Most pharma organisations have already bought the technology. They have the authoring platforms, the DAM, the review systems, and now a queue of AI pilots. And yet the numbers that matter — content reuse, MLR cycle time, time to market — have barely moved.
In this session, Jindrich Borovsky, who spent more than ten years rebuilding MSD’s global content ecosystem before joining Shaman to lead its Content Excellence Practice, argues that this is not a tooling gap. It is an operating-model gap: fragmented ownership, no end-to-end mandate, and change management that never got funded. Maurice van Leeuwen, CEO at Shaman, puts the questions.
What you’ll take away
- Nobody owns the content supply chain end to end. Commercial creates, medical affairs reviews, IT selects the technology, local teams adapt — and the budget is split across all of them, so no single leader owns the outcome.
- Technology is the easy part. The barriers are executive mandate, change management and stability. Without a mandate, adoption stays optional and the platform goes unused.
- Content reuse keeps being built as a software feature instead of a governance change. Tagging, taxonomy and search matter, but so does whose KPI reuse actually is — today it usually sits with the global team, not the local one that has to do it.
- MLR delays are an upstream quality problem. Buying faster review technology means non-compliant content gets rejected faster. Guardrails belong in the authoring tool, not at the approval gate.
- Modular content fails without traceability. If you cannot see where modules are used, how they are adapted and how they perform, modularity becomes administrative overhead rather than leverage.
- AI is the last step, not the first. Map the value chain, fix governance and standardise the inputs — then automate. Otherwise you get a faster version of the old process.
Chapters
Why the investment isn’t showing up in the numbers
Jindrich’s diagnosis starts with ownership. Pharma has optimised the content supply chain in pieces — a better review platform here, a tagging project there — while the chain itself belongs to nobody. Commercial teams create, medical affairs manages review, IT chooses the technology, and local markets adapt. Budgets and decisions are split across those silos, and, as he puts it, most senior leaders don’t fully understand the end-to-end chain in the first place.
That is why the reported numbers deserve scepticism. Content-utilisation figures are easiest to measure in CLM, hardest everywhere else, and they exclude the large volume of content that entered the workflow and stalled before final approval — content that was produced, paid for, and never used.
Technology itself never drives the change or transformation. It’s simply an enabler.
Jindrich Borovsky · 9:36
What actually blocks adoption
Three barriers come up repeatedly in the session. The first is the missing executive mandate: without leadership making a global standard non-optional, local markets keep going to local vendors and the central platform goes unused. The second is consultant-driven hype around new tooling, adopted without a view of the end-to-end chain, so technology lands in silos rather than in how people actually work. The third is instability — when the toolset changes too often, people disengage and stop using it at all.
The counterweight is change management, and it is chronically underfunded. Jindrich recalls an industry rule of thumb he is careful to flag as anecdotal: for every dollar spent on technology, five to seven should go to change management, capability building and process redesign.
Reuse, MLR and modular content: the same root cause
Each of the three classic bottlenecks turns out to be a governance problem wearing a technology costume.
Reuse. Discoverability is real — inconsistent manual tagging, no harmonised global taxonomy, market-by-market metadata deviation. But so is user behaviour, and so are incentives: neither agencies nor internal teams are rewarded for searching before they build. When finding an approved asset takes longer than making a new one, people make a new one.
MLR. Most efficiency work happens downstream, at the approval stage. The bottleneck is upstream: creators without embedded compliance guardrails put content into the pipeline that was never going to pass. Jindrich points to two practical fixes short of a full rebuild — ready-for-review checklists that let creators catch their own errors before submission, and systematically capturing rejection reasons so the same mistakes stop recurring.
Modular content. It works on paper and stalls in practice whenever the operating model and agency workflows stay unchanged. Without end-to-end traceability of where modules are used and how they perform, modularity adds complexity instead of removing it. Maurice adds a practical filter: decide deliberately where modular makes sense and where it does not, because the upfront cost only pays back if the reuse actually happens.
We create content reuse as a software feature, instead of changing the culture, the governance and the operating model.
Jindrich Borovsky · 15:43
What the audience said
Two live polls ran during the session. Asked to name the biggest barrier to scaling content excellence in their own organisation, the audience’s clear winner was global-to-local — which is, in effect, content reuse by another name. Asked about AI readiness for content, the most common answer was piloting, but the foundations aren’t there yet.
Where AI actually fits
The second poll result sets up the closing argument: you cannot automate chaos. Before AI can produce compliant output at scale, the inputs have to be structured — a unified global taxonomy, consistent tagging and metadata, consolidated channel requirements and SOPs, with as little brand and market deviation as possible, in a form a model can actually read. Without that, AI simply creates content faster and moves the traffic jam to MLR.
What changes for people is the role, not the headcount. Jindrich describes the shift from content builders to content directors, with two responsibilities: building and programming the foundations — standards, design systems, taxonomies — and then acting as the quality filter that validates AI output for accuracy, medical compliance and brand alignment before anything reaches review.
Treat AI as the last step, not the first. Otherwise you’re just using new tools to create a faster version of your old programme.
Jindrich Borovsky · 36:57
Maurice closes on the same structure Shaman uses with customers: people, process and technology as three buckets, where the value sits in the connections between them. Invest heavily in one and ignore the others and the total outcome disappoints — which is a reasonable description of the last decade of content technology in pharma.
Go deeper
This session builds on our written interview with Jindrich, Why Pharma’s Content Operations Problem Is Bigger Than Technology, which covers the same ground in more depth and includes three short self-assessment scans — for your MLR bottleneck, your content reuse, and your AI readiness.
If you want to put numbers on your own operating model, the Content Production ROI Calculator models the time and cost difference between your current production model and a self-service approach.
Full transcript
Auto-transcribed and lightly edited for readability. Timestamps link into the recording. Speakers: Maurice van Leeuwen (CEO, Shaman) and Jindrich Borovsky (Director, Content Excellence, Shaman).
Welcome: why this conversation 0:00
Maurice
Welcome to our webinar on scaling content operations in life sciences. We’re organizing this as Shaman. Shaman is a content production partner in life sciences. We deliver a technology platform and offer strategies to scale content operations in life sciences.
And in this QA session, I will speak with my new colleague, Jindrich Borovsky, who is our Director of Content Excellence, and we will speak about real-world scaling challenges. And depending on your role, I’m sure that you have experienced some of these as well. Content is and always has been very important in pharma marketing, and today we see obviously for some years increasing content production, content numbers, and also with our customers and beyond, we see a growing investment in content excellence. Focusing on process simplification, dissemination, maybe, fewer vendors, content reuse, streamline MLR, and using AI.
And today we wanna discuss on why technology is important, but definitely not enough to make this happen, and what else is needed to fix it. Some time ago, I had the pleasure of interviewing Jindrich, and I will now and then refer back to that interview. We will also share the link. You can see it online.
My name, by the way, Maurice van Leeuwen. I’m the CEO at Shaman. We are over ten years partner in content production for pharma. We work for over twenty-five pharma, animal health, and med tech organizations.
Jindrich, could I ask you to maybe tell a little bit about your background?
Jindrich’s background: ten years rebuilding MSD’s content ecosystem 1:57
Jindrich
Absolutely happy to do that. So hi, everyone. JB here. I’m still relatively new to Shaman.
I joined them a few months ago, and what I do here is that I lead our content excellence practice. And what we believe that is important, or we believe that the real content transformation in pharma is mostly about the operating model and what we actually do. So what we do ultimately is that we help pharmaceutical companies to transition from these slow and fragmented content creation to a streamlined, scalable content operating model. And rather than layering AI onto broken workflows, what we do is that our approach, in our approach, we map the end-to-end value chain in order to resolve the core friction between content production or production costs, MLR review delays, and local execution.
We do it by establishing these clear governance and and modular processes before adding the AI, as I mentioned already, or automating the-- or deploying the automation. So with that, we really enable the teams to deliver the content which is personalized, compliant, and highly engaging. So instead of delivering it in weeks, they can deliver it in days. While at the same time, we’re cutting the production spend by up to seventy or eighty percent.
So before joining, before joining Shaman, I spent slightly over ten years at MSD, where I’ve been responsible for building, modernizing, and innovating their global content ecosystem and operating model. I managed the full content lifecycle across strategy, creation, MLR review, publishing, analytics. So that’s myself. Nice to meet you all.
Back to you, Maurice.
The technology paradox: why investment only moves the needle so far 4:03
Maurice
Now, Jindrich, that’s great. And, I think that is a great basis coming from where you, where you’ve been and from our side, servicing all of these different customers to see a lot of elements in the process and how the implementation is in different markets. Let’s start with, with the what you call the paradox of technology. So from your interview, you mentioned that every pharma company, they’re working to improve excellence, and they use technology for that.
But how does that progress, and what does that actually visually deliver?
Jindrich
It does, but only partial. I mean, only partial improvements, when it comes to the content excellence. And the main reason, at least how I see it, is that the organizations only see those partial results because they only address the specific steps or stages rather than looking at the content supply chain holistically for from the start to finish. And right now, also the ownership, if you look around, is completely fragmented.
So you have commercial teams creating the content, medical affairs who manages the review. You have IT who’s responsible for selecting, making the choices when it comes to the technology. And then you get the local teams that are handling the adaptation. And because the budgets and the decisions are split across these silos, that there’s no single leader who owns the entire process.
And frankly speaking, most senior leaders in pharma don’t fully understand the end-to-end content chain. So all this, we saw this happen in past already. And with just an example of modular content where they optimize the individual steps, but they have, they haven’t fixed really the underlying operating model. So hence the operating model, again, being very important obviously in this setup.
And actually what we are seeing is that company that they’re actually making the exactly the same mistake with the AI, as they did in past maybe with modular content and so on and so forth. And I was just thinking about the how to maybe. Prove these, my point of view here. I believe that you can clearly see that disconnect in the numbers.
So if you look at the key KPIs and those key metrics like content reuse or MLR cycle time or time to market. So you can see that they’re barely improved. So that they’re pretty much still the same. And I’m looking at this slide right now with this impressive number.
I’ve seen it multiple times. I’ve heard about it. I was just thinking about the reasons, and whether this applies to entire content world. It focuses on the CLM content.
So in that era, I believe that it’s a little bit easier to track the usage, utilization of the content. It’s not the case for the other channels. It gets more and more complicated unless you have, like, super-duper digital publishing in place. So when usually it’s very hard to track, and in most of the times, this number, it’s really just an estimate.
And it doesn’t count. I know that there’s a lot of content that sits in the flow that hasn’t reached the, that final stage being full, fully approved. So it got stuck somewhere in the flow, and it’s quite significant amount of the content that’s not counted here at all. So it’s the content that reached AFP, but it gets stuck there, it sits there, and nobody’s progressing it farther.
In case of the, in case of the reasons what I’m thinking is that definitely discoverability, and I believe that’s one of the topics that we’ll uncover or discuss more later, so that will be one. And speaking of the sales and marketing, there’s this quite natural friction between the teams, and I can. I believe that that also contributes, like the disconnect between sales and marketing to this number as well. So that’s my perspective on this.
“Technology is the easy part”: the real barriers 8:16
Maurice
It’s interesting. So you mentioned things like discoverability, a number of days to review or validate content, so these KPIs. And it looks like technology could actually help solve this. I mean, we could index content better.
We could have a semantic search as we do have in our platform. So you would actually type, and it would definitely be, like from a technical perspective at least, easy to kind of solve that. Now in your same interview, you mentioned that tech is actually the easy part. That’s a pretty bold claim.
Not only if you, if you go and work with Shaman, but also if you, serve an industry, if you talk about an industry that in fact does spend on tech. So what are these organizations getting wrong in your mind?
Jindrich
I remember making that statement in the interview, and I definitely stand by that completely. So for me, technology is the easy part. The hard part is getting the operating model Because technology itself, and that’s what I’ve been witnessing and observing for a very long time, even before joining MSD. So it’s.
Technology itself, again, it never drives the change or transformation. For it’s simply just an enabler. I believe, that the real barriers are more like organizational or culture or operational. Just to give you a couple of examples, again, based on my observation, my experience.
So as a first one, I would call out the lack of executive sponsorship and mandate. So because without strong leadership driving a clear mandate, adoption stays optional. So people just keep doing things the old way and the technology goes unused. And in my previous roles back in Microsoft, for example, at MSD as well.
So what’s, what we’ve been doing there is that we’ve been creating those, the, this core digital backbone. So with the hope that we will create these global standard solutions that the local markets will be using instead of, like, going to their local vendors for a local solutions and stuff like that. But again, if there’s no clear mandate for, from the, coming from the higher-ups or from the leadership, then it’s very hard to get the meaningful adoption and realize the ROI of that investment.
Maurice
It’s about change.
Jindrich
Yeah, it’s about the change. A change will get to it. That’s definitely number one for me. But I just wanted to mention a few others.
So another thing that I’ve been experiencing myself as well is that many companies, many of us get distracted by this consultant-driven hype. So that they create around the latest shiny new tools that everyone wants to get. So that they’re creating this hype without, however, that they’re, they’re not looking at it, again, holistically. So they don’t have this holistic view of the end-to-end content supply chain in their mind when they’re, when they’re presenting or showing these new technologies.
And that’s obviously related to the fact that sometimes we’re missing this end-to-end vision. So technology is deployed in silos rather than integrated seamlessly into how people actually work daily. And as you alluded to already, for me, the most important is the poor change management and lack of stability. So because driving real adoption, that requires stability and long-term commitment.
You can’t just expect that you purchase something, you threw it there to the field, and you let them to figure out how to make it work, figure out, set up all the processes, and simply realize that ROI for, from the in-investment. And following on that comment about those shiny tools, new shiny tools. So when technology changes too quickly, and that’s been happening quite frequently. So then people get obviously tired and they give up or they stop using it.
And I remember that once, one very clever guy told me that there’s. There’s this. Allegedly there’s this industry benchmark that said some-something like that for every invested dollar on the technology, the companies should be investing, I don’t know, between five to seven dollars in change management, capability building, and process redesign to actually, again, realize its value.
Poll: the biggest barrier to scaling content excellence 13:02
Maurice
Maybe it’s interesting to use a poll and to also get some feedback from the attendees. What would you say in your mind, in your organization, what is currently the biggest barrier for scaling content excellence?
Jindrich
It’s a good and tough question. I already have a candidate, so I’ll
Maurice
Where’s
Jindrich
Once we, once we reveal
Maurice
Where’s the select all option. All of the above.
Jindrich
Where is that option? Yeah, I don’t see it, but yeah, let’s see. What do
Maurice
Cool. Can we, can we move to the results? Global to local. That’s the winner, yeah.
And I think actually that is a great example where technology is not the problem. It’s about this change behavior and I guess about this notion that if you have a budget for innovation, technology should be one part, yes, but probably not the biggest part. You should. It should a total solution, and you probably should spend more time and money on adoption and getting all stakeholders on board rather than just getting a tool and that’s gonna be the solution, and miraculously everybody’s going to use it.
Jindrich
And if I’m able. Actually, if I’m able to vote to answer that question. So for me, it would probably be the governance and the operating model. I believe that there’s been few.
There’s been some percentage, some votes for it as well. ’Cause for it really. It’s really that foundation or root system. ’Cause, ’cause all the remaining options that I’ve been being s-- that I’ve been seeing there, they really function as, downstream symptoms or direct dependencies of that core operating mode.
Content reuse: why platforms and tagging haven’t fixed it 14:57
Maurice
Cool. You know, thinking about this global to local, what was mentioned, in a way it is content reuse. It’s the fact that create once and reuse everywhere. And it sounds that it makes a lot of sense because we’re talking about global brands that have the same medical background, that should have, to a very large extent, the same medical benefits.
So why do you think that after a lot of years in platforms and tagging, in tools, are we still talking about content reuse as a, as an issue?
Jindrich
And still not getting there, where we’d like. So listen, for me, it’s, it’s really simple. So the reason for me is that we create content reuse as a, as a software feature. Again, instead of changing the culture, governance and the operating model.
So content reuse, it’s a multi-layered issue. It’s the same old issues that we’ve been dealing with and talking about for some time already. So I mentioned it already. It’s the.
It’s one of the reasons, it’s the discoverability. So meaning that content is hard to find, and we’ve been constantly hearing from the users that search is broken, the technology doesn’t work. But for me, it’s not just about the search technology. ’Cause it’s also about user behavior ’cause when you look at how users conduct the search.
Well, what keywords they’re entering into the search bars. So it’s not. Definitely not only the technology to blame here. So besides the discoverability, obviously content tagging.
So we’ we know that if it’s done manually, it’s not always the best values that we get there. So without really standardized tagging, these global tagging taxonomies and clear metadata. Without as little as pos- nuances between the local markets. So ideally fully harmonized set of values and metadata fields.
And these intuitive search tools. So if we don’t have these elements right, then finding an existing asset, it actually takes longer than creating or making a new one.
Maurice
Yeah, and to your. Sorry. And to your point, I think that some people may say, "I cannot find it," but are they actually searching. So i-is it maybe.
Is it the real reason? So I think if the operating model is. It’s not there, then you can sub-optimize all smaller parts to make it better. But first we need to make sure that people actually go in and look for stuff.
Jindrich
And then that brings me back to my comment about the mandate, which is missing here as well. There’s no real incentives, neither for the agencies or the internal teams to go and search for and recycle, reuse the existing content. So that’ very important part of that equation.
Maurice
No, exactly. And I think that this thinking pattern of saying, "Okay, if everything is there and it will be easy to search, then automatically will people will use it. " But I think that’s the mistake in that thinking. Very often the percentage of content reuse is a KPI for a global team, but it should be a KPI for a local team.
To your point. So anyway, let’s, let’s look at the next topic. There’s so many-- so much to discuss here, which is, which is MLR. Often, a long MLR timelines, they get, they get blamed on team capacity.
But again, going back to the interview, you mentioned that that’s, that’s a symptom rather than a cause. Can you maybe explain what you, what you meant by that?
MLR: treating the symptom instead of the cause 18:13
Jindrich
Yeah, absolutely happy to do that. So, most pharma organizations, they focus their MLR efforts downstream. So what they do is that they’re buying all these new platforms, technologies. They’re throwing in AI at-- however they throw it or they apply it onto the final approval stage, to clear that traffic jam or the bottleneck, that is there.
However, as you, as you said correctly, it’s like they’re not treating the symptoms. The-- they’re, they’re treating the symptoms, they’re not in-instead of the cause. So for me, the core issue is what happens upstream, at the creation stage. That’s where I see the real bottleneck with that upstream quality gap.
And what that means is that the creators in. At that stage, they don’t really have embedded compliance guardrails. So what happens is that the bad content is being created, and it enters the pipeline. And with the, with all that, automation technology and AI solutions that they’re implementing, they just gets rejected faster, all these bad inputs, or submissions into the MLR.
Ideally, what should happen is that these brand local regulatory teams and all the other SOPs requirements that w- that we have around the, for these final tactics, final assets, those should be ideally embedded directly into their tools. So ultimately, they’re creating compliant content from the very start. Or as we say now, it’s, it’s this new word that’s being used, creating compliant content by design. So that’s.
In terms of the, in terms of the other fixes that can be implemented here in the later stage, like more closer towards the MLR itself. So it’s the. And what I’ve been seeing, in the industry, in this space is that the. Some companies, they’re creating and using these ready-for-review checklists.
So which help them or which help their content creators to catch some of these mistakes before they even submit the work for the MLR. There are companies that are, like, gathering the. And collecting the reasons why the job got rejected. So it can serve as an input to improve whatever they’ve identified that’s constantly being done and delivered wrong,?
Maurice
If you make sure. Because otherwise, it’s gonna all pile up at the end. And if you are aware of these guidelines and you can apply them earlier, then obviously the whole. It will streamline the whole process.
So also, I think from your interview, we got some questions that you, that you delivered to kind of. You can ask yourself to get a little bit more idea. Indeed, is my MLR bottleneck, is it actually an MLR problem, or is it a content production problem? So I think here is another one which is about content reuse.
We talked about that as well. All right, so I wanna touch on a couple more topics, and I see it’s already. We are already thirty minutes in, so I’ll try to. Let’s maybe try to keep it a little bit more condensed to make sure that we cover all the topics.
Modular content, obviously. So it has been a topic for a long time. It’s going, sometimes up and down a little bit, but obviously the concept is promising and makes a lot of sense. However, a real huge implementation in production in pharma is still rare.
So why do you think that’s the case?
Modular content: why large-scale implementations stay rare 22:48
Jindrich
First of all, I don’t want to. I don’t want to keep repeating myself, but a lack of mandate, is part of the problem here as well. So modular content, it’s that strategy that sounds great on paper. But it constantly fails if you don’t completely change your operating model and the agency workflows.
So if you don’t redesign how teams actually author and approve content, then modularity will only create extra work and unnecessary complexity here. So it also breaks down even faster if you cannot track the process for. From the start to finish, without end-to-end traceability, meaning the ability really to track where modules are used, how they’re adapted, and how they actually perform in the market. So without really rebuilding the workflow for.
From the ground up and tracking module performance, what sounded like a streamlined strategy quickly turns into an expensive administration nightmare.
Maurice
I think part of the idea is that you do more work upfront, and then because of the reuse, you’re gonna save later. But if you do not save more later, then the total time is gonna be more. So a really important question is, are we actually going to reuse this? Do we have the channels where we’re going to reuse this?
So to kind of say that modular content may not be the solution for everything, and so having this framework upfront where you say, "Okay, where does it make sense and where does it not make sense? " is, apart from the whole operating model and the governance that I agree on, is probably something you need to. Purposefully, because it’s not, it’s not a one-size-fits-all. It’s not a solution that will help you everywhere.
Fantastic. So let’s turn to the inevitable AI, obviously. AI being the biggest investment probably in a lot of areas, including content production, MLR review. There’s a lot of talk about it right now.
There’s a lot of projects, pilot projects. However, taking it from a shiny pilot where you did a few test cases to a real production kind of environment is a different question. So maybe you can talk a little bit to this. Oh, before we do this, and before you talk about what it needs to take for a company to take a shiny pilot to a production, maybe we do a poll first, because we have a poll exactly on this topic, which is AI readiness.
So how would you describe the AI readiness in your organization today? And then specifically for content. So AI content readiness, which means content production. It can mean tagging, searching, discoverability, MLR checks.
So everything that happens in the content production space. All right. Curious to see what, where we are with this audience today. So this is great.
And it may mean that AI readiness is not that bad. Doing pilots, having solid foundations, that’s great. But to talk a little bit more about readiness, so what can you. You know, what do you see in this area?
AI: from pilot to production, and the readiness poll 24:45
You can’t automate chaos: the foundations AI needs first 26:22
Jindrich
Well, I’ll just build on the result that we’ve just seen. So piloting, but foundations not there yet. So that’s exactly what needs to be fixed first. ’Cause you can’t really automate the chaos.
So you can’t layer or put the AI on top of the broken processes that may be there already today. So it’s the foundations first, which means that. It’s the standardization. Which is non-negotiable.
So that means clean, standardized inputs like unified global taxonomy, consistent tagging and metadata. You need to consolidate the channel requirements, all the other requirements, SOPs, or content standards. And ideally, without this brand or market deviation. So the more consolidated, the more standardized, the better results you’ll get from the AI, unfortunately.
So if the inputs aren’t structured here, AI can’t really generate meaningful and compliant outputs. So that’s, that’s number one. Strong metadata models, clear taxonomies, standardized workflows. If you don’t have those, the same will happen as it.
As we’ve already mentioned. So it’ll move content faster. It will be creating content faster, but it will be stacking and creating this huge bottleneck at the MLR phase. So I believe that the opportunity is significant and high here, for the AI.
But before you jump onto the train, you really need to. Besides fixing and building those foundations, you really need to look at your end-to-end, the entire content value chain holistically and completely redesign it for, from this AI perspective.
Maurice
That makes sense. And it almost looks like vision and strategy is more like C level, and then ultimately it comes down to, the work floor. And I think we also see this. There’s a lot of ambition, but ultimately if these processes and structures are not in place, it’s gonna be very difficult to execute on it.
So this is coming from a Deloitte research, interviews with 500 people. I added this little gray bar, which is not inside this interview, but it is something that we are working on in Shaman. We call it Shaman Atlas. And what we here try to do is to have the foundation for AI, amongst others, around the brand.
So having a very structured knowledge graph, knowledge engine that contains all information about the brand, the guidelines, the strategy, the personas, and then moving to medical, which is claims, which is previously approved content that can be reused, up to gathering regulatory authority information, SmPC and PI information. And we strongly believe that when you have this well-structured in place and you then can give AI access to it will definitely help your readiness in terms of creating content that is compliant and will, survive your MLR review.
Jindrich
And that’s really strong foundations for the AI here.
What a mature content operating model looks like 29:49
Maurice
So another checklist that I think is very interesting to look at, so please visit our website and check it out. And then at Shaman, we also talk a lot with our customers about the maturity of our customers’ content supply chain. And so from your perspective and your experience, can you talk a little bit to what that maturity is about? What does a mature model look like?
And then maybe as a sub-question, what impact will, amongst others, but predominantly AI have on the future of a mature content supply model?
Jindrich
So talking about the mature operating model and how I see it. A strong content operating model as an end-to-end system that connects all those stages, phases of the content life cycle, like strategic creation, MLR, localization, distribution, and the measurement. So the key shift here is the key shift is for from, like, the producing individual assets, like the, like those final finished tactics to treating content really as reusable governed components that can be adapted across channels, markets. And the operating model should bring together also the right people, process, agencies, and technology to make content easier to find, to reuse it and scale it, while maintaining all the compliance.
So having all those guardrails implemented, embedded in the tools is definitely the best way to go. You know, ultimately, the goal is to create the right content for the right customer and channel as efficiently as possible and continuously learn and optimize based on the performance. So it should be a really the closed loop here. You
Maurice
Yeah, and I think it also alludes to your earlier point that if you make an analysis in your company, where are we in this operating model, and you analyze those buckets. You look at, people, what is the, what is the ownership? What is their adoption? What are incentives, et cetera?
Then you look at the process column, you look at the technology, and these are not aligned, or maybe one is really strong because you invest a lot there, but you don’t invest in people or processes. Then the total outcome is gonna is going to be disappointing. So I think this is a real nice model to think about. It’s also.
It makes a lot of sense, but to kind of analyze and say, "Okay, where are we and where should we actually invest to improve this? " So how do you see AI impact all of this?
From content builders to content directors 32:53
Jindrich
That’s the second part that I wanted to answer, the future with the AI. So it’s like the most visible change that’s happening right now. It’s this fundamental shift towards an AI augmenting operating model across the entire content life cycle. So it’s obvious, or at least we hope, that the AI isn’t replacing the human expertise here.
So instead of spending hours on the manual execution, they can focus on high-level strategy. So creating creative problem-solving and critical decision-making. So humans, they stay right in the loop. And as I mentioned in the interview is that they’re really moving for-- from being the content builders in this case to becoming the content directors.
So what it means in practice, it’s that it’s like the role of the humans, it divides or splits into two critical responsibilities. So first is there to really help establish and build those foundations, and set up and program the engine itself. Defining those foundations. This means, like, organizing those con- content standards, all the SOPs, establishing design systems.
It’s very important as well. Defining global marketing and tagging taxonomies and all the other stuff. All this, actually all these inputs, they need to be well-structured, and they needs to be, like, machine-readable or LLM-readable in that format, so the AI can, can read it and understand it. Second role of the humans in that future AI-driven or led operating model, it’s the validation and refinement.
So that they’re, they’re looking at the results, that they’re validating those. Ultimately, that they become the, this quality filter. So verifying the AI output for accuracy, maybe also for the medical compliance, brand alignment, before anything goes to the MLR or out to the market. So that’s how I see the future evolving here in the, in this space.
Maurice
What I like about this, to kind of visualize it like this is that when you think about how AI is going to improve the content production process, again, we think about technology. If you don’t have the people that know how to work with this AI, how to instruct this AI, how to evaluate the outcome of what AI proposes, if you don’t have the right processes in place, if you don’t have the data machine-readable, then the technology by itself will probably do more harm than it will do good. So I think this basic idea of we have these three buckets and the value is really in connecting them. And if you focus on one and you don’t focus on the other, it’s not gonna work.
So what I really like about this is that, yes, AI is gonna change a lot, but it’s not gonna change this idea of these buckets and the connection you need to manage and think about between them. So look, JB, thanks a lot for this. It’s always a big pleasure to talk with you and have these conversations. Today we’re a little bit on the clock.
I was wondering, is there any questions from the attendees that we can maybe take as a, as a closing,? Is there something. Yes, I’ll share a question with you. I just got it here.
All right. So the question is, JB, is for you. If you could. So the question is, if you could change one thing that pharma organizations are doing today that’s holding content excellence back, what would it be?
The one thing to change: treat AI as the last step 36:55
Jindrich
Oh, that’s a, that’s a great question. So I would say it would be all about the AI at the moment, probably. So I would say that treat AI as the last step, not the first. So start by mapping your entire content value chain to find out where delays and high costs are actually coming from before buying new tools.
And use those insights to fix your governance and operating model first so you know who owns the process end-to-end and how MLR reviews will work. So only then you should start layering on the technology or AIs. Otherwise, you’re just, you’re just using new tools to create faster version of your old program. So that would my recommendation.
Maurice
Fantastic. Really clear, and I think there’s a lot of truth into that. So thank you so much, everybody, for joining today. JB, thank you so much.
Always a pleasure, and I hope you enjoy the rest of the day.
Jindrich
Thank you. Thanks for having me. Bye, everyone.
Maurice
Bye.
Jindrich has nearly 20 years of experience across enterprise platforms and content operations, including 10+ years at MSD (Merck), where he led a global team of 40+ IT professionals and drove global rollouts of Veeva Vault PromoMats — and before that, Microsoft’s Global Marketing Operations. He leads Shaman’s Content Excellence Practice.
