top of page
  • Writer's pictureKarsten Schmidt

Fundamentally remodel omnichannel customer journey concept comprehensively integrating AI

Providing the right content through the right channel to the right customer at right time continues to be a challenge in today´s more complex omnichannel customer engagements with HCPs. We have asked Giacomo Porzio, Founder of Hyntelo, how Artificial Intelligence (AI) can assist field forces in building better journeys for their customers.

Karsten: Thanks for taking the time for this interview, Giacomo. To start with, could you please break down for us a bit the important elements that enable building better AI-supported customer journeys?

Giacomo: Thanks for having me. First of all, in order to adopt AI-supported journeys, organizations should embrace a customer-centric, data driven strategy. Embracing it means to dedicate work and executive commitment to the development of a vision, a culture and a corporate mindset, all aligned with such strategy.

Secondly, a dedicated plan should be prepared on the implementation of the above customer-centric vision, with the objective to leverage data to drive the creation of better customer engagement through personalized journeys.

Another important factor is to build internal consensus and educate (or upskill) internal teams about customer centricity: teams have to trust technology and make the best use of it.

At this point, you need enabling assets from sales, marketing and medical teams to generate a content mix (promo/non-promo, channels, modules...) that allows personalization for different customer personas and steps along the customer journey.

Regulatory, Legal and Compliance Support also play a key role in success. Finally, there’s the tech part: the skills and tools that enable the execution of the above vision.

Illustration: Foundation for building better AI-supported customer journeys

Karsten: From your experience, how much customization is needed to make AI solutions fit individual Life Sciences clients?

Giacomo: AI is often considered a buzzword for the Machines-Can-Do-Anything concept. The truth is, machines are quite "stupid" and you need to tell AI what to do. Teaching AI to do its job is the “training” phase of an AI project, which is essential for the good result of any project.

To make it work in Life Sciences, three steps need to be taken into account:

  1. The first is to choose an appropriate set of algorithms that perform a specific task. Predicting a future behaviour by looking at past events (time series) is different from segmenting and clustering a group of individuals or analysing semantic entities in a document. Each of these three tasks are performed differently for different AI models.

  2. Secondly, each of these models should be trained to maximise the performance or predictive power in a context-specific environment (i.e., Pharma & Life Sciences). Of course, having a set of pre-trained models is a competitive advantage for tech vendors.

  3. Last, but not least, each company has access to different data, and customer behaviour may vary from one market to another. Therefore, re-training a model for a specific customer context is often preferable to maximise the model's performance.

Even though all this may seem quite technical and complex, the good news is, nowadays there are tools and skills that make AI solutions more transparent so there is very limited impact on the company.

Karsten: Field teams and marketeers in Life Sciences want to understand whether their customers are changing their behaviour and are moving up or down the adoption ladder based on the engagement feedback that they are providing. How can such a dynamic segmentation be obtained and adjusted over time?

Giacomo: First of all, there must be a strong focus on first-party data (i.e., data that is generated directly by the company interacting with its customers) and, to allow profiling and segmentation of audiences, a great deal of attention should go into obtaining consent for data collection in line with GDPR (if in the EU). Currently, CRM systems do a good job of collecting data on customer interactions. However, CRM systems are often not enough to draw a 360° picture of a customer's behaviour. We, thus, need to gather data from external data sources as well (both structured and unstructured), bring them together, and build unified customer views, providing business teams with the information they need to monitor the changes in HCP engagement status.

Such data needs to be refreshed periodically, ideally weekly or daily, so some work must also be done by IT departments.

Part of our work is helping customers design a roadmap to get to a unified and holistic view of customer engagement.

Our approach could be summarized as follows:

  • Definition of an ambition

  • Analysis of the current status

  • Roadmap and steps to achieve the ambition

  • Collection of required data from different channels into a unique system (e.g., F2F visits; Approved Emails; phone calls & web calls; website interactions; traditional customer profiling & segmentation; marketing email campaigns; mobile app logins & interactions; SMS / WhatsApp interactions; events attendance; call center interactions; content-specific views; tagging data of all customer-facing contents, both promotional and non-promotional)

  • Customization of models (we have built a solid methodology to measure and track customer engagement in a channel-specific way)

  • Collection of feedback from Field Reps

All of the above is utilised to build a unified and consolidated customer view that helps to understand customer beliefs, needs and preferences in terms of content, channels, etc. This not only allows the evolution of customer engagement to be tracked across channels and the results of marketing or sales actions to be measured, but also helps to understand how to improve customer engagement over time.

llustration: Roadmap for a unified and holistic view of customer engagement

Data collection strategies and approaches are also very important. For instance, moving away from the traditional push approach (which is still used by the majority of Life Science companies), where the company sends mono-directional messages to the customer, is one key area of improvement. With the objective of understanding customers' needs, behaviours and preferences, it is also helpful for the journey to include a pull approach (where the customer can make requests and interact in a bidirectional way) and feedback collection.

Finally, tools can be deployed to support field teams with AI-driven suggestions on the Next Best Action at the individual HCP level.

Karsten: Field forces tend to be quite sceptical when an AI solution suggests Next Best Actions or Content. What is the most promising way to overcome such scepticism?

Giacomo: This is a classic. Field Reps are convinced they know the customers best and how to do their jobs better than others. And they are right in thinking so! The best way to succeed in an AI-driven solution that suggests the Next Best Actions to Sales Reps is to learn from the best. And that is what we do, in a very peculiar way.

Our methodology for the roll-out of NBA involves (without going into too much detail) selecting the best performing Reps on each business line in order to customize solutions that somehow replicate their behaviour (and this is especially beneficial for others who have less experience or lower performance levels) and provide useful tips that may have otherwise not been considered.

In terms of content, we need to make sure the system is trustworthy. We achieve this in three ways:

  • Proving that the suggestions are useful and win-win for all departments. This means Reps get to their targets faster, marketing is pleased because their strategy is executed effectively and customers are engaged with content relevant to them.

  • Suggesting specific content, down to a single slide or email, to be shared with a customer and showing the motivations that generated the suggestion (i.e., "the HCP has done similar research lately").

  • Providing an explainable AI. Nowadays purely “black box” models are not necessarily the best. Leveraging a mix of models and technologies that enable solid performance without losing the possibility of explaining why a certain action was suggested is the key to success.

Illustration: Ensuring trustworthy Next best action suggestions

Karsten: I understand that you worked with other industries before building AI solutions for Life Sciences companies. What were important learnings from these other industries that are also applicable to Life Sciences?

Giacomo: Our first expertise is in the Financial Services Industry. You can imagine the enormous amount of data that is available there and has been for a few decades.

For this reason, FSI has been a pioneer in many fields of AI and data analytics in general - also because the potential returns were massive.

We are bringing to the Life Science industry a lot of expertise from the years of work that we did (and still do) in FSI, especially in terms of:

  • Systems for data management with no compromise on performance

  • Statistical and data-driven models to segment clients based on behaviour

  • Predictive and prescriptive models

  • Building digital, customer-facing solutions that enable the optimization of operations and an increase in revenues by leveraging data.

In recent years we have learned a lot. In particular, AI solutions that are not designed in collaboration with end users have a very high probability of failure. If there is no evident need for it, don’t build it.

Furthermore, we believe that AI performs better than simpler models. If, however, complexity (e.g., there is not enough data, the systems in place are not there yet) and costs are significantly higher, it is worth considering alternative, simpler approaches that may solve (at least for some time) a specific pain point.

Karsten: AI-supported omnichannel content orchestration is often primarily aimed at supporting Commercial Operations teams. In your view, what importance does this topic have for Medical Affairs?

Giacomo: Medical Affairs, in my opinion, are gaining more and more importance and have been neglected by tech vendors of Next Best Action / Content solutions far too long. This is the primary reason we are bundling our solution with a set of external data sources that enrich the informative base of Life Science Companies, allowing them to build suggestions that are not only useful for Sales Reps, but potentially also for MSLs.

For instance, these external data sources allow for the identification of emerging KOLs (very often not even in CRM targets), affiliations that are becoming a reference for the scientific community, networks of scientists to maximise the impact of MSL information, and so on. Tracking medical content distributed to customers is currently a challenge to the improvement of customer centricity and we want to play a key role in creating an all-in-one integrated solution.

Karsten: Thank you for shedding some light on this rather complex topic, Giacomo. I think that our readers now have a better understanding of cross-functional efforts required and cultural change needed to start building AI-supported customer journeys.

To sum up we can say that first-party as well as external data sources are needed to build a unified and consolidated customer view to better understand their beliefs, needs and preferences across different channels. This will be the basis to measure the impact of sales and marketing actions which ultimately shall lead to improving customer engagement over time. Training AI systems by learning from behaviour of high-performing representatives helps generating Next Best Action recommendations for other field team members. Trustworthy Next Best Actions need to be useful, explainable and must provide specific content at specific moments during the customer engagement journey.

Which other elements are fundamental from your perspective to build AI-supported customer journeys? Please feel free to comment below.

For further reference please contact or


Xeleratio Consulting Ltd.

We help Life Sciences executives improve sales performance with innovative best-in-class Business Excellence tools and methodologies . Expertise in Business Excellence has been gained with over 12 years of working in different global and regional roles in the Life Sciences industry.


Please feel free to reach out and request a free strategy session whenever convenient. See all articles on Xeleratio Consulting Blog


bottom of page