|
A breakthrough technology rarely wins because the invention looks impressive. A breakthrough wins
when it makes a costly form of coordination cheap. I’m a student of history, especially the transformative effects of disease and technology on how societies and commerce function. I read many posts on LinkedIn discussing Agentic AI and there is considerable handwringing in some of those posts. Taking a look at history, I suggest the adoption pattern for Agentic AI will be similar. Transportation (canals, railroads, cars, fiber optics) provide a clean reference case for what happens when a technology reduces friction in moving value across distance. Canals illustrate the pattern Canals mattered because they turned water movement into controlled flow. Water always moved, but canals added engineered controls and governance—locks, channels, rules, and tolls—that enabled steady, predictable freight transport across distance. That reliability reduced friction and lowered the effective cost of coordination. Industry reorganized around the new throughput. Commerce soon moved more per unit of time. Railroads repeated the pattern Railroads repeated the pattern with speed and reliability. They moved more per unit of time. Additionally, railroads had a knock-on effect. Rail wear made durability a requirement, not a preference. That demand helped scale steelmaking into mass production. Steel became available in high volume, on quick command that enabled many other infrastructure developments. Cars repeated the pattern again Mass-produced cars repeated the pattern with point-to-point mobility. Cars increased destinations per hour. Cars also changed the map of work because commuting time became a variable people could buy down. Fiber-optic cables repeated the pattern again Fiber-optic cables repeated the pattern with data bandwidth. Fiber increased information moved per unit of time and reduced latency. Early builders laid far more fiber than networks needed, so large stretches sat “dark”—installed but not lit—because demand had not caught up. As data traffic grew, operators lit that capacity and throughput rose without re-digging the ground. Global coordination became routine because distance had little to do with what was delivered per unit of time. Agentic AI belongs in the same category Agentic AI reduces time (and cost) of moving work across systems. Agents increase completed work per day by reducing handoff time between intent and action. Coordination work becomes cheaper than the work it supports. Regulatory writing provides a clear view of the shift. What becomes cheap Most regulatory teams do not suffer from a lack of expertise. Many teams suffer from expensive coordination. Coordination shows up as time spent:
The three-layer framework that predicts outcomes History suggests that durable transformations depend on three layers. 1) Infrastructure layer Canals needed locks, surveying, financing mechanisms, and governance. Agentic AI needs:
Canals needed schedules, toll rules, standards, and dispute handling. Agentic AI needs algorithms to define:
Canals reorganized commerce and rewarded the builders of complements. Agentic AI may reorganize regulatory work and reward teams that deliver:
Where we are right now Most organizations sit in an early “canal” phase: impressive boats, missing locks. Many teams are experimenting with generative drafting while the foundation remains weak:
A strong agentic approach begins with structure: define what the agent may access, what the agent may produce, and what a reviewer must be able to verify. What will likely happen next in regulatory writing Several shifts appear likely. 1) Value will migrate from prose to architecture Sponsors will pay less for “first drafts” and more for:
2) QA and governance will become differentiators Auditability will separate helpful systems from unsafe ones. A defensible workflow will require provenance, versioning, and repeatable checks. 3) Agent design will look like information design
The practical takeaway Canals did not replace judgment. Canals rewarded builders who reduced friction with consistent water flow and rules. Agentic AI will likely do the same. Teams that build the complements—provenance, workflow control, evaluation, and traceability—will define the next standard for decision-ready regulatory writing. Join the conversation on LinkedIn: https://www.linkedin.com/pulse/agentic-ai-feels-new-pattern-old-lessons-from-canals-writing-cuppan-fft1c
0 Comments
Teams often endlessly add to reg submission documents what they call “relevant” content to feel complete. But keep mind, regulatory reviewers need “important” content to make decisions. There is a clear distinction between important and relevant. This mismatch drives reading friction, reviewer questions, and avoidable misunderstanding.
I want to define for you how I distinguish important versus relevant, then explain why the distinction matters. Relevant information Relevant information has a logical connection to the topic. Relevant content may be accurate, helpful, and necessary somewhere in the dossier. Common forms of relevant content include:
Relevant content answers: “Does this relate?” Important information Important information changes, supports, or constrains a regulatory decision. Important content earns priority because the reader must act on it. Important content usually does at least one of these jobs:
Important content answers: “So what for the decision?” A working rule helps teams move faster:
Why the distinction matters in real regulatory review Regulatory reviewers read Module 2 to decide, not to accumulate information Module 2 supports fast, selective decision-making. A reviewer navigates toward decision points, not toward completeness. When sponsors treat relevant as important, reviewers must do some triage. That workload shift creates friction and delays. Relevant overload hides the message you most need the reader to see Dense, undifferentiated text encourages skimming. Skimming changes what a reviewer notices. Further, skimming reduces cognitive recall: readers do not remember what they read. A buried bottom line often becomes a missed bottom line. A missed bottom line becomes a question, then a RFI. Interpretive space becomes sponsor risk Ambiguity invites inference. Inference becomes a working story inside the regulatory review team as they backfill interpretive gaps. A sponsor then spends precious time and intellectual capital rebutting the story. A better document prevents that story from forming. Trust drops when prioritization drops A regulatory reviewer who must hunt for meaning may assume weak reasoning. I suggest a reviewer who must assemble the argument may assume instability in the thinking of the sponsor. This is the 3rd level of information design: Affective. Confidence erodes when the document does not lead the reader tot he "So what?" Additionally, decision time rises when confidence drops. Teams lose time debating sentences instead of decisions Teams often defend paragraphs that they argue are “relevant.” When in turn, the paragraphs are just “frosting on the cake". The reality is teams should align on the regulatory decision questions, rank them, and then construct cogent responses. Prioritization is governance. A shared prioritization method reduces rewrite churn. A practical method: The Decision Test Use this test on any paragraph, table callout, or subsection. The test works in Module 5 and Module 2. Ask 5 questions:
When removal changes little or nothing for the regulatory reader, then relevance exists but without importance. That content then should be subordinated, in a table, in an appendix, or in Module 5. I keep arguing to the teams that I work with: “you gotta stop putting background (relevant information) in the foreground where ONLY important should reside.” To this day I have people tell me they always remember my analogy of treating documents like real estate: “never build cheap nice-to-know information” into your “high rent” districts of the document. I argue that a team should treat the decision test as a permission slip to move text. A team may also treat the test as a permission slip to delete text. A second method: The Section Job Rule Importance is not universal. Importance depends on the job of the section. A simple prompt helps: “What must a reviewer understand after reading this section?” Section purpose determines what deserves top position. Section purpose also determines what should move downstream. In an efficacy conclusions section, baseline detail is often supporting. The bottom line should lead. In generalizability, baseline imbalances may become decision-driving. Those details may become important. Patterns that signal “relevant pretending to be important” Watch for these signals during authoring and reviewing:
These patterns do not mean the content is wrong. These patterns mean the content is poorly ranked. Remember information design is driver of reader understanding. Where relevant content should go Relevant content still matters. Relevant content should not compete with decision-driving messages. Tables for breadth, text for meaning Put breadth in tables.
Appendices for completeness Appendices protect readability in core sections. Appendices also preserve traceability for audits and reviewers. A reviewer may access detail when needed. A reviewer should not be forced to wade through detail. Cross-references to Module 5 for depth Module 2 should guide and interpret. Module 5 should provide underlying evidence detail. A crisp reference often beats a long restatement. Restatement increases bulk without improving decisions. Progressive disclosure within a section Lead with the conclusion.
This order respects how reviewers navigate and reduces misinterpretation. Closing thought Regulatory readers do not need more facts. Regulatory readers need ranked facts that point to decisions.
Sponsors who master the distinction write less, but communicate more. Sponsors who master the distinction also reduce avoidable regulatory inquiry. I strongly suggest that most sponsors still write as if FDA reviewers read slowly, linearly, and generously. My experience says the opposite: the FDA reviewer is a time-starved, human decision-maker. The reading model I constantly talk about starts from that premise. In turn, I argue that the model I espouse should change how you design every page in every document you submit to FDA and EMA.
I always argue to get the "So what?" up front (I am practicing what I preach in this article). “What planet are they from?” On a client call last week, I was challenged with this query: “Well Greg, your model of reading behavior is interesting. But how do you know it applies to people working at FDA?” That question triggered a quiet “WTF?” for me—not because the question was rude, but because of what sat underneath the query. I hear variations of the same line in my workshops:
My answer never changes: “Trust me, all regulatory agency reviewers are a carbon-based life forms like you and me. They are not from some exoplanet identified by the Kepler telescope.” The deeper problem is not skepticism about what I present as a reader behavior model. The deeper problem is a persistent myth: the FDA reviewer is an alien whose behavior sits outside normal human reading constraints and we cannot predict how they will react. This myth feels convenient. Mythology helps development teams avoid harder questions:
Many teams carry unspoken mental models of the regulatory reader:
These models create comfort stories:
The deflecting exhortations protect the development team from added discomfort. In reality, your FDA reviewer:
What my reading model really describes The model I use to explain how expert readers move through complex documents under pressure grew out of the reading research community and the evidence we have collected formally and anecdotally in conversations with regulatory agency reviewers. The observed and anecdotal behavior is clear:
The model I use is a description of high-stakes human reading: Triage → Navigate → Sample → Trace → Cross-check → Decide. If that sounds familiar to your own behavior as a clinical lead, safety physician, regulatory writer, or statistician, then I call it good. This is the point. Why our reading model applies to FDA reviewers Back to the question: “How do you know your model applies to people working at FDA?” Here is the short answer: nothing about the FDA reviewer exempts them from human cognition or organizational pressure. Four reasons matter. Shared cognitive hardware FDA reviewers are expert clinicians, statisticians, and pharmacologists. They are not endowed with extra RAM. They still:
Keep in mind: Professional training refines judgment and process. Training does not rewrite the basic limits of attention and memory. Structural pressures of the role Consider the environment:
These pressures force triage and selective reading. No one in such an environment reads every paragraph with equal depth. The reading behavior model we use represents a rational coping strategy used by agents in the FDA and EMA. Consistency across contexts What I argue as reading behavior appears in:
When context looks the same, (complex evidence, compressed time, meaningful risk) reading behavior converges. FDA reviewers are not the exception. They are the most visible example. If the reading model I describe does not apply to FDA reviewers, then this would imply the "agency" discovered a way around human cognition that the rest of us lack. No evidence supports that belief. People at FDA are indeed carbon-based life forms from good ol' Planet Earth. “How should we structure this document if our success depends on a busy human understanding the core message in one pass?” This is the question to be addressed in every document planning meeting. A final thought: If your organization wants to stress-test your current draft documents against a robust reading and decision-efficiency model, then reach out to me. How Decision Efficiency Redefines Quality
Regulatory writing has evolved more in the past 5 years than in the previous 23 years when the first eCTD guidance reshaped expectations for how sponsors structure and submit information. The move to electronic documents led to wholesale “infobesity” in submission documents. Once submissions shifted from binders to electronic modules, the constraints that once forced concision vanished. Sprawling narratives, dense paragraphs, and an overwhelming volume of “just-in-case” content could be created at virtually no direct cost. In turn, regulatory and clinical research readers did not gain clarity—they gained cognitive load. This shift exposed a deeper truth: regulatory and medical writing was not suffering from a lack of information. Rather the writing was suffering from a lack of decision-oriented design. Our industry has been evolving ever since, moving from text density toward lean writing and now toward decision clarity. Text Density — When Volume Masqueraded as Rigor The early years of application of eCTD created a quiet but powerful shift: once page limits disappeared (or ignored), volume became a proxy for thoroughness. Teams got into the “story telling” business and filled sections with extensive background, descriptions, and narratives about development choices. Exhaustive restatements of data already visible in tables appeared in text whenever the opportunity arose. Text density produced the illusion of rigor. A 200-page Clinical Overview or a 450-page Summary of Clinical Safety looked complete, but regulatory readers had to work hard to locate relevance, reconstruct logic, and discover answers to their questions. The writing provided information, but not orientation. Reviewers encountered paragraphs thick with detail yet thin on meaning. Text density also obscured reasoning. Critical comparisons, implications, and interpretive cues were often buried deep within paragraphs, leaving regulatory readers responsible for making connections the sponsor should have made explicit. This created a systematic drag on regulatory review: more content meant more friction, not more understanding. This era revealed a foundational problem that still affects submissions today: information without structure does not support decisions. Lean Writing — Necessary, but Not Sufficient The industry’s first corrective action to address text density was and still is “lean writing.” Teams recognized that dense passages and exhaustive narrative created reader friction, so the solution became minimalism (a topic I have written about here on multiple occasions). The shift has helped, but only at the surface level. Lean writing reduced noise, but it did not facilitate understanding. Many documents now look cleaner but still lack decision orientation. A protocol may present short, well- edited sentences yet hide the operational logic reviewers and sites need. A Clinical Overview may be concise yet still require the reviewer to infer why a finding matters or what the sponsor means when they use the term “clinically meaningful” 23 times in the document. A Summary of Clinical Safety may use lean prose but avoid interpreting the patterns revealed in its own tables. Lean writing is a valuable improvement—no question. But the gains remain shallow when the deeper reasoning structure stays unchanged. Lean prose without interpretive clarity still forces the reader to supply missing logic, rebuild context, or hunt for warrants behind claims. This era revealed a second truth: reduced text volume does not guarantee reduced cognitive burden. Lean writing improves readability, but not necessarily regulatory decision-making. Decision Clarity — The New Standard The next stage in the evolution of regulatory writing is not shorter text—it is supporting clearer decisions. Decision clarity shifts the writer’s goal from reducing words to reducing the reviewer’s cognitive load. The measure of quality becomes simple: How efficiently can a regulatory reader reach a defensible conclusion? Decision clarity begins where lean writing ends. Once excess words are removed, the real work starts: structuring the logic behind claims, interpreting evidence directly, and guiding the reviewer through the reasoning that connects data to decisions. Decision clarity requires writers to treat every paragraph as an opportunity to make meaning visible. Documents that achieve decision clarity share three characteristics: Lead with the conclusion. Sections begin with the “So what?”—the decision-relevant point—followed by the reasoning and evidence that justify the position. Reviewers never have to hunt for the point the sponsor is trying to make. Make reasoning explicit. Interpretation is not implied or buried. Writers connect effect size to clinical relevance, exposure to safety implications, uncertainty to risk boundaries, and subgroup findings to generalizability. The reviewer is never forced to supply missing logic. Create visible decision cues. Reviewers see the logic trail: purpose → evidence → interpretation → implication. Headings signal meaning, not merely topics. Tables are interpreted directly. Comparisons are explicit. Warrants behind key messages and claims are stated, not assumed. Decision clarity transforms submission documents from passive repositories of information into decision tools. Reviewers gain momentum instead of friction. The thinking of the writing team is traceable. This era reveals the third and most important truth: clarity of reasoning—not speed to final draft, not minimalism—is the core determinant of regulatory decision efficiency. +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ We solved the wrong problem for 20+ years. eCTD made it easy to add content—but not to create clarity. The next leap isn’t faster drafting or cleaner sentences. It’s decision clarity: writing that reduces friction and accelerates reviewer reasoning. Check out my new article in the Decision Efficiency series: www.linkedin.com/pulse/from-text-density-decision-clarity-gregory-cuppan-bvwvc The real performance of a regulatory briefing book is not measured by page count or polish. It is measured by how efficiently a regulatory reader can follow the logic, test it, and reach a defensible position on the sponsor’s ask.
Unfortunately, many briefing books still make the reviewer work too hard. They are organized by how teams built the data—not by how regulators make decisions. That misalignment sounds minor, but it has profound cognitive consequences. The Hidden Cost of Misalignment When a document mirrors internal development logic instead of regulatory reasoning, the reader must reverse-engineer the argument. The regulatory reader sifts through design history (“the largest study ever planned in this patient population”) to locate decision relevance. They assemble fragments of details and data to reconstruct the sponsor’s logic trail. Each mental step adds friction—time, effort, and uncertainty. The concept of friction is not to be underestimated. Legal scholar Brett Frischmann describes friction as resistance or drag built into documents used in legal and regulatory decision processes. Cass Sunstein and Richard Thaler refer to unnecessary complexity in documents as sludge—“bad friction” that drains cognitive energy and delays decisions. In regulated markets, economists have quantified these effects. Studies in The Quarterly Journal of Economics and the American Economic Review show that even modest “choice frictions” in Medicare decisions significantly slow and distort judgments. The parallel for regulatory writing is direct: when reasoning is hard to trace, decision friction rises. I often tell clients that their briefing book delivers information but fails to deliver understanding. Common Patterns of Friction
Tables without interpretive anchors that leave reviewers to construct meaning. Massive appendices meant to prove thoroughness rather than aid reasoning. Each of these choices forces reviewers into reconstruction mode. They also trigger what cognitive-load theorists call the split-attention effect—when information that must be mentally integrated is physically separated across pages, tables, or sections. John Sweller’s research in applied cognitive psychology shows that split attention increases working-memory load, slows comprehension, and makes readers less confident in their conclusions. In briefing books, the effect appears when a data table sits ten pages from its interpretation or when related safety and efficacy evidence live in separate silos. The reader’s mind does the integration work the document should have done. And when reviewers must reconstruct logic, decision friction increases and confidence declines. Reviewers start to question not only the data, but the sponsor’s grasp of its own argument. A document that cannot show its reasoning clearly invites more information requests, longer review cycles, and less trust in sponsor judgment. The Decision Pathway Decision-enabling design exists to solve this exact problem. My vision of information design reorganizes briefing content around the decision pathway—the same mental sequence reviewers follow when testing an argument: Issue → Evidence → Interpretation → Position → Request This sequence is not arbitrary. It mirrors both the FDA Benefit–Risk Framework (issues → evidence → appraisal → conclusion) and the structure used in legal briefs (Issue → Rule → Application → Conclusion). Across disciplines, this logic flow has proven to reduce cognitive friction and increase confidence in reasoning. When that sequence is built into the structure, reviewers stop searching for logic and start evaluating reasoning. That shift—from retrieval to evaluation—is the defining mark of a well-designed briefing book. The Shift: From Information Display to Decision Design Teams often assume that completeness equals clarity. It doesn’t. Regulatory readers don’t reward density—they reward traceable reasoning. A decision-enabling design builds its logic trail around the reviewer’s task. Each section of the book should answer one implicit question: What decision does this content support, and what evidence makes it reasonable? When that sequence is visible, reviewers stay oriented. When it’s hidden, they reconstruct it—usually with lower confidence and more questions. The Hallmarks of Decision-Enabling Design A. Purpose-Anchored Structure Each major section begins with the So what?—why the issue matters and what decision is needed now. This tells reviewers not just what they’re reading, but why it matters. B. Progressive Disclosure Information flows from summary → key evidence → traceable detail. Reviewers can engage at any level of depth without losing sight of the reasoning line. C. Decision Cues at the Point of Use Tables and figures are placed where the decision occurs—not buried in appendices. Each display answers a single regulatory question and closes the loop with interpretation, not adjectives. The Result: Cognitive Efficiency, Not Rhetorical Flourish When design mirrors the decision process, cognitive load drops. Readers no longer have to infer your logic; they can inspect it. Review becomes faster, more confident, and less adversarial. That is the quiet power of decision-enabling design: the approach turns information into defensible reasoning—and does so on the regulatory reader’s terms. The Takeaway A decision-enabling briefing book doesn’t argue harder. It argues smarter. It lets structure do the heavy lifting and keeps reasoning in full view. When regulatory teams master this design discipline, meetings change. Regulatory readers spend less time asking “Where does this fit?” and more time discussing “What does this mean for the development program, the patient and the label?” That’s the point—clarity that enables decision. Join the conversation on LinkedIn: https://www.linkedin.com/pulse/decision-enabling-design-how-briefing-books-should-think-cuppan-3pa7c Regulatory documents are judged not only by what they say, but by how efficiently the documents help regulatory readers decide. Let’s keep in mind that every document you produce that ends up appearing on the computer screen of a regulatory agency reader is intended to be an advisory document. A document intended to aid the reader in making decisions.
However, my experience suggests that many development teams do not fully appreciate this point. Teams tell me they are in the business of reporting. Teams measure quality by format compliance, data completeness, or adherence to templates — not by decision efficiency. You are not in the reporting business. You are in the business of advising. This judgment gap on “what documents do” explains why many submissions are technically correct but cognitively exhausting to read. Most regulatory documents fail not because they are incomplete, but because they are opaque to reasoning. Regulatory readers are not struggling to find data — they are struggling to follow the thinking behind the data. Every table, paragraph, and conclusion should serve one purpose: to let the reviewer see how evidence informed our thought processes and contribute to decision-making. Regulatory reviewers do not need more information. They need better reasoning access — clear logic, visible connections, and writing that mirrors how decisions are made. Decision efficiency measures how well a document enables a regulatory reader to:
Keep in mind that regulatory readers are auditing reasoning associated with study design, conduct, and data sets. When insight into sponsor reasoning pathways are nonexistent or blocked, the cost is real. This is where the friction comes into play—the reviewer should not have to backtrack, infer, or guess. Reviewers spend time reconstructing logic that writers should have made transparent. Dense text, redundant tables, and poorly signposted arguments force readers to think about the document rather than through it. Each moment of confusion compounds uncertainty. Confidence in the sponsor’s reasoning drops, and the agency’s questions multiply. This is not inefficiency of time — it is inefficiency of thinking. The inefficiency of thinking is the far more dangerous dimension. It erodes trust, clouds the evidence trail, and converts clarity into doubt. In an environment where regulatory review cycles are compressed and AI tools assist human readers, the ability to think clearly through a document — not merely within it — is the new competitive advantage. Decision efficiency evaluates the clarity and logic density of a document — not the volume of its words. At McCulley Cuppan, we define decision efficiency through three observable dimensions:
Why This Metric Matters Now Regulatory authorities worldwide are integrating AI-assisted review tools that rely on structured reasoning and clear metadata. Decision efficiency determines whether those tools (and human reviewers) can interpret content without manual intervention. In this context, poor writing is not just a style flaw — it is an information-access problem. A document with low decision efficiency hides logic behind noise. Moving from Compliance to Cognition Teams that focus only on compliance write to satisfy checklists. Teams that focus on decision efficiency write to support thinking. This shift transforms how review rubrics, templates, and training are used:
This is the next frontier of quality in regulatory writing. Decision efficiency is emerging as the missing metric — the measure that predicts whether a document will accelerate or delay regulatory decision- making. Future articles in this series will show how to design decision-efficient documents, measure them with structured rubrics, and visualize reasoning clarity through AI-enabled analysis dashboards. Behind every table, figure, and p-value lies one purpose: persuading regulatory readers that your interpretations are built on logical evidence-based analyses. In the workshops I facilitate, participants always hear me invoke: “regulatory writing is not neutral—it is strategic.” Each justification is an argument that the data are reliable, the analyses reproducible, quality is consistent, and the benefit–risk balance acceptable.
Persuasion in this context is not simply rhetoric. It is “confidence engineering”—helping regulatory readers reach well-supported decisions quickly and with trust. Why Persuasion Matters in Regulatory Writing Briefing books and Module 2 submission documents go beyond summarizing data. They exist to justify scientific and development choices—to explain why a development program, design, or conclusion deserves confidence. Regulatory reviewers approach every document with professional skepticism. They must confirm that claims are supported, methods are sound, and limitations are acknowledged. Writers who anticipate these needs—by shaping information to mirror how reviewers think—make decision-making easier. The goal is not to impress regulators with volume (I have clients who still want to “bulk up” documents). Rather, it is to enable clear judgment through structure, logic, and transparency. Start Where the Reader Starts — Lead with the Conclusion Regulatory readers read for certainty. Lead with your conclusion, then show how the evidence earns it. Use a top-down flow:
selective, and purpose-driven. Signal Your Logic, Don’t Bury It Regulators look for explicit markers of reasoning. Framing phrases act as cognitive signposts:
and reinforce transparency—a hallmark of credibility. Comparison Is the Language of Persuasion Regulatory readers judge claims in context, not isolation. Comparative framing strengthens justification by positioning your evidence within a known landscape.
relevance. Readers Follow Logic, Not Chronology A strong justification follows a predictable rhythm: Why → How → What.
interpretive burden. Regulatory writing, done well, aligns with how reviewers think, process, and decide. That is the science of persuasion: clarity as method, structure as reasoning, and trust as outcome. For years, eCTD v3.2.2 has been the backbone of global regulatory submissions. It gave structure,
consistency, and a common language for sponsors and regulators. But today, our products and the data behind them have outgrown the rigid folder hierarchy of this eCTD platform. eCTD 4.0 is not just a technical upgrade. It is a shift toward data-centric, decision-ready submissions designed to match how regulators actually read and decide. Why eCTD 4.0—Built for modern product complexity For eCTD v3.2.2, the backbone of a submission is a fixed folder tree. Each document must be placed in a rigid location with a pre-defined name. A concept that works well for small molecules and many biologic submissions with predictable sections. But this rigidity creates friction when you add new data streams. For example, emerging therapies like CAR-T cell products and CRISPR-based gene editing make this challenge sharper. These products generate new evidence categories. Such as, chain-of-identity and chain-of-custody records for each patient’s cells, highly specialized potency assays, and genomic off- target analysis for edited cells. Under eCTD v3.2.2, sponsors often wedge these data sets into ill-fitting sections or invent new folder labels The result is added complexity for the regulatory reviewer. Instead of following an established mental map, regulators must pause to interpret where the information sits, why it’s there, and potentially how it links to other evidence. That extra interpretive work most likely increases cognitive load—the mental effort needed to find and integrate meaning. As I have discussed in other articles, high cognitive load slows review, raises the risk of misinterpretation, and invites clarifying questions or requests for additional information. eCTD 4.0 reduces these problems with a metadata-first model. Instead of forcing content into a physical place, each document or data set is described by rich metadata. The metadata provides terms for data type, data source, and regulatory intent. These metadata elements become the “coordinates” for regulators to find what they need. On top of metadata sits “Context-of-Use” tagging. This is a way to declare why the document exists and how it is meant to be used in the dossier review. For example, a potency assay file might carry context tags such as:
In information design, as I teach it in my workshops, an advance organizer is a signal placed before complex material to help the reader build a mental map. By explicitly stating why a document or data set exists and how it will be used in decision-making enables Context-of-Use tags to work in the same way. The tags prepare reviewers to integrate what they are about to read, reduce search time, and anchor interpretation. Instead of scrolling between sections or inferring meaning from file names, the reviewer sees at a glance the analytical or regulatory task the content supports. This is a cognitive service. Over the course of a 200,000+ page submission, reducing even small interpretive friction compounds into meaningful efficiency and fewer avoidable queries. The result: a scalable submission structure. You can introduce new modalities and data types without breaking the CTD map. Reviewers can filter and navigate by function and relevance, not just by where you tucked the file. And you can maintain submission integrity over time — updates to metadata or context can clarify meaning without re-life cycling the entire file. Instead of duplicating entire documents across IND, NDA, BLA, and global variations, eCTD 4.0 uses unique document identifiers and machine-readable metadata to let the same content unit be updated and referenced throughout a product’s lifecycle. This reduces redundant authoring, helps maintain consistency across submissions, and supports automation for publishing and review. Lastly, richer metadata, stable document identity is now critical as regulatory agencies pilot AI tools for comparison, safety signal detection, and labeling review. A well-tagged eCTD 4.0 dossier will be easier for both humans and machines to navigate. Global Status
For organizations ready to adapt, eCTD 4.0 isn’t just compliance work. It’s a chance to modernize regulatory writing, improve clarity and traceability, and align better with how reviewers — and soon their AI assistants — read and analyze submissions. Further Reading
So what? The way regulatory and medical writers design document hyperlinks now influences not only how efficiently human reviewers read but also how effectively AI tools, like FDA’s Elsa, parse, cross- reference, and retrieve content. Link design is no longer a simple way to manage information economy as defined within lean writing precepts. Link design directly shapes comprehension for humans and interpretability for machines.
The Human Reader Challenge Many biopharma document users must read under time pressure. They work within and across long documents with dense appendices, moving between summary sections, data, methods, and supporting detail. Poorly structured links force them to jump back and forth with little context. That jump is not harmless. Cognitive psychology highlights two distinct challenges:
The result: reviewers not only lose the argument thread but also misinterpret the evidence when links overshoot or underserve their purpose. Poor link design magnifies two distinct risks—split attention and misdirected attention—each undermining comprehension in different ways. The AI Reader Challenge Hyperlinks can both improve and challenge Natural Language Processing (NLP) parsing in large technical documents by providing additional context and structure while also introducing noise and formatting complexities. For modern Large Language Models (LLMs), hyperlinks can be invaluable resources to enhance the quality of text-based applications like Retrieval-Augmented Generation (RAG). AI tools such as FDA’s Elsa approach links differently. Machines do not skim, infer, or guess. They parse hierarchies and rely on structure. A vague cross-reference like “see above” or “refer to Appendix 1” leaves a machine with no anchor point. For AI:
The anchor text of a hyperlink—the clickable words—often provides a concise, semantically meaningful summary of the linked content. An NLP model can use this information to better understand the linked document’s topic and relevance. Internal links that connect different sections of the same document can act as a roadmap, informing NLP models of the document’s structure, similar to how a table of contents functions. But not all anchor text is descriptive. Hyperlinks with generic text like “refer to Section 6.4.2” or “see Appendix 1” introduce noise for NLP systems, as they provide no semantic information. Another factor is technical documents in various formats, such as PDFs or legacy file types, may lack consistently marked- up hyperlinks—posing a major challenge for accurate hyperlink extraction. Without consistent structure and metadata, AI parsing has constrained value. Tools may mis-index or mis-categorize evidence, leading to gaps in automated review or flawed analytics. The Dual Design Challenge The challenge for regulatory writers is designing links that serve two audiences at once. Human-centered design:
These principles are complementary, not competing. Links designed well reduce cognitive load for people and improve interpretability for machines. Why This Matters Now Two shifts make link design urgent:
the regulatory argument weakens. Design Principles Going Forward
Future-ready regulatory documents must support two modes of reading: fast, context-seeking human review and precise, structure-dependent AI parsing. Writers who treat links as part of information design—not just formatting—reduce cognitive load for reviewers today and build trust with AI systems tomorrow. The real question is not whether your documents contain links. The real question is: Are your links designed for humans and machine? https://www.linkedin.com/pulse/designing-document-links-humans-machines-getting-right-gregory- cuppan-veeec Readability determines whether your document works—or fails—in practice. A protocol, briefing book,
or other regulatory submission document may be perfectly accurate, but if readers struggle to navigate the document, they most likely waste time, may make mistakes, and likely lose confidence in the work. Working definition: readability is the degree to which a document enables the intended readers to quickly find, understand, and apply the information with minimal cognitive effort. In regulatory submission documents and protocols, readability is not optional—it is risk mitigation. High readability:
what’s understood becomes dangerous. The Hidden Cost of Poor Readability Poorly designed documents:
hardest to fix. Readability in technical and regulatory documents is not a cosmetic feature—it is a competitive advantage. As Saul Carliner observed, this is the 2nd level of information design: enabling documents to perform reliably in real-world use. He also suggests that when your documents reduce cognitive strain, you build trust with readers. The most successful submission documents I’ve reviewed are not only scientifically rigorous, they are designed to be read. The Writer’s Responsibility Patricia Wright stresses the author’s role in designing readable documents: “The message is that the onus for achieving successful communication cannot be safely left to the reader. Writers need to see themselves as catalysts for the strategies that their readers adopt; and they need to be aware of the design features that promote the selection of particular strategies.” Wright’s insight shifts accountability squarely onto the author’s shoulders. Too often, medical and regulatory writers assume that expert readers will “figure it out” even if a passage is dense or disorganized. That assumption is dangerous in regulatory contexts, where readers work under time constraints, juggle multiple documents, and must reach reliable conclusions. Readable writing is not about lowering standards—it is about following document design standards as an act of responsibility. Authors shape the strategies readers use and they create conditions that promote consistent, accurate, and rapid comprehension. |
AuthorGregory Cuppan is the Managing Principal of McCulley/Cuppan Inc., a group he co-founded. Mr. Cuppan has spent 30+ years working in the life sciences with 20+ years providing consulting and training services to pharmaceutical and medical device companies and other life science enterprises. 
Archives
December 2025
Categories |
RSS Feed
Services |
Company |
|
