AI‑Powered Social Media: Creating at Scale Without Breaking Compliance in Pharma and Other Regulated Healthcare Settings

Stewart Gandolf (Healthcare Success): Hello and welcome to the healthcare success podcast. Welcome back if you're a loyal listener and welcome to the first time if it's your first time here. Today we're going to be interviewing Nikki Wolfert who is strategic partnership lead at Marketbeam and Pushba Ithal as the CEO of Marketbeam. Welcome to both of you.

Nikki Wolfert (MarketBeam): Yeah, thank you.

Pushpa Ithal (MarketBeam): Thanks for having us.

Stewart Gandolf (Healthcare Success): Very good. So Pushpa, why don't we start off by talking a little bit about what's unique about marketing. We're obviously going to talk in a few minutes much more about what's going on in AI, social media, and the pharma world. But give us a little sense of what your company does, what you guys do, so that the audience has an idea of what it is we're talking about today.

Pushpa Ithal (MarketBeam): MarketBeam is an end-to-end social media management platform specifically built for the health industry, keeping compliance, regulation, all those things in mind. But it's not only just sprinkling, it's embedded from step one to all the way to auditing and then how even the monitoring is handled within the pharma industry.

So of course now, know, AI is making it much more robust, faster, and then also like, you know, putting the guardrails is very important in this AI world.

Stewart Gandolf (Healthcare Success): Very good. So let's talk about when you start at the beginning here. And obviously on this podcast, we talk about AI a lot, but for people that aren't as familiar, what are the main types of AI that exist today?

Pushpa Ithal (MarketBeam): So generative AI is the most common one that we are all familiar with, right? So whichever industry that we work on, and then even like, know, personally, we started using generative AI for the past probably even like, you know, two to three years now, where we create content, ideation, content in terms of like, even like text and media and all those things. So the other types are, you know, predictive AI is another type of AI, is again, it kind of like existed in another form in the software world, right?

So what we learn, what are the results, and then how do we build that whole feedback loop? And then now it's at 100x speed and then also accuracy. That is the predictive AI that we are building as an industry.

So another one is the agentic AI. We keep hearing about agents and then like, you know, many people have that notion of, what is an agent? Is this a task base or is it end to end? The way that I would define this is agents are little oompa loompas, right? So each one of them would have their own task. But at the end of the day, they all have to work together to complete the whole workflow.

So in many cases, we do think that it's all going to work smoothly, right? But there are still so many gaps that I think we're going to probably like deeper dive into those things is where it's going to kind of like block or like where are the gaps and where it's going to fail.

That's agentic AI. Just one more layer added to that is industry-trained model. So industry-trained models are super critical as industries mature as software matures and then we do need that. So these are the four types that I would define the AI types as.

Stewart Gandolf (Healthcare Success): Very good. So, Nikki, how do you see generative and agentic AI fundamentally changing the role of social media? Because today our audience, by the way, includes pharma, includes multi-location businesses, private equity, memory care, know, a whole bunch of different businesses. But not everybody on our podcast is marketing driven, but a lot are. And certainly social media is an area that everybody's interested in continuously. I'd love to hear a little bit about social media, how it's impacting social media teams, and both paid and organic, and just help us talk about the big picture here.

Nikki Wolfert (MarketBeam): Yeah, absolutely. So really the role of social media and how social media teams specifically in pharma, it's just constantly evolving, right? On any given day, a social media person might be educating an internal stakeholder on where kind of social fits within that broader patient or HCP journey. Another day, you know, they're advocating to legal or regulatory about why a certain trend matters or how to participate in it compliantly.

And so when a new platform emerges, suddenly you're becoming the student, right? TikTok was once new to everybody. Facebook, how to use Facebook in Pharma was once new. And so that adaptability has always really kind of defined Pharma social media teams, right? That constant evolution.

AI doesn't change that. It really is just kind of the other tool in that toolbox. So I like to look at it as like generative AI, as Pushpa mentioned, it can absolutely kind of help reduce, you know, the mundane operational things like of course, drafting baseline copy, summarizing medical papers, pulling key claims, etc.

And that's obviously really, really helpful. You know, because if AI can really take on more of that repetitive day to day tasks, your social media teams can really spend more of their time where they're most valuable, right? And that's where being strategic, advising cross-functional partners, teams. And then the key thing here, I think is it's about infusing that real human perspective, right? In pharma and med tech, that human layer is absolutely everything.

So while, know, agentic AI, you know, might be a really helpful know, creative starting point for content. I think it's really important that, you know, it's considered very foundational and that that human element is really like building off of that.

Stewart Gandolf (Healthcare Success): Totally makes sense. And it's funny, when we work with life sciences, you mentioned something there a moment ago, is just consolidating all the different sources of data and studies and pulling together what's the messaging that's appropriate for social media translating that is really intriguing. So this is a topic we speak about in our agency a lot with the balancing AI between the sort of the quality of it, the quantity, and then the other part is speed. So like, Nikki, any comments on that?

Nikki Wolfert (MarketBeam): Yeah, yeah, I have a lot of thoughts on this here. You know, I think that, like I mentioned, AI has that potential to improve certainly the quantity, right? How much content you're making. I do think that it has the potential to improve the quality. And that's always been this trade-off, right? If you're gonna scale content well, the quality is not gonna be great. Or if you have really good quality, it's gonna cost a lot. So you might not get as much out of that.

And so it's always kind of this trade-off, this sacrifice, but AI can help protect quality within the pharma space. So I know I don't wanna steal too much of Pushpa's thunder here, but AI has thrown around a lot, especially within pharma right now, it is that hot topic.

And from the quality perspective, where I see it fitting in most is not to replace all of our brilliant copywriter friends that we work with, but it's really helpful in things like helping make sure that your content is going to pass MLR muster, right?

Having those automated pre-checks, certainly reference validation, annotation support, et cetera. So I think those things alone really kind of help the efficiency, certainly the end point quality.

I always say at minimum social content generally takes 12 weeks to get reviewed. And then to the point that it's approved and ready for use. And I think with AI, it's going to allow those teams to kind of expedite those workflows so that they can actually create scalable social content, but keep everything really, you know, quality at the same time.

The last thing I'll mention, because like I said, I have a lot of thoughts on here, is how we can use social media to create better strategy and how we can use AI to sort through all of the incredible social content that's out there. And that's really by having social listening be incorporated and at the start of your strategic journey, essentially.

And there's just, there's a lot of really great information out there. Typically it's been really tedious for us to kind of look at and distill and use. And so there's an element there too, where I think AI is really helpful at, know, flagging those mentions, looking at holistic trends, identifying your audiences, et cetera. And that's another thing that I think is really helpful and will, you know, continue to expand how helpful it is to pharma and making better content at scale.

Stewart Gandolf (Healthcare Success): Very good. That's a lot there for sure. And it's funny, well, we work with life sciences, both pharma and device and other kinds of life sciences companies. And obviously compliance is number one, two and three. Like everybody wants to know if you've ever worked with pharma. It's like, yes, we've worked with pharma.

 So it's interesting too. I'm not surprised that in pharma it's pretty predictable actually that anything that's new is scary, right? So we're talking about sudden AI and my gosh, what are we going to do? But if the checks and balances are there, I'm assuming it's going to be like the way you generate may be different, but you still need to go through compliance in similar ways.

So Pushpa, I'd love you to expand upon that. how you, because I understand your tool works for other industries that are regulated as well, but what's unique in pharma now from your perspective?

Pushpa Ithal (MarketBeam): You think in pharma, in again, like when we're talking about compliance regulations, those things are not changing. Right. So that is still the same. So what we're doing right now is creating content at scale. you know, so even since like, you know, 2019, 2018, every year we are increasing content by 20%, at least 20 % continuously. And then now this is going to like, you know, even explode even more.

So what is happening is what we're not paying attention to is like all this content still needs MLR review. And then is that creating a choke point? Right?

So we are hitting a manual wall at that point. A lot of it, it needs to be still reviewed by human beings. Of course they have software to review it, but it's still like, you know, we call software almost like manual at this point, right? So we'll pass that stage of software is automation.

So what we see here is like, over 50 % of the content, even historically, just sat there and then never made it to HCPs or patients because it just passed the shelf life. And then it was not relevant anymore to even like get to the audience because it just got stuck. It got created, but got stuck in the review process.

But now what we're creating is a very expensive waiting room. it's the rate at which MLR has actually improved over a period of time is 1.5x, while the content is 5x. So we know that we are creating a choke point, right? So unless we change that, any regulated industry, right? So even financial services, it's going to be exactly the same problem.

So just content creation is not going to solve the problem of like, okay, we are getting more content to the end user. That's not going to happen. So unless we take care of this compliance things and then automated more, at the end of the day, we still need regulators to take a look at it as of today, that might even change in like a year or two. So as of now, we still have that, but creating that is very, very important. How easy it is to get through the MLR process. all know, and like Nikki was mentioning, it takes about 12 weeks. Yeah, 30 to 40 days is very, very common. So I'll give you a quick example. So we were working with a pharma company who launched their new market in Eastern Europe and then Western Europe at the same time. And then there's a social media campaign that they launched.

And then they're like, you know, hey, you know what, I'll get back to you in a couple of weeks because we know all the issues once we launch in this new market. And then he got back saying that, Hey, why does it like, you know, just changing the button in my Facebook ad is taking me 40 days.

It's like, yes, that's the problem we're talking about. And then they just like really realize and say, okay, what can we do about it? So unless we solve that, just creating more content is not going to help this industry.

Stewart Gandolf (Healthcare Success): That totally makes sense. It's funny because if you haven't worked with, any of our listeners, haven't worked with AI and content, like our strong opinion on this because we use AI is like it's still a tool, right? You can put in brand guidelines, you can put in compliance guidelines, you can put in things to get you, but you still need human judgment. So it's not like for our company, we're not just, you know, doing less work for our clients, we're doing more, we're getting a lot more done for them for the same dollars.

But I can see in this case, you guys can put with AI, you can put theoretically unlimited amounts of content much, much faster. So compliance can't keep up. It's like, what's the point, right? Okay. I have to figure out how to pace that. So Nicky, pivoting over to the future on this space, what's exciting to you and what's concerning to you?

Pushpa Ithal (MarketBeam): Exactly.

Stewart Gandolof (Healthcare Success): So Nikki, pivoting over to the future now in this space, what’s exciting to you and what’s concerning to you?

Nikki Wolfert (MarketBeam): So, you know, if AI can help with the things like I mentioned, like automating pre-checks, reference validations, even monitoring comments and, you know, the scary thing, the potential, you know, adverse events or product quality complaints that are so prevalent on social.

My hope and what, you know, I'm most excited about is that it's going to give medical, legal and regulatory teams a lot more confidence. And that confidence is going to help, you know, marketers open the door to participating more fully on social so that I like to say it allows them to be social on social, right?

It's not just being used as a one-way push, but my hope is, is that the MLR teams can be, you know, feel confident that there's, you know, stop gaps in place where they can be compliant, but from the marketer perspective, and I think most importantly, from patients and HCP's expectations, they're going to get that kind two-way engagement that everybody goes to social media to seek in some way, or form. So that's what I'm the most excited about as it relates to pharma incorporating AI.

Stewart Gandolf (Healthcare Success): That makes a lot of sense. So Pushpa, let's talk a little about your company. What are you building at MarketBeam? What is the need that you're trying to fulfill?

Pushpa Ithal (MarketBeam): So the new, best again, like AI, right? So as we were talking about, how do we remove that choke point? Or at least like, make it go faster, right? So the choke point has to be like, bigger.

So domain-trained reasoning layer. So that's what we call it. So domain-trained reasoning layer is some layer that actually understands the requirements of this industry. So what we're bringing, the data sources that we are bringing together is, if for a piece of content, where the content is like, know, it gets more complex in terms of social media because it is text, it's the links and the hashtags and your videos and everything has to come together.

And then also, of course, younger and younger generations are making a lot of their decisions looking at social media content. So the responsibility of those content being really accurate and then also seeing the right things and then no confusion. So we're like, you know, the safety information and all those are super, super critical.

So what we're building is again, like, you know, to remove that a checking point. And then that starts from when you are creating the content, not wait till the waiting room that I was talking about is while you're building the content, you already have access to what's the right thing to say or not.

So this layer has the data sources from label information, it has disease state information, and then also like FDA, FTC guidelines. And then also mainly it is kind of like what I call the tribal knowledge that's sitting on people's desks, right?

So because many times it's the same kind of feedback that MLR gave multiple times because it's the different teams submitting the same mistakes again, which we hear very, very often. So that never makes it to automation. So we are pulling all those things through even like, you know, building our own groups of people who can provide that information to us.

And another new one that we are adding to our data source is warning letters. What previously five to 10 years of warning letters and then title letters to. What if they really asked these companies to change in their communication, going to the details of the safety information was in a link that was not visible inside your YouTube video? That is as specific as it can get. And then companies get warning letters off of it.

So bringing all those things to AI and then making it the first ever robust domain-trained layer. That's what we're building. that again, like, you know, this is going to reduce the number of back and forth with MLR. And the goal is to get the first pass.

Stewart Gandolf (Healthcare Success): Very good. That makes sense. What you were just saying there too, it's like the AI is only as good as you feed it, right? Otherwise it's just giving you what you tell it. But if you're feeding it data point after data point, you're going to get a better result. Nikki, if you were to give one piece of advice to form a medtech marketing leaders trying to incorporate AI effectively, what would that be?

Nikki Wolfert (MarketBeam): Yeah, I think my main piece of advice is to really look at AI, not so much as a threat, right, to your capabilities, but as an extension of your team or kind of that force multiplier. And I think especially within, within medtech and biopharma teams are typically pretty lean. The resources are a little bit more limited than typically within, you know, big pharma.

And so if some of that kind of those repeatable daily time-consuming tasks can be automated using AI, I think it's not a threat of your team's capabilities and potential. It's actually going to allow your team to kind of stay more focused on the strategy and driving that impact.

When AI is used and implemented in the right ways with the right compliance, like we've said kind of over and over again, it's like, it's about unlocking a level of scale that the industry has not really seen before, but while maintaining quality.

And of course, as Pushpa has talked about, you know, like eliminating a lot of bottlenecks too, because yes, you know, content at scale, yes, quality, but if not, your intended audience isn't seeing it, that's a problem too, right?

Like I said, think amplifying your team's capabilities is one piece, but there's another thing that we haven't touched on is that how AI can act not just as a force multiplier for your team, but in amplifying your team's voice. I think that's a thing that we haven't really seen commonly spoken about within pharma, but using AI as that way to kind of act as a virtual assistant, if you will, as it relates to how your team can use social media, it's so important to tee up your team, whether it's marketers, whether it's field reps, I know specifically, you know, as we're talking to a lot of, you know, med tech and hospital groups here. How can you find ways to use it to intelligently distribute the right kind of content, whether that's therapeutic areas, geographic regions, et cetera.

How can we use that in that way as well? And so that you're building credibility in your team's trusted voices, it's gonna make your content stand out while being efficient and yet kind of like more impactful as well.

Stewart Gandolf (Healthcare Success): All right, that makes sense. So Pushpa, we're wrapping up here. If a pharma company is trying to evaluate AI for social media, what are some of the questions they should ask vendors as they're thinking through trying to figure out how to get this done correctly?

Pushpa Ithal (MarketBeam): Don't lose your edge. Don't make it lukewarm, your content. for that, you just need the industry layer to understand that so that you can still have that edge. At the same time, you can still be compliant. For example, putting guardrails around just like creativity. We were looking at one of these examples where an image that has a child playing outside football, but they're talking about hemophilia. Okay, that doesn't make sense. because you know, a child playing outside is not, not, not something that you want to talk about the hemophilia.

So keeping those things together. you just need one is make sure that your AI understands your industry. The next is make sure that you have the guardrails. The AI has the guardrails and the term most important thing is accountability because once AI takes over accountability is something at risk too. Who actually created and then who produced, who published, and then nobody is accountable, and then make sure that everything is audited and then kept track of so that you know what went wrong if something went wrong.

Stewart Gandolf (Healthcare Success): Excellent. Well, this is a great discussion. I would just say from our experience, some of the things that we see as well is just the idea of you mentioned it a little bit earlier. think, Nikki, is the institutional knowledge base too, because that's the problem is things can become really siloed as you're trying to scale. So not only having the knowledge of the regulation, having the brand book and the compliance guidelines, but the institutional knowledge can get lost and making sure we're thinking about the voice, which Pushpa, you just mentioned.

So, Nikki, Pushpa, this was a great interview. Thank you guys for your insights about how to scale and how to effectively use AI in the world of pharmaceutical social media. There is a moving target for sure and compliance is going to go away, but maybe we can get our content better, faster, compliant more quickly without overwhelming the compliance team at the same time. So thank you guys. I appreciate your time.

Pushpa Ithal (MarketBeam): Thanks for having us.

Nikki Wolfert (MarketBeam): Yeah, thanks for having us.