Why graphs ways of thinking and working fail to land and what it really takes to show their value in the Life Science industry.
When the Message Doesn’t Land
Back in 2017, the demo went well. We had worked hard to articulate a clear and sound logic, supported by concrete value and numbers. We deliberately avoided a technology-first narrative. Instead, we focused on real use cases, connecting clinical, regulatory, and real-world data, breaking silos, enabling a more comprehensive understanding, reducing blind spots, and revealing previously invisible connections to generate new insights and value.
The room nodded, and then came the question that stopped me: “What you say is true… but how is it relevant?”
In that moment, my mind went blank. Time seemed to stretch. Looking back, it was more revealing than it first appeared. It did not signal disagreement, it revealed a gap: a gap between innovative ways of working and the reality of how organizations actually create value.
In pharmaceutical companies, as in most large organizations, the limiting factor is rarely technology. Extracting value from graph-based ways of working is not primarily a technical challenge. It is by order of importance a matter of 1) people, 2) processes, 3) data, with 4) technology playing, paradoxically, the smallest role.
Beyond Technology: Where Value Is Really Won or Lost
It is striking to realize that, in the Western world, the notion of “progress” is often intersubjectively reduced to technological improvement (see article about “Techno-Messianism“). There is a widely shared belief that better tools will naturally lead to better outcomes.
My experience with knowledge graphs in the pharmaceutical industry has led me to a different conclusion: organizations do not fail to innovate because they lack solutions, they fail because they cannot absorb them.
To understand how to demonstrate and unlock value, we must shift the focus away from the graph itself and toward the system in which it operates:
-
- People, who must trust and adopt it
- Processes, which must integrate it
- Data, which must make sense within it
Technology, in this equation, is simply an enabler, and often the least problematic component.
Real progress is driven by changes in mindset and culture, incremental improvements in processes, and the ability to define, disambiguate, reconcile, and manage data, metadata, information, and knowledge so that they make sense to human decision-makers.
This has been true so far. With the rise of language-model-based AI, it is possible that “making sense to humans” may no longer be a strict requirement, but we are not there yet.
1. People’s Organization: The Invisible System That Shapes Everything
Innovation does not fail in org charts, it fails in everyday behaviors.
At work, people operate within a set of implicit and explicit constraints that shape how they respond to challenges. These behaviors are rarely rational and often protective.
Comfortable routines provide stability in complex environments. Fear of change reflects real risks to delivery, reputation, or expertise. Individual goals and KPIs encourage local optimization rather than collective improvement. Hierarchical pressure reinforces alignment over experimentation.
In such a context, innovative ways of thinking and working (such as connecting the dots instead of querying isolated records) are not perceived as breakthroughs, but as risky disruptions.
A clinical expert may wonder whether their hard-earned intuition is being replaced. A manager may worry about taking responsibility for adopting something not yet proven. A team may simply lack the time to learn a new way of working. None of these reactions are about the technology itself, they are about what the technology implies.
For new ways of working to take root, several conditions must be deliberately created:
-
- People need to see how the change enhances, rather than threatens, their role
- They need space to experiment without immediate performance pressure
- Incentives must align with collaboration and reuse, not just individual output
- Leadership must signal that understanding and adoption are priorities, not optional efforts squeezed between meetings
Most importantly, innovation must be co-created, not imposed. Adoption is not a rollout, it is a social process (see “The Psychology of Knowledge Management“).
2. Process: Where Innovation Meets Reality
If human collaboration is the invisible system, processes are how it materializes.
In pharma, processes are not just operational; they are deeply structured by internal SOPs, regulatory requirements, and domain-specific good practices (GLP in laboratories, GCP in clinical, GMP in manufacturing). These constraints are not optional, they are foundational to trust, compliance, and safety. And yet, this is precisely where many innovation initiatives stumble.
Technology is often introduced as something that will “transform” the organization, without fully accounting for how work is actually performed. In doing so, it unintentionally becomes a constraint rather than an enabler, adding complexity instead of removing it.
Graph-based ways of thinking and working that do not fit into existing workflows, validation frameworks, or documentation practices will remain parallel systems, interesting, but unused. To create value, innovation must embed itself into the fabric of existing processes (see “Navigating Complexities in Life Science Business“).
This starts by identifying where decisions are made and where friction truly exists, not abstractly, but concretely, for example in protocol design, regulatory submission preparation, or safety signal detection. From there, the question is not “how do we deploy the technology?” but rather: “How does this change the way this specific process works without breaking compliance or overly disrupting how people work?”
This typically requires incremental integration:
-
- Starting with narrowly scoped use cases
- Aligning with validation and audit requirements
- Ensuring traceability and explainability
- Involving stakeholders early
- Communicating, communicating, communicating
3. Data: Making Sense Before Making Connections
At its core, graph-based ways of thinking and working promise to connect everything. But highlighting connections without demonstrating how they add value is pointless.
In reality, what we call data, metadata, information, and knowledge are all, fundamentally, sequences of bits in computer memory handled via complex transformations. The distinction lies not in their structure, but in the meaning we assign to them, and meaning is not universal.
What makes sense to one team does not necessarily make sense to another. A “study,” a “patient,” an “endpoint”… these concepts may appear shared, yet carry subtle differences across domains, functions, or systems. This is where many initiatives underestimate the challenge.
Building a knowledge graph is not just about linking data, it is about aligning understanding. It requires defining concepts clearly, disambiguating terms, and reconciling different representations of the same reality. It becomes the foundation for conversations (sometimes long, sometimes difficult) until a shared understanding emerges.
In this sense, knowledge management is not a technical discipline; it is a collective sense-making exercise. A knowledge graph is simply its formalization.
From Truth to Relevance
When executives ask, “How is it relevant?”, they are not dismissing the idea. They are pointing to a missing link, not between data sources, but between a proposal they do not fully grasp and the operational reality they are accountable for under constant pressure.
Their concern is not whether the solution works. It is whether it matters now, within their constraints: timelines, budgets, regulatory exposure, and competing priorities. In that context, anything that is not immediately actionable or clearly impactful is perceived as optional.
More importantly, value becomes tangible when it is observed elsewhere. Executives are far more receptive when they see:
-
- Peers gaining measurable advantages
- Competitors accelerating timelines
- Other organizations reducing risk or cost
At that point, the conversation shifts. It is no longer about curiosity, it becomes about positioning. This is where a powerful, often unspoken driver comes into play: FOMO, the fear of missing out. Not adopting an innovation is rarely perceived as a risk, until someone else proves its value. Once that happens, inaction itself becomes the threat:
-
- “Why are we not doing this?”
- “Are we falling behind?”
- “What opportunity cost are we accepting?”
Activating this shift is harder than building the tech stack. It requires storytelling, evidence, peer benchmarking, and strategic framing. But this is precisely what accelerates adoption.
Because in the end, innovation is not adopted when it is technically correct, it is adopted when not adopting it becomes the bigger risk.