestration in monotonous tasks.
Traditional AI
democratization
Revolutionizing processes with unparalleled precision, harnessing immense amounts of data to perform tasks beyond human reach (Predictive maintenance, forecasting, optimization)
Generative AI
diffusion
Widely adopted for simplifying various tasks thanks to a variety of specialized assistants in everyday life, through information retrieval, document production, synthesis…
Agentic AI
ascension
Agentic AI mimics human reasoning and autonomously explores complex problems, executes actions and makes decisions.
What is Agentic AI?
12
As the healthcare ecosystem reinvents itself with new demands from professionals, patients, and pharmaceutical industries, technologies are also going through a decisive moment of transformation. In this context, Artificial Intelligence enters a new phase with the emergence of Agentic AI.
This approach combines:
Generative AI that interprets complex contexts and generates decisions in natural language.
Autonomous agents that execute tasks with minimal human intervention.
Real-time data that enables contextualized and adaptive decision-making.
Agentic AI systems combine RPA and Generative AI
for agents to perform tasks autonomously
Agents
Complex reasoning
LLM
Task orchestration Integration
with tools
Automation
Dynamic decision- making
For ad hoc tasks that require great adaptability
PERCEIVE
RACIOCINATE
To orchestrate manual tasks
reliably, with high adaptability
ACT
Based on deterministic
Pre-defined processes
For your routine tasks that require high reliability and efficiency.
Cognitive tasks
Complex tasks
Manual tasks
What is Agentic AI?
13
Furthermore, Agentic AI solutions differentiate themselves by integrating technologies such as RAG (Retrieval-Augmented Generation) that allow agents to consult external sources in real- time before making decisions, increasing accuracy; AI pipelines that can automate the complete flow of ingestion, processing, and decision, connecting data to action; and orchestration that coordinates various agents simultaneously, ensuring they operate synchronously and efficiently.
This combination allows systems to perceive, plan, and act, transforming analytical operations into active and continuous value generation, something previously very complex to achieve with approaches using traditional AI, for example.
Operational Capability Conventional AI Agentic AI
Execution Method Static / Reactive
Continuous / Proactive
Input Type Fixed and structured data
Dynamic prompt + real-time context
High (adapts decisions according to the environment)
Adaptability Low (ad hoc, with human intervention)
Limited to static training; no feedback during execution
Iteration and Learning Acts, observes, adjusts, replans
Dynamic, with RAG and APIs connected in real-time Data Integration Limited and pre-configured
In the current Healthcare scenario, where the complexity of interactions with HCPs continuously grows, the ability to transform processes becomes essential to ensure efficiency, agility, and accuracy.
The operational routine is often marked by reactive actions, system fragmentation, excessive manual tasks, and alignment challenges between areas. Identifying the right intervention points is the first step to simplify, scale, and improve the journey, engagement, and support for HCPs.
What is Agentic AI?
14
The ideal process should include Example of application with Agentic AI
1. Manual, chronological, and repetitive tasks
Automatic sending of scientific newsletters or event reminders to HCPs on predefined dates
Automated analysis of interactions with HCPs to generate personalized insights based on history and profile
2. Large volume of data to be processed
3. Multiple platforms without added value
Integration of CRM, email system, and medical repository into a single HCP service interface
4. Many stakeholders involved (e.g., medical, legal, compliance)
Automation of material approval flow before sending to HCP, with alerts and input from relevant stakeholders
5. Different areas involved (medical, regulatory, commercial)
Automatic alignment of scientific messages with promotional material, reducing rework and inconsistency
Organizations that have adopted GenAI-based solutions have already reported significant gains, such as a reduction in operational task execution time by up to 40% and a 30% increase in the efficiency of medical and commercial teams by minimizing rework, according to a study published by McKinsey & Company (2023).
In addition, AI-driven platforms have improved the personalization of interactions with healthcare professionals by up to 50% (Accenture Digital Health Tech Vision, 2023), promoting greater engagement and accuracy in scientific campaigns.
With the emergence of Agentic AI, these impacts are expected to be amplified, given its potential to orchestrate complex tasks autonomously and adaptively.
What is Agentic AI?
15
Chapter 3
Challenges & Opportunities with Agentic AI
- i
16
The Four Critical Challenges in the Relationship with HCPs And How Agentic AI Can Overcome Them
The relationship between the healthcare industry and healthcare professionals (HCPs) has never been simple. It demands scientific precision, human empathy, communication consistency, and adaptability. And despite digital innovations in recent years, many fundamental challenges remain relevant they have only changed form.
In this chapter, we will examine four of these structural challenges and how Agentic AI can offer disruptive and scalable solutions.
1 Volumetry
How to reach and connect with all HCPs?
Yesterday
The traditional model was centered on in-person visits by medical representatives. Although effective for certain profiles, this model was limited by logistics, high costs, and low scalability. This limitation is especially critical considering that 68% of doctors prefer online communication, with 80% using smartphones and 75% participating in exclusive networks for doctors [Doceree]. The traditional model simply cannot meet this digital demand.
Challenges and Solutions
17
Today
Digital marketing seeks to overcome this limitation with email campaigns, programmatic media, and webinars. But the content is still cold, generic, and often irrelevant to the HCP’s current moment. This generic approach ignores the vast and diverse universe of healthcare professionals in Brazil, estimated at almost 4 million by their respective federal councils, making it practically impossible to create messages that resonate with each individual in such a broad market.
Tomorrow
We can imagine a solution where each HCP interacts with a digital agent specifically trained for their clinical profile, specialty, and journey. This “Sales Rep Agentic Hub” would be available 24/7, with medical language, personalized and updated content, and could scale coverage intelligently, without losing the depth of human interaction.
2 Quality
Are we delivering the right information to the right HCP?
Yesterday
The focus was 100% on the product. Communication followed a “one size fits all” logic, centered on benefits and technical characteristics. This generic approach ignores the fact that 57% of doctors changed their perception of a drug based on information seen on social media, and 41% altered their prescribing habits as a result [Sermo and LiveWorld]. “One size fits all” communication misses the opportunity to positively influence clinical decisions.
Today
With CRM and marketing automation tools, we are beginning to move towards personalization, segmenting by specialty, digital behavior, and stated preferences.
Challenges and Solutions
18
Tomorrow
Real-time hyper-personalization will become the new standard. Using historical, contextual, and predictive data, intelligent agents will be able to deliver individualized Next Best Actions (NBAs), whether with scientific content, an invitation to a congress, or a clinical update. For example, if a doctor has recently attended a webinar on a new treatment for diabetes, the agent could send a relevant scientific article on the topic, an invitation to a specialized congress, or a clinical update on the latest advances all automated, yet relevant and reliable.
3 Frequency
How to maintain ˝top of mind˛ a among so many competing interactions?
Yesterday
Frequency was addressed through constant visits, regardless of effectiveness or need. This generated saturation and low marginal return. This intrusive and impersonalized approach contributes to the information overload that affects 62% of healthcare professionals, who feel overwhelmed by the volume of promotional content, and to the negative perception of “spam,” reported by 65% of doctors, according to research. Frequency alone does not guarantee engagement, and can even be counterproductive.
Today
Commercial teams attempt to optimize their routes with prioritization and segmentation tools, but frequency still depends on the physical capacity of the teams.
Tomorrow
Omnichannel orchestration with Sales Rep Agentic Hub will allow continuous and contextual presence, integrating messages via email, WhatsApp, medical portals, and even voicebots, creating journeys that keep the brand active even between in-person visits always in the right tone, on the right channel, and at the right time, avoiding unnecessary repetition of messages across different channels.
Challenges and Solutions
19
4 Trust
How to build a long-term relationship and real partnership?
Yesterday and Today
Gifts, events, and congresses are still the cornerstones of emotional connection and trust. But these practices face increasing ethical and legal restrictions. This reliance on traditional relationship methods makes it difficult to measure ROI and optimize campaigns, making it hard to determine which actions are truly generating results and preventing efficient resource allocation.
Tomorrow
Building new services and digital experiences useful to the HCP will be the new driver of trust. With Agentic AI, companies can create agents that help doctors in their daily lives suggesting scientific articles, integrating with prescriptions, monitoring patient adherence, and even tracking patient progress after the consultation. These agents could collect data on prescription adherence (did the patient adhere to the treatment?), how the medication is being taken (is it following instructions correctly?), the evolution of the clinical picture (has there been improvement?), and the appearance of possible side effects. By providing this continuous and valuable monitoring, the company reinforces an image of a real partner and not just a supplier. Agentic AI offers a data-driven and results-based alternative, enabling companies to be 37% more likely to measure ROI compared to those using general-purpose AI tools [Live Ramp] and increasing ROI by up to 44% [PwC].
Yesterday
Today
Tomorrow
Volumetry
Low-scale, high-cost in- person visits.
Generic, poorly personalized digital content.
24/7 personalized digital agents for all HCPs.
Quality
Uniform communication focused solely on the product.
Basic personalization using CRM and segmentation.
Real-time hyper- personalization with tailored actions.
Frequency
Frequent visits that lead to saturation and rejection.
Limited frequency due to the physical capacity of field teams.
Omnichannel orchestration with continuous and contextual presence.
Trust
Relationship based on events and gifts.
Ethical restrictions make it difficult to measure ROI and optimize actions.
Digital services that support and engage HCPs on a daily basis.
Challenges and Solutions
20
Chapter 4
Building the Sales Rep Agentic Hub
21
The Future of Medical Representation is Autonomous, Multichannel, and Scientifically Relevant
The routine of the medical representative is becoming increasingly complex. With more selective healthcare professionals, expanding digital channels, and a growing volume of scientific information, maintaining relevance in interactions requires more preparation, intelligence, and agility.
It is in this context that the Sales Rep Agentic Hub emerges:
An integrated set of intelligent tools created to accelerate and empower the sales rep’s work. Instead of operating alone, the representative now relies on the support of specialized agents who work in the background analyzing data, suggesting next steps, adjusting messages, selecting scientific arguments, and identifying the best channel for each contact. In certain contexts, there could even be direct autonomous interactions with HCPs, such as the automatic sending of specific scientific content by email.
In the era of Agentic AI, the sales representative is no longer just a link between the company and the doctor. They transform into an enhanced professional, equipped with a digital ecosystem that delivers more precision, scale, and much more value in each interaction with the HCP.
Sales Rep Agentic Hub
22
From Representative to Intelligent Ecosystem
The Sales Rep Agentic Hub is, in practice, an orchestrated solution of multiple autonomous agents, each playing a specific and highly coordinated role within the HCP’s journey.
This solution combines:
Journey Intelligence Agent
Powered by data from CDPs, CRMs, and digital channels, it continuously monitors and interprets the HCP’s stage in relation to the target, which covers awareness (first contact with the topic or product), interest (demonstration of curiosity and information seeking), consideration (evaluation of evidence and comparison with alternatives), adoption (start of use or prescription), and loyalty (recurrent use and advocacy of the solution), in addition to identifying possible moments of disengagement.
Synthetic Persona Agent
Acts as an intelligent interpreter of the HCP’s 360º view, crossing their data with patterns of similar professionals to predict interests, anticipate objections, detect opportunities, and help other agents operate with greater precision and context.
NBA/NBO Agent (Next Best Action / Next Best Offer)
Responsible for suggesting the next most relevant engagement action for each HCP based on their profile, history, clinical moment, and behavior patterns whether sending a scientific article, initiating a natural voice chat, recommending an explanatory video, or an invitation to a congress.
Sales Rep Agentic Hub
23
Optimal Channel Agent (Next Best Channel)
Defines the best channel for each interaction (WhatsApp, email, medical portal, voicebot, etc.) based on previous preferences, historical performance, and availability context.
Tone of Voice Agent
Adjusts the interaction language to the HCP’s individual profile more technical, more empathetic, more concise, or more educational ensuring that the content is delivered naturally, engagingly, and appropriate for the doctor’s specialty and style.
Scientific Agent (Argument Selector):
Navigates a dynamic database of clinical studies and publications, selecting the most relevant and updated scientific evidence to support the arguments presented, ensuring rigor and technical credibility in each interaction.
Multimodal Generation Agent (GenAI)
Builds content (text, audio, images, short videos, infographics) in real-time, aligning the selected arguments with the appropriate tone of voice and defined channel, delivering personalized and contextualized experiences.
This modular architecture not only exponentially expands coverage capacity but also creates a significantly superior experience from the HCP’s perspective who begins to interact with a “digital partner” capable of understanding their clinical needs, responding scientifically and respectfully, and offering continuous and non-intrusive value.
Sales Rep Agentic Hub
24
The Scientific-Contextual Engagement Machine
Imagine…
A cardiologist in São Paulo who has just participated in a congress on new therapies for heart failure.
>>
The Journey Intelligence Agent detects their presence at the event. >> The NBA agent suggests sending a clinical summary of a relevant study presented at the congress. >> The optimal channel is identified as WhatsApp. >> The tone of voice is adapted to a technical and objective language. >> The scientific agent locates a new meta-analysis published in JACC. >> The content is transformed by the GenAI agent into an interactive infographic and sent in less than 2 minutes. >> The doctor interacts, clicks, consumes the content. >> The system collects implicit feedback (reading time, interaction, clicks) and feeds it back into the database for the next suggestion. All without direct human intervention.
Detects your presence at the event
Interested in summaries of scientific studies
Suggests sending a clinical summary of a congress study
WhatsApp is the optimal channel
Dr. Laura Cardiologist
Adapt the tone of voice to technical and objective language
Sales Rep Agentic Hub
In São Paulo, she has just attended a congress on new therapies for heart failure.
Locates a new meta- analysis published in JACC
All the content collected is transformed into a personalized infographic for the HCP and sent in less than 2 minutes
Sales Rep Agentic Hub
25
Strategic Gains for the Industry
The applications of the Sales Rep Agent go far beyond promotional marketing. They allow:
Scaling coverage with quality
Increase the number of engaged HCPs without overloading human teams, maintaining the level of personalization and technical depth.
Increasing perceived value by doctors
Provide truly relevant scientific content, in the preferred format and channel, with compatible language useful for clinical practice.
Measuring with greater precision
Monitor each interaction in detail, with sentiment analysis, implicit feedback, and a continuous learning cycle.
Optimizing compliance
Agents operate within approved flows, with traceable, auditable logs, reducing regulatory risks.
Reducing operational costs
Automating repetitive tasks and freeing up human teams for high-value strategic interactions.
Recommended Technical Architecture
The implementation of a high-performance Sales Rep Agent involves:
Integration of a robust CDP, with structured and unstructured HCP data;
Sales Rep Agentic Hub
26
Use of LLMs connected via RAG for real-time consultation of scientific databases (e.g., PubMed, Scielo, internal databases);
Orchestration of agents via frameworks such as LangGraph, CrewAI, or AutoGen Studio, with central governance;
Compliance and approval layers integrated into the automated workflow;
Multimodal capability with GenAI for text, image, audio, and video content;
Executive dashboards for real-time performance, engagement, and ROI monitoring.
Final Considerations
The Sales Rep Agentic Hub does not come to replace the representative, but to make their work more efficient, focused, and relevant. By automating what is repetitive and providing real-time intelligence, it frees the sales rep for what really matters: building more strategic, useful, and aligned conversations with each HCP’s moment.
Sales Rep Agentic Hub
27
It is a practical way to deal with the complexity of the current scenario, helping the representative to prepare better, act with greater precision, and respond with greater agility without sacrificing the human touch that remains essential in medical relationships.
More than technology, the Hub represents a change in the modus operandi: less effort to organize, more time to connect. And, with that, more real impact on each visit.
By adopting this approach, the healthcare industry moves towards a new paradigm: less promotional push, more scientific pull, and a more genuine partnership. The future of medical representation is autonomous, yet even more human because it understands the HCP better than ever.
Traditional HCP engagement models have reached their limits. With Agentic AI, we gain the ability to deliver predictive and truly personalized value at scale. This enables us to build strong trust that goes far beyond isolated interactions and genuinely influences clinical decision- making. This is an opportunity to lead the next generation of physician relationships, ensuring a competitive and differentiated ROI while reinforcing our commitment to patient care.
Ricardo Kim, Senior Data Consulting Manager
Sales Rep Agentic Hub
28
Chapter 5
Real Application Cases
29
How Companies Are Implementing Agentic AI in Practice
In this chapter, we will explore examples of how companies in different sectors are adopting Agentic AI in practice, transforming their processes and improving their results. We will present real use cases that illustrate how technology has been applied to optimize operations, improve customer interaction, and enhance efficiency in product development. These practical examples provide valuable insight into the opportunities and challenges faced by organizations, as well as demonstrating the real impact of AI implementation.
Use Case #1
Summary of interactions of HCPs
For a global company of CPG, we created a summary of interactions of HCPs, very similar to a summary of medical bills.
Commercial challenge
Before a face-to-face visit, medical representatives need to gain insight into previous interactions with HCP.
The relevant data is scattered, obtaining and cross-referencing it is manual
The main indicators of interest from HCP are within reach, but are not readily available
The process is time-consuming and results in inconsistent results
Results of implementing GenAI
NPS of 75 for the solution in the medical representatives team
Time savings and improved relationship with HCP are two of the main recurring benefits reported by MRs
Real Application Cases
30
Summary view of interactions offline and online
KPIs and important insights about brand, topics, channel preferences based on previous interactions
Recommendations on next steps, suggestion on best time to share content (e.g. an email)
Why invest?
We are used to creating strategic tools for to increase compliance, optimize engagement and improve patient outcomes. We will use our knowledge and experience to create a tool tailored to your personas and needs.
GCS
(Data storage)
Upload summaries to the user interface
Code
Creating Streamlit image Storing summaries
App Engine
(Streamlit) Vertex GenAI Cloud Creation
Send the interactions
and request in the LLM
Image implementation Upload the image
to the artifact registry GCS (Data Extraction) Registering artifacts
The product was delivered with a product-focused approach and using agile methodology, through sprints to ensure stakeholder alignment with priorities.
This product was delivered on a GCP architecture – we have the technical experts and certified to work on a similar version on AWS or Azure.
Technical challenges
Availability and quality of data to reconstruct all past interactions with HCP (solved by a separate 360° data governance project from HCP)
Obstacles in integrating with the existing CRM tool, requiring an agile approach with iterative sprints to develop the required functionalities.”
Key learnings
Positioning GenAI as an assistance and training tool is key to adoption
It is important to record feedback from the user and set up an evaluation structure to continuously improve the tool
Having well-structured data with clearly defined and unambiguous column descriptions dramatically improves performance
Real Application Cases
31
Use Case #2
Medical Writing Assistant
Getting drugs to patients faster by accelerating authoring of key regulatory medical documents by 30%.
Business Challenge
Writing medical documents, such as Clinical Study Protocols and Informed Consent Forms, is often very time-consuming:
It requires reading and analyzing dozens of reports over 100 pages long.
Requires manual replication of each change throughout lengthy documents
Causes delays in patients’ access to medicines
GenAI Implementation results
30% reduction in time spent preparing medical documents
Use of common knowledge, use of proven terms and standardized templates
This assistant is used by medical writers to speed up the preparation of reports via GenAI (notably by identifying and synthesizing documents, and generating content).
Medical writers retain control over the curation of source documents and the final output, being able to analyze and interrogate texts and tables. The outputs are accurate, offering sources directly within the documents to validate the responses.
Why invest?
We know how to ingest and process large volumes of data efficiently to deliver accurate results. We’ve also developed a similar use case focused on Clinical Study Reporting (CSR) for drug R&D – we know how to start fast and scale with agility.
Architecture Schema
Real Application Cases
32
Azure OpenAI Service
Call through Ragify or manually LiteLLM for prediction
Upload on GCS
Direct Upload
Pipeline de Indexação (mesmo do IB)
rd-import bucket (1 per env)
Parsing-on- the-fly RAG not needed Sections:( Risks and Side Effects, Possible Risks
Cloud workflow + run jobs on icf-rd GCP project
Call through Ragify or manually LiteLLM for embedding & prediction
Vertex AI
All Other Sections
*Weaviate: use team-managed instance on processing-rd or SaaS version depending on advancements
Medical Writing (IFC) Application
Cloud Run Service
Deployedon ict-rd-GCP project
˝Business˛ monitoring: Prompts and Answers + User Feedbacks
Example for the Informed Consent Form (ICF):
Retrieval Augmented Generation architecture, able to ingest 20k+ pages of context
OCR (image recognition) to identify and interpret tables in pdfs
Secure solution (no data/request/prompt sharing outside of the clientˇs env.)
Technical challenges
Finding the right solution with multimodal capabilities & satisfactory accuracy levels to identify and embed tables
Building a Medical Authoring AI enabler replicable on multiple reports
Key learnings
The underlying tech must be tailored to both varying document types (IBs, ICF, CSPs& ) and phases (e.g., initial drafting vs. regulatory consistency checks)
Ensure users retain full control through a human-in-the-loop design
Custom prompting workflows are needed to support existing cross-functional roles
Seemingly simple features like document drag & drop can make or break the toolˇss adoption & success
Real Application Cases
33
Use Case #3
AI Persona Generator
The AI Persona Generator we built for a CPG company can be inspiring for targeting Marketing.
Business challenge
Rapid shifts in consumer expectations and behaviours, an evolving competitive landscape and fragmented data make it increasingly complex for marketers to understand their audience and deliver personalized, targeted campaigns that meet consumers where they are.
GenAI Implementation results
Empowering marketing teams with AI Persona tool resulted in:
5-10% market share gain in the Coolers segment
Improved customer engagement & satisfaction
Faster insights and accelerated time-to-market
Generate personas to identify high-value target audiences and adjust your campaigns before launching them
Chat with personas to better plan your campaign strategies
Personalize content for specific audiences
Simulate campaigns on your synthetic audiences to adjust targeting and content selection based on results
Why invest?
Not only can we help you generate personas based on both structured and unstructured data, we can also help you synthesize customer data to talk to your personas, ensuring your campaign content is truly personalized to your audience.
Persona Examples
Real Application Cases
34
Dr. Tereza Andrade The townˇss doctor
Experience City size
- rient (46+) Sm | all cities
Place of work Representativeness
- ital 100% 25 | %
Experience City size
Dr. Caio Fernandes The on-call specialist
Experiente (46+) Large cities
Place of work Representativeness
Hospital (91,7%), Clinic (8,3%)
17%
Experience City size
Dr. Marina Silva The popular
Little Exp. (31 45) Small cities
Place of work Representativeness
Hospital (75,4%), Others (24,6%)
10%
Dr. Rebeca Moraes The clinic reference
Experience City size
- rient (46+) Sm | all cities
Place of work Representativeness
- c 100% 15 | %
Experience City size
Dr. Roberto Kim The server
Little Exp. (31 45) Large cities
Place of work Representativeness
Hospital (61%), Clinic (5%), Others (33,9%)
13%
Experience City size
Dr. Sam Moraes The miltipurpose doctor
Experient (73,4%) + Little Exp. (26,6%)
Small (59%) + Large (41%)
Place of work Representativeness
Clinic (18,9%), Government (26%), Others (54%)
20%
Real Application Cases
35
Chapter 6
Ethical & Compliance
36
Ethical and Compliance Considerations in Agentic AI for Healthcare
As artificial intelligence advances to fill sensitive roles in interactions between healthcare companies and professionals (or patients), ethics, transparency, and compliance cease to be parallel themes and become core components in the design of these solutions.
Agentic architecture being autonomous, adaptive, and customizable brings unique governance challenges. This chapter addresses the main risks and how to mitigate them from the outset.
Privacy and Continuous Consent
Challenge
Agents operate in real time with personal data, often sensitive (interactions, preferences, clinical patterns). It is essential to ensure that all processing is done based on informed, granular, and auditable consent.
Good practices
Dynamic and revocable consent.
Clear separation between identifiable and aggregated data.
Use of synthetic data or privacy-preserving AI techniques (differential privacy, homomorphic encryption) for model training.
Ethical & Compliance
37
Accountability: Who is responsible for
automated decisions?
Challenge
If an agent recommends clinical content, a channel action, or a sensitive response who is responsible? The company? The developers? The medical teams?
Recommendations
Create a decision log accessible and auditable by humans.
Adopt “human-in-the-loop” frameworks for critical decisions.
Map risks by agent type (content, channel, journey) with impact matrices.
Bias and Representativeness
Challenge
Models trained on historical data can perpetuate biases (by specialty, region, HCP profile, etc.).
Good practices
Evaluate agent performance by demographic group and clinical profile.
Adopt fairness-aware ML practices.
Review prompt and message content for inclusive and accurate language.
Transparency and Explainability
Challenge
Agents need to be understood by users and regulators especially in medical interactions.
Ethical & Compliance
38
Good practices
Provide simplified explanations of agent actions (“why did I receive this?”).
Display scientific sources or references when content is clinical.
Create internal dashboards for agent traceability, with confidence and accuracy metrics.
Recommended Governance Frameworks
EU AI Act (2024): defines risk categories for AI. Medical decision or prescription agents are ˝high risk˛ a and require rigorous documentation.
Good Machine Learning Practices (GMLP FDA): guide for ethical AI development in healthcare.
ISO/IEC 42001:2023: ew international standard for AI system governance.
Recommendation
Create an Agentic Governance Committee with members from Legal, Compliance, Medical Affairs, and Tech to evaluate, validate, and supervise the evolution of agents, ensuring continuous security and regulatory alignment.
Ethical & Compliance
39
Chapter 7
Implementation Roadmap for Agentic AI
- i
40
Adoption Roadmap: How to Implement Agentic AI
Artefact can support organizations’ journey in implementing Agentic AI, focusing on value capture in a structured and sustainable way. Successful adoption begins with a clear roadmap built around three key pillars: Individual Level, Use Cases with Real Impact, and Transformation at Scale.
In this chapter, we will explore each of these pillars in detail, addressing the key steps to effectively implement Agentic AI.
We will start with the Individual Level, where the focus is on empowering employees to make the most of available agents, promoting knowledge democratization.
Next, we will discuss Use Cases with Real Impact, which help the organization prioritize and develop high-impact use cases to generate value quickly and tangibly.
Finally, we will talk about Transformation at Scale, which prepares the organization to expand AI use technically and culturally, structuring a program that ensures the scalability, robustness, and security of the agents.
Adoption Roadmap
41
How we can support your journey with Agentic AI – To capture value in a structured and sustainable way, we work on three pillars:
Individual level
We enable individuals to make the most of the agents already available
Use cases with real impact
Select 1 process: Purchasing, HR, GTM…
Developing an agent for a specific activity, helping to demonstrate value quickly
M a t u r i t y
Scale Transformation
We help design a structured program to map opportunities and prioritize initiatives across the organization
Adoption Roadmap
42
Deep Dive: Individual PoV
Individual level
We help people harness the power of Agents through:
Democratization of knowledge: Training people to leverage these tools in their daily work
1.
2.
Tools: Implementing a secure and private. GPT in
your environment
We allow individuals to make the most of the agents already available to them
Governance: Allow self-service creation of agents with tools such as n8n and Zapier
3.
Deep Dive: Use Cases with real impact
Use cases with real impact We accelerate your Agentic transformation through pilots of use cases:
Prioritize and launch agent pilots within a few weeks based on value potential
1.
2.
Transform key functions to achieve gains in
productivity, lead time and NPS
Development of an agent for a specific activity, helping to demonstrate value quickly
Apply AI to real processes, proving the impact and laying the foundations for scaling the use cases
3.
Adoption Roadmap
43
Deep Dive: Scale transformation through a
transversal program
Scale Transformation
We help you create a structured program to map opportunities and prioritize initiatives across the organization
We prepare your organization to scale Agentic AI, both technically and culturally:
Structuring the program (~2 months) AI platform readiness
Identify priority functions and processes 1.
Create or improve AI foundations (AgenticOps platform)
1.
Sizing up the business value at stake 2.
Ensure agents are scalable, robust and secure
2.
Design and launch an operating model to scale the use of Agentic AI
3.
Incorporate a cultural change, promoting adoption and long-term impact
4.
While implementing Agentic AI may appear complex, Artefact is here to make the journey actionable and grounded. With our structured methodology and focus on concrete results, we offer the necessary support to empower your team, accelerate transformation, and ensure scalable and secure adoption. Together, we will help transform your business and harness all the benefits of AI effectively and sustainably.
Adoption Roadmap
44
Artefact
Artefact as Your Strategic Partner
- r
45
Why Artefact is the right partner
In the era of generative artificial intelligence and data-driven marketing, choosing a partner with real experience makes all the difference. Artefact has already supported major pharmaceutical and biotech companies in Brazil and worldwide, helping these companies transform their relationship with healthcare professionals in a smarter, more personalized, and efficient way.
Our differential lies in the unique combination of AI, product, and business teams. This allows us to design solutions that connect strategy and operation, with deliveries that go beyond technology: we create adaptive journeys, personalization at scale, and actionable recommendations with direct impact on the day-to-day of marketing and sales teams.
Why trust Artefact:
1. Proven industry experience
We have vast experience in the Healthcare sector, delivering practical solutions and measurable results for leading companies. We operate at all stages of the value chain, with an emphasis on innovation, efficiency, and sustainable growth.
2. Structured and Personalized Methodology
Our approach combines detailed analyses with tailor-made solutions, balancing strategic vision and practical execution. We continuously monitor results to ensure our clients’ success.
Artefact
46
3. AI Vision as a Business Unit
We incorporate the vision of Data and Artificial Intelligence as value-generating assets, so that technology serves the business. Focusing on solving business area challenges, applications must provide tangible and measurable results, whether through more assertive decision-making, process optimization and automation, or the generation of actionable insights in real time, positioning your company ahead of the competition.
4. Commitment to Concrete and Sustainable Results
Our goal is to deliver solid and far-reaching results, promoting continuous innovation and strategic efficiency to position your company as a market benchmark.
Our proven impact
Over years of partnership in driving business through data and AI solutions and strategies, weˇve made a significant impact consistently adding value to the methodologies and growth initiatives of our clients and partners.
35%
Increased Operational Efficiency
Implementation of customized AI solutions to optimize internal processes at a major consumer goods company, resulting in a faster and more efficient operation.
20%
Reduction in operational costs through the application of advanced automation technologies
40%
Improved demand forecasting accuracy using AI-based predictive analytics
Artefact
47
What our clients say about us!
Our projects have already delivered significant gains to our clients. Here are a few testimonials from those whoˇv ve seen the results up close.
Demand and Supply
Forecasting
Stockout
Forecasting
Some of our clients
Some of our clients: We work with some of the most recognized brands in the sector. We partner with over 1,000 clients worldwide, including more than 300 major international brands.
Artefact
48
WE OFFER END-TO-END
DATA & AI SERVICES
Strategy
& Transformation
Data & AI Strategy
Data & AI organization
Data Maturity Assessment
Corporate Training
Hackathons
Data & AI Days
GenAI Academy Artefact AI Summits
AI
Acceleration
AI & Gen AI Factory
Data & AI for Operations
AI for Customer Care
Data & AI for Private Equity
Data
Foundations & BI
Data Governance & Management
Data New BI Self Business Intelligence
Data for Sustainability
IT
&
Data Plataforms
Data-Centric IT
Cloud Services
Tech-Agnostic Solutions
Smarter Decision- Making
Optimized IT Operations
Marketing
Data & Digital
Data Valorization & Category Management
Consumer Data Environment
Measurements (MROI) & Insights
Marketing Analytics
GMP Certified Reseller Data-driven Sales
Expertise by Industry:
FMCG | Retail & E-commerce | Luxury & Cosmetics | Healthcare | Banking & Insurance | Telecommunications | Sports & Entertainment | Travel & Tourism | Public Sector & Government | Real Estate | MANUFACTURING & Utilities
Artefact
49
Conclusion
From Supporting Roles to Key Players
50
The Future of Influence in Healthcare with Agentic AI
We are witnessing a turning point in the relationship between the pharmaceutical industry and healthcare professionals. The adoption of Agentic AI is not just another technological evolution it is a profound redefinition of how we deliver value, build trust, and support clinical decisions in an increasingly complex world.
Throughout this eBook, we explored the transformations in the sector, the current challenges in interactions with HCPs, and how Agentic AI can act proactively, personalized, and scientifically relevant. We saw that traditional methods still dominated by in-person interactions, generalized approaches, and slow processes
can no longer cope with the speed and sophistication demanded by the current scenario.
The future, however, is not distant. It has already begun to materialize in the initiatives of pioneering companies that are applying Agentic AI to automate journeys, amplify the performance of representatives, and deliver meaningful and ethical experiences to HCPs. This is the new normal: an operation orchestrated by intelligent, autonomous, and always compliant agents.
Conclusion
51
The adoption of this new approach requires vision, preparation, and a commitment to responsible digital transformation. But the benefits are clear: greater efficiency, genuine engagement, and, above all, more value delivered to professionals and the patients they serve.
If there is one certainty at this moment of transition, it is that companies that begin to experiment and scale Agentic AI now will be shaping the next chapters of healthcare history.
Conclusion
52
Links & References
Scientific and Industry Articles
DOCEREE NUMBER ANALYTICS
The role of personalized communication in HCP marketing Available at: https://blog.doceree.com/ role-of-personalized-communication-in-hcp- marketing. Accessed on: 4 jul. 2025
7 AI stats for ad targeting success. Available at: https://www.numberanalytics.com/blog/7-ai-stats- ad-targeting-success. Accessed on: 4 jul. 2025.
FIERCE PHARMA
How can pharma avoid spamming HCPs? Synchronized campaigns and storytelling marketing. Available at: https://www.fiercepharma.com/ marketing/how-can-pharma-avoid-spamming-hcps- synchronized-campaigns-and-storytelling- marketing. Accessed on: 4 jul. 2025
PATTEN, S. et al
ChatGPT MD: Characterizing the Adoption and Experiences of Users of AI-Generated Health Information A Cross-sectional Survey Study. ResearchGate Preprint, 2024. Available at: https:// www.researchgate.net/ publication/376254560_ChatGPT_MD_Characterizi ng_the_Adoption_and_Experiences_of_Users_of_AI- Generated_Health_Information_-_A_Cross- sectional_Survey_Study_Preprint. Accessed on: 27 may 2025.
FRANKLY PHARMA
Social media influence on HCPs. Available at: https://www.franklypharma.com/blog/social- media-influence-on-hcps. Accessed on: 4 jul. 2025
MAHER, C. et al
Use of ChatGPT to obtain health information in Australia in 2024: insights from a nationally representative survey. Medical Journal of Australia, v. 222, n. 4, 2025. Available at: https:// www.mja.com.au/journal/2025/222/4/use-chatgpt- obtain-health-information-australia-2024-insights- nationally. Accessed on: 27 may 2025.
PR NEWSWIRE
New Jasper research reveals early AI wins for marketers in productivity & ROI, but key gaps remain. Available at: https://www.prnewswire.com/news- releases/new-jasper-research-reveals-early-ai-wins- for-marketers-in-productivity–roi-but-key-gaps- remain-302392725.html.Accessed on: 4 jul. 2025.
Links & References
53
Reports, Studies, and Regulations
ACCENTURE The economic potential of generative AI
The economic potential of generative AI: The next productivity frontier. Available at: https:// www.mckinsey.com/capabilities/mckinsey-digital/ our-insights/the-economic-potential-of-generative- ai-the-next-productivity-frontier. Accessed on: 26 may 2025.
Technology Vision 2023: When Atoms Meet Bits. Available at: https://www.accenture.com/br-pt/ insights/technology/technology-trends-2023. Accessed on: 26 maio 2025.
FEDERAL COUNCILS
MEDICINE NURSING The future of field teams in pharma
DENTISTRY
FEDERAL COUNCIL OF MEDICINE; FEDERAL COUNCIL OF NURSING; FEDERAL COUNCIL OF DENTISTRY; FEDERAL COUNCIL OF PHARMACY. Estimate of the total number of healthcare professionals in Brazil in 2024. Available at: https:// portal.cfm.org.br; http://www.cofen.gov.br; http:// www.cfo.org.br; https://www.cff.org.br. Accessed on: 4 jul. 2025.
EVALUATE
Available at: https://www.evaluate.com. Accessed on: 26 may 2025.
PHARMACY Available at: https://www.mckinsey.com/ industries/life-sciences. Accessed on: 26 may 2025.
UNIÃO EUROPEIA
Regulamento (UE) 2024/1689 do Parlamento Europeu e do Conselho, de 13 de junho de 2024, que estabelece regras harmonizadas em matéria de inteligência artificial (Lei da IA) e altera determinados atos legislativos da União. Jornal Oficial da União Europeia, L, 2024/1689, 12 de julho de 2024. Available at https:// artificialintelligenceact.eu/the-act/. Accessed on: 28 may 2025.
ISO/IEC
INTERNATIONAL ORGANIZATION FOR STANDARDIZATION (ISO); INTERNATIONAL ELECTROTECHNICAL COMMISSION (IEC). ISO/IEC 42001:2023 – Information technology Artificial intelligence Management system. Available at: https://www.centraleyes.com/iso-iec-42001/. Accessed on: 28 may 2025.
MCKINSEY & COMPANY
Commercial model evolution in pharma
Available at: https://www.mckinsey.com/ industries/life-sciences. Accessed on: 26 may 2025.
US/CANADA/UK
U.S. FOOD AND DRUG ADMINISTRATION (FDA); HEALTH CANADA; UK MEDICINES AND HEALTHCARE PRODUCTS REGULATORY AGENCY (MHRA)
Good Machine Learning Practice for Medical Device Development: Guiding Principles. Available at: https://www.fda.gov/medical-devices/software- medical-device-samd/good-machine-learning- practice-medical-device-development-guiding- principles.Accessed on: 28 may 2025.
Links & References
54
Glossary
Accountability APIs
In the context of Agentic AI, refers to determining who is responsible for decisions made by automated agents, especially in sensitive medical interactions.
Application Programming Interfaces: Used by Agentic AI solutions to dynamically integrate data, allowing agents to connect to various systems.
Adoption Roadmap Artificial Intelligence (AI)
A structured plan based on three pillars (Individual Level, Use Cases with Real Impact, and Transformation at Scale) to ensure the successful adoption of Agentic AI.
A field of computer science that develops systems capable of performing tasks that typically require human intelligence, such as learning, decision-making, and pattern recognition.( Bias and Representativeness: Ethical challenge of ensuring that models trained on historical data do not perpetuate biases (by specialty, region, HCP profile, etc.).
Agentic AI
A type of artificial intelligence that adds autonomy and decision-making capabilities to systems, allowing them to act proactively, learn from experience, and execute complex tasks with minimal human intervention. It mimics human reasoning and autonomously explores complex problems, executes actions, and makes decisions.
CDP -Customer Data Platform
A robust platform for integrating structured and unstructured HCP data, crucial for feeding the Journey Intelligence Agent in a Sales Rep Agentic Hub.
AI Pipelines
Automated flows of ingestion, processing, and decision, connecting data to action.
Compliance
A set of rules and regulations that companies must follow to ensure legal and regulatory compliance, especially in the Healthcare sector. In the pharmaceutical industry, this is fundamental to ensure ethical and legal operations.
Links & References
55
Customer Relationship Management ERP
Systems used to manage and analyze customer interactions and data throughout the customer lifecycle. In the context of HCPs, CRM tools assist in initial personalization and segmentation. Integration challenges with existing CRM tools are highlighted in the document.
Enterprise Resource Planning: Systems used by clinics that can offer valuable inputs to personalize interactions, predict needs, and accurately measure engagement.(
EU AI Act (2024)
Data Governance Law that defines risk categories for AI. Medical decision or prescription agents are considered “high risk” and require rigorous documentation.
A set of processes and policies that ensure the quality, security, and compliance of data use within an organization.
Frameworks
Data Security (e.g., LangGraph, CrewAI, AutoGen Studio): Tools for orchestrating multiple AI agents, ensuring they operate synchronously and efficiently.(
Practices and technologies used to protect sensitive information against unauthorized access, leaks, or cyberattacks.
Digital Transformation
The process of integrating digital technologies into all areas of a company to improve operations, efficiency, and customer experience. In the healthcare sector, this involves changes in patient behavior and expanded access to data about HCPs.
GenAI
Generative Artificial Intelligence: A subfield of AI focused on creating new content, such as text, images, code, and audio, from models trained on large volumes of data. It is widely adopted to simplify various tasks through information retrieval, document production, and synthesis.(
Digitalization of Interactions: GMLP
The digitalization of interactions with healthcare professionals has created a highly fragmented ecosystem, but also one full of opportunities.
Good Machine Learning Practices – FDA: A guide for ethical AI development in healthcare, recommended for governance.
Links & References
56
HCPs Machine Learning (ML):
Health Care Professionals: The central actors in the healthcare ecosystem, including doctors, who play a unique and strategic role as trusted intermediaries between the industry and the patient.
An AI technique that allows systems to learn from data and improve their performance without explicit programming. Fairness-aware ML practices are recommended to address biases.
Human-in-the-loop Multimodal Generation Agent
A recommended framework for critical decisions, where human intervention is part of the automated decision-making process.
Builds content (text, audio, images, short videos, infographics) in real-time, aligning the selected arguments with the appropriate tone of voice and defined channel, delivering personalized and contextualized experiences.
ISO/IEC 42001:2023
New international standard for AI system governance, recommended for adoption.
Journey Intelligence Agent
NBA/NBO Agent
Next Best Action / Next Best Offer: Responsible for suggesting the next most relevant engagement action for each HCP based on their profile, history, clinical moment, and behavior patterns.
Continuously monitors and interprets the HCP’s stage in relation to the target (covering awareness, interest, consideration, adoption, and loyalty), in addition to identifying possible moments of disengagement.
OCR – Optical Character Recognition
LLMs – Large Language Models
A technology mentioned in a use case for identifying and interpreting tables in PDF documents.
Large-scale language models that use advanced neural networks to understand and generate natural language, enabling applications such as chatbots and virtual assistants. They are combined with autonomous and interactive decision flows in Agentic AI.
Omnichannel Orchestration
The ability to integrate messages through various channels such as email, WhatsApp, medical portals, and even voicebots, creating journeys that keep the brand active even between in-person visits.
Links & References
57
Optimal Channel Agent Real-time Hyper-personalization
Next Best Channel: Defines the best channel for each interaction (WhatsApp, email, medical portal, voicebot, etc.) based on previous preferences, historical performance, and availability context.
A new standard of communication where intelligent agents deliver individualized Next Best Actions (NBAs), such as scientific content, invitations to congresses, or clinical updates, based on historical, contextual, and predictive data.
Privacy and Continuous Consent
RPA Ethical challenge of ensuring that all processing of personal data, often sensitive, is governed by informed, granular, and auditable consent.
Privacy-preserving AI
Robotic Process Automation: Technology that uses software to automate repetitive business processes, increasing operational efficiency. It focuses on automating back-office operations with simple and static logic orchestration in monotonous tasks. Agentic systems combine RPA with Generative AI for autonomous task performance.
Techniques such as differential privacy and homomorphic encryption, used to train models while protecting sensitive data.
Return on Investment
Prompts A metric used to evaluate the efficiency of an investment. Agentic AI is expected to generate an ROI increase of up to 44%.
Inputs used to guide AI models in content generation. Prompt personalization is crucial to support existing cross-functional roles.
Sales Rep Agentic Hub
RAG
An integrated set of intelligent tools created to accelerate and empower the sales representative’s work, enabling autonomous, multichannel, and scientifically relevant medical representation.
Retrieval-Augmented Generation: A technique used in generative artificial intelligence systems, where the model combines text generation with retrieval of relevant information from a database or external documents to increase the accuracy and relevance of responses. It allows agents to consult external sources in real time before making decisions, increasing accuracy.
Links & References
58
Scientific Agent – Argument Selector Telemedicine
Navigates a dynamic database of clinical studies and publications, selecting the most relevant and updated scientific evidence to support the arguments presented, ensuring rigor and technical credibility in each interaction.
The practice of providing healthcare services remotely, using technology, which has become part of the new value offering of HCPs.
Tone of Voice Agent
Scientific-Contextual Engagement: Adjusts the interaction language to the HCP’s individual profile more technical, more empathetic, more concise, or more educational ensuring that the content is delivered naturally, engagingly, and compatibly with the doctor’s specialty and style.(
The concept of a “machine” that orchestrates hyper-personalized and scientifically relevant interactions with HCPs, using various AI agents.
Strategic Gains
Benefits obtained from implementing the Sales Rep Agentic Hub, including scalability, perceived value by doctors, accurate measurement, compliance optimization, and reduced operational costs.(
Transparency and Explainability
Ethical challenge of ensuring that agents are understood by users and regulators, especially in medical interactions.(
Synthetic Data Voicebots
Data that is artificially generated rather than collected from real-world events, used for model training while preserving privacy.
AI-powered conversational agents that can interact with users through spoken language, used in omnichannel orchestration.
Synthetic Persona Agent
Acts as an intelligent interpreter of the HCP’s 360° view, crossing their data with patterns of similar professionals to predict interests, anticipate objections, detect opportunities, and help other agents operate with greater precision and context.
