Marshall Thompson knew he was pushing the envelope when he set out to deliver an AI-powered creative campaign for Alnylam Pharmaceuticals’ recent disease state awareness campaign. The idea was to help healthcare providers better understand primary hyperoxaluria type 1 (PH1), a frequently undiagnosed metabolic disease. But using AI to deliver on the campaign’s promise added an additional wrinkle.

Now that the campaign is live, Thompson, Alnylam’s Senior Manager, Digital Marketing – Global Rare Disease, sat down with Kinara co-editor Ahmed Elsayyad to assess the company’s efforts and discuss what’s next for life sciences marketers looking to add AI to their toolkits.

Elsayyad: I know everybody in every industry is thinking about how to use AI these days. How do you think about it at Alnylam?

Thompson: We know our principal rare disease audiences: Undiagnosed and diagnosed patients and the HCPs who can help them. But within those two principal audiences are many niches whose needs and consumer patterns we study. We look for patient subgroups and HCP subspecialties and we think about how best to reach them. Patient subgroups can include patients with common demographics or circumstances, like parents and caregivers of minor patients.

While we view AI as a new tool that can help get messages to our audiences in novel ways, we didn’t want to drive a generative AI tactic too soon just for the sake of jumping on the “shiny new toy” bandwagon. We try as much as possible to test and derive a market insight before we act. This helps control costs – both agency fees and the opportunity cost of reversing course on something which proves premature. If we have a valid business reason to employ a new tactic like GenAI, then we examine whether executing said tactic ultimately serves our patient and HCP communities.

Elsayyad: So how did the PH1 campaign originate?

Thompson: We had two key insights. First, we heard repeatedly from nephrologists that they had probably seen PH1 patients during their career, but at the time did not know to test them for PH1. Some chalked this up to the rarity of the disease and a lack of understanding what testing should be done when faced with a potential PH1 patient.

Second, we knew that people of Middle Eastern and North African descent were more likely to have PH1 than the general population. But we weren’t currently featuring any people from this group in any of our marketing materials. To address this gap quickly and efficiently, we used GenAI to create digital assets we needed to build the campaign and connect with audiences in an authentic way.

Elsayyad: What – or, I guess, who – did you instruct GenAI to create?

Thompson: We used a generative AI image generator and prompted it to create patient imagery from its stock photo library that would be representative of people of Middle Eastern and North African descent. We worked to situate those stock images in a context that made sense for a PH1 clinical interaction. We saved on out-of-pocket production costs because we relied heavily on existing software that we already had access to, and an updated license gave us unlimited iterations. We were thrilled we were able to create this compelling imagery in house without having to contract an external agency for a photoshoot.

Elsayyad: What did you do to address concerns about transparency and privacy?

Thompson: We labeled every image as “Created with AI. Not a real patient” and made that text conspicuous, plus we centered the entire campaign around this concept with the “Have You Seen PH1?” tagline. By sharing AI-generated imagery of patients – none of whom our HCP audience has actually seen, because they aren’t real people – we drove home just how easy it is to miss the signs of PH1.

Using AI in our marketing in a legally compliant and ethically sound way is nothing to hide from. We aren’t hoping that no one notices that these aren’t actual humans; we choose to state plainly that they are not.

Elsayyad: How did your MLR colleagues help shape this campaign? Are there lessons you’d share for others taking GenAI to a MLR review for the first time?

Thompson: Before we even reached the MLR stage of review for these social media ads, we had to present the concept to the newly appointed AI board at Alnylam, which consists of employees in IT, legal, compliance, regulatory, medical, patient advocacy and more. Once we told them how we planned to utilize the imagery we created, the board enthusiastically approved it. The ensuing MLR review of the ads was fairly straightforward. They realized that using GenAI to create these images is not too dissimilar from using stock photography to represent real patients.

Elsayyad: How do you decide whether to use GenAI or actual people – patients or HCPs – in your storytelling?

Thompson: Real patients are uniquely compelling. They have real stories, real experiences, real results from therapies. They can be photographed or filmed repeatedly as their healthcare journey progresses. They can answer questions. They can speak to other patients and HCPs in person.

There are ethical implications as well. Take PH1: It can occur in people of all ages, including young children. By generating imagery of children on hospital beds, we exemplify a typical setting where HCPs might see these children – creating a more effective campaign and lasting impression in the HCP’s mind – without a child needing to participate in a full-day photo shoot with their parents’ consent.  Additionally, we know people of certain backgrounds may not like to have their photograph taken due to cultural or religious beliefs.

With GenAI’s help, we can be much more sensitive and empathetic and capture these moments of vulnerability in an ethical way. And since these are not actual people, we don’t need permission to use their likenesses.

Elsayyad: How do you see AI changing the future of the brand/agency relationship in rare disease?

Thompson: With the ease of use and value created by this new generation of generative AI tools, many brands may consider bringing creative development in house. Smaller creative projects – social media posts, banner ads, CRM emails – can be designed by in-house marketing teams, while larger campaigns might be handled by a brand’s agency of record.

I would, however, strongly recommend a brand sticks with its AOR for distributing, managing, and reporting on performance of content through digital channels. I’ve heard it said: “If content is king, then channels are queen.” Creating educational, engaging, and sometimes even entertaining content is vital to connecting with our audiences, but it’s useless if we can’t get the content in front of their eyes.

Elsayyad: So what’s next?

Thompson: GenAI video is the next frontier, with plenty of compelling tools coming to the market for commercial use. Will we be using text-to-video GenAI at this time next year? Almost certainly.