Scaling Noise or Building Trust? HCP Engagement in the Age of AI

HCPs experience one company, not your organisation chart

Karim opened with something an HCP told him: every company has the same website, the same emails and the same slide deck, and the doctor could not tell them apart. His diagnosis is that pharma does not have an engagement volume problem. It has an engagement value problem.

Inside a company, a global email, a local email, a medical follow-up, a congress activation and a sponsored post all belong to different teams. The HCP experiences them as one company, and as the sum of all its interruptions. Each interaction can be correct on its own while the whole experience stays repetitive, inconsistent or disconnected. Plenty of internal alignment meetings do not add up to a coherent experience on the outside.

That is why omnichannel cannot simply mean more channels. Real orchestration asks what happens next: a conversation in the field should improve the next digital experience, and digital behaviour should sharpen the next conversation with a person.
 

What AI does to a system that is already fragmented

As Karim sees it, AI can produce content, create variants, support targeting and surface patterns, and it makes the whole engagement machine much faster. Speed is not direction, though. AI arrives in a system that already contains duplicated activity, competing incentives and metrics that reward visible activity, and it amplifies whatever operating model it is given. If the strategy is undifferentiated, AI scales undifferentiated engagement. If the organisation rewards volume, AI produces more volume. The bigger risk is that AI industrialises the blind spots that are already there.

Manuel Mitola puts it more simply: garbage in, garbage out. Access to AI is becoming a commodity, the way access to the web once did, so what will set companies apart is how they use it. He sees two good places to start. Use AI to listen better, by extracting insight from more data than any person could work through. And use it to tailor engagement to the individual HCP, while the thinking stays with people and only the execution is handed over.
 

What your metrics are teaching the organisation

Opens, clicks, impressions, calls and cost per engagement tell you whether the machine moved. They do not tell you whether the customer moved. Karim recognises that engagement cannot be linked to prescribing for compliance reasons, but argues that a change in behaviour can still be assessed. He suggests asking of every interaction:

• Did it resolve a meaningful need?

• Did it remove a barrier?

• Did it increase the HCP's confidence?

• Did it make the next interaction more relevant?

• Did we know when not to engage?

Metrics are never neutral. They teach the organisation, and increasingly the algorithm, what is valued.
 

Earn the right to talk about your product

Dan Buckland's concern is that scaling today's engagement with AI could damage HCP trust for good. What HCPs say they value is relevance, trustworthy and balanced information, and content that is clear and concise, far more than any particular channel.

Context is where engagement most often goes wrong. Picture a doctor who has just finished a shift and is sitting in her car. If you tapped on the window, you would not open with your mechanism of action or your latest data. The same holds when she is reading email or scrolling social media. Start by helping with the problem she has right now, without your product, and earn the right to make a recommendation later.
 

From complex journeys to a simple funnel

Dan asks a question most brand teams find hard to answer: how does a doctor who does not use your product today actually become a customer? Instead of elaborate omnichannel journey maps, he proposes the kind of marketing funnel other B2B industries have used for years. It moves engagement from push to pull, and every step can be measured.

• Attract. Offer something genuinely worth engaging with, such as a resource, educational content or a regular newsletter, in exchange for an email address and two or three answers that tell you what the HCP needs.

• Nurture. Design a specific sequence that takes a consented subscriber to customer, built around the problem they told you about.

• Onboard. Help the HCP succeed with their first patient, through case studies and contact with experienced peers.

• Advocate. Confident HCPs help co-create content, which in turn attracts the next ones.

Listening is built into every stage. With a model like this in place, AI becomes genuinely useful, because it can generate the regular content and insight the funnel runs on.
 

Segment by the problem, not the persona

Asked about digital segmentation, Dan's advice is to keep it simple. Identify the few problems in your therapy area that your product can address, and let each one lead a genuinely different story: overall survival, response rate, safety and tolerability, or patient expectations. Every download or webinar registration is a chance to ask which problem matters most to that HCP. That gives you a segmentation you can automate, without chasing a one-to-one message for every doctor.

Manuel sees more room for AI here, for example an always-on assistant that gives HCPs reliable answers whenever they need them, provided it is fed with relevant data. Dan's counterpoint is that doctors already have AI tools tailored to them, so a pharma tool only adds value if it offers something they cannot get anywhere else.
 

Fixing the complexity inside the company

More agencies, AI vendors and tools can add to the fragmentation instead of solving it. Karim's view is that ways of working need to be redesigned for the system as a whole, rather than one partner at a time:

• Clear roles. Who owns the customer need and the outcome, and who creates, adapts and approves.

• One shared brief. Common evidence and aligned success measures, so agencies, AI partners, medical, marketing and field teams work from the same interpretation of the objective.

• Global and local on the same side. Global provides the strategic platform, reusable assets and guardrails. Local brings customer insight, cultural relevance and execution.

• Organised around the customer. Structure the work around the problem to be solved, not around functions, channels or suppliers.

Manuel adds a practical point: choose AI partners who have a strong understanding of pharma.
 

About the speakers

• Karim Istafanous is Global Oncology Strategic Marketing at Boehringer Ingelheim, with over 17 years of experience in oncology and haematology. He has led cross-functional teams through multiple oncology product launches at Boehringer Ingelheim, Eli Lilly, and Novartis, building deep expertise across strategic marketing, market development, and HCP engagement, including the design and management of Key External Expert programmes across complex therapeutic landscapes. Karim is a passionate advocate for rethinking how pharma engages healthcare professionals in the age of AI, bringing a frontline commercial perspective to a topic the industry is still figuring out.

• Dan Buckland teaches CELforPharma's The Pharma Customer Engagement Course. He is Agency Director at Brandcast Health, leads an award-winning agency recognised for its customer-centric, inbound content marketing in the pharma industry. Dan is an influential speaker and digital opinion leader, known for bringing fresh, practical perspectives on omnichannel engagement and content strategy.

• Manuel Mitola teaches CELforPharma's The AI for Pharma Marketing Course. He is Partner & Managing Director at ctcHealth, a consultancy specialising in transforming pharma marketing, sales, and commercial excellence through AI adoption. Manuel is an experienced speaker on AI at international pharma conferences, and an active voice on LinkedIn and YouTube, where he shares practical case studies.