Field notes from the regulated-writing frontier.
Articles, essays, and technical writeups from the Asthra team and our advisor network. Written for medical writers, regulatory leads, and the people who hold them accountable.
The expedited lane is now the default. This has an implication on regulatory drafting.
More than 70% of last year's new drugs took an expedited path, and most of this month's designations are Fast Track or Orphan. The compressed timeline lands on a part of the process nobody plans for — the drafting.
2026
FDA reversed its CRL pause in a week. That's not the part worth watching.
The FDA paused public release of Complete Response Letters, then posted 14 new ones days later. The policy will keep swinging. Meanwhile there were 455 rejection letters already in the public database when I checked, and almost no one drafting submissions is reading them.
2026
Solving hallucinations won't make an AI regulatory-writing tool work
Everyone asks whether the AI hallucinates, and assumes that if it doesn't, the product works. A regulated document is a big, coherent argument grounded in real data — and getting a tool to actually help write one takes more than accuracy.
2026
FDA Pauses the Public Release of Rejection Letters: What a World of Publishable CRLs Asks of the Writing Organisation
FDA has paused its proactive public release of Complete Response Letters while HHS works toward a formal disclosure rule. Whether CRLs remain confidential or become routinely publishable, the writing implication is immediate: sponsors should write submissions and CRL responses as if FDA's description of any deficiency could eventually be read by investors, competitors, journalists, litigators, and patient groups.
2026
One Batch, a Dozen Tables: The Consistency Problem in CMC Module 3
A CMC / Module 3 dossier is mostly interlocking tables, and the same batch, impurity, and limit turn up across many of them. Keeping those appearances consistent, and traceable, is what makes Module 3 hard to automate — not the writing. Here's how we think about it.
2026
Casgevy's 53-Day Approval: What a National Priority Voucher Review Asks of the Writing Organisation
FDA approved expanded use of Vertex's Casgevy for patients aged 2 years and older with sickle cell disease or transfusion-dependent β-thalassemia, after a highly compressed review under the Commissioner's National Priority Voucher pilot. When the review window is measured in weeks, the dossier has to be review-ready at filing: internally consistent, cross-referenced, and pre-answered against foreseeable questions on pediatric extrapolation, conditioning risk, genome-editing safety, and long-term follow-up.
2026
The House China Probe and Trial-Site Provenance: When Data Geography Becomes a Submission Risk
A House China Select Committee inquiry into major drugmakers' clinical-trial activity in China has put trial-site geography under new scrutiny, including sites in Xinjiang and institutions linked to the Chinese military. For regulatory writing teams, the lesson is provenance: sponsors need to know where pivotal evidence came from, which claims depend on which sites or regions, and whether the benefit-risk case still stands if a regional data subset is challenged.
2026
When Pivotal Data Gets Retracted: What the Tavneos and Cancer-Timing Retractions Mean for Dossier Provenance
NEJM retracted the pivotal ADVOCATE trial publication supporting Tavneos, while Nature Medicine retracted LungTIME-C01, a high-profile China-based trial on time-of-day immunochemotherapy in NSCLC. For regulatory writing teams, a retraction is a provenance event: the organisation must know quickly which claims, sections, documents, and arguments relied on the withdrawn source — and what remains supported without it.
2026
Regulatory Review: An FDA/ICH Audit Inside the Draft, Grounded in Real openFDA Precedent
Asthra's Regulatory Review runs a section-by-section FDA/ICH audit right inside the document — grounded in curated ICH E3 expectations and real approved-product labeling from openFDA — and writes specific, actionable regulator flags inline where the writer works.
2026
When the Confirmatory Trial Fails: Sarepta, Traditional Approval, and the Hardest Submission a Writing Organisation Handles
Sarepta's ESSENCE confirmatory study for AMONDYS 45 and VYONDYS 53 did not meet statistical significance on its primary endpoint, but FDA has accepted sNDAs seeking conversion from accelerated to traditional approval, with a February 28, 2027 target action date. This is one of the hardest regulatory-writing cases: the submission has to acknowledge the failed confirmatory trial, reconcile the original dystrophin-surrogate rationale, and make a totality-of-evidence case without pretending the miss does not matter.
2026
AbbVie's $10.9B Move for Apogee: What Happens to the Dossier When a Clinical-Stage Program Changes Hands Mid-Development
AbbVie is acquiring Apogee Therapeutics for $10.9B, centered on zumilokibart / APG777, a half-life-extended anti-IL-13 antibody with Phase 2 atopic dermatitis data and broader potential in asthma and EoE. The regulatory-writing problem is not ordinary bolt-on cadence. It is a living-dossier handoff: open IND history, IB revisions, active trials, safety updates, and agency correspondence must move into AbbVie's system without losing continuity.
2026
Operation TrialBlazer and Phase-Appropriate Requirements: What FDA's Modernization Push Asks of the Writing Organisation
FDA's Operation TrialBlazer aims to accelerate drug development from the pre-IND / IND transition through late-stage trials, with a Phase 1 IND Navigator, updated Phase 1 CMC resources, dose-selection guidance, master-protocol updates, and an explicit focus on phase-appropriate requirements. For regulatory writing teams, the implication is direct: every section now has to be scoped to the phase deliberately, with clear deferral language and reviewer-visible rationale.
2026
Regenxbio Resubmits Its Hunter Syndrome Gene Therapy Without a New Study: The Rare-Disease Resubmission Pattern Takes Shape
Regenxbio plans to resubmit NAVSUNLI / RGX-121 for Hunter syndrome without enrolling additional patients or running a new study, after FDA indicated the existing data could support accelerated approval. Coming days after FDA reopened uniQure's AMT-130 path, the move signals a more flexible rare-disease posture. The resubmission dossier now has to reconcile the prior CRL, make the existing evidence load-bearing, and carry a credible confirmatory plan.
2026
Broad Win, Subgroup Miss: The Label-Writing Problem Behind Exelixis's Zanzalintinib Review
Pfizer's sigvotatug vedotin failed a late-stage lung cancer trial outright. Exelixis faces a harder writing problem: zanzalintinib plus atezolizumab met the OS endpoint in the ITT colorectal cancer population, but missed in the non-liver metastases subgroup while FDA reviews the NDA. Mixed-efficacy results are where benefit-risk, subgroup interpretation, and label scope have to be written with unusual discipline.
2026
The Safety Analysis Moved to ITT. Now Reconcile 200 Pages.
A late change to the safety analysis population recomputes every adverse-event table, exposure summary, and benefit-risk paragraph in a CSR. Here's what that costs by hand — and what it costs with an agent that re-grounds every number in the source.
2026
FDA Has Seen 1,000+ AI-Related Submissions. Regulatory Writing Has to Catch Up.
At DIA, FDA officials said CDER and CBER have received more than 1,000 submissions incorporating AI elements. FDA is also advancing drug-repurposing work using real-world evidence, while industry is asking for more clarity on AI in manufacturing. None of this directly names regulatory writing. But all of it creates new writing surfaces: context of use, model credibility, RWE integration, and Module 3 descriptions of AI-supported quality systems.
2026
Grace Science and the CMC Gate: Writing Under Rare-Disease Runway Pressure
Grace Science is developing GS-100, an AAV9 gene replacement therapy for NGLY1 deficiency, and says FDA's plausible-mechanism framework has not resolved an unresolved manufacturing-data issue. The company is running short on capital and is seeking a path forward. For the writing organisation, the hard problem is CMC under runway pressure: making the existing manufacturing evidence, residual-risk argument, and post-approval commitment plan as clear and defensible as possible without pretending writing can replace missing data.
2026
Jazz–AbCellera's T-Cell Engager Pact: What Pre-IND Writing Looks Like Before the Modality Is Locked
Jazz Pharmaceuticals and AbCellera announced a staged T-cell engager collaboration that could reach multibillion-dollar value if multiple programs advance and milestones are met. The programs are preclinical, but the writing work starts now — construct vocabulary, mechanism description, safety logic, dose-finding rationale, CMC language, and partner-to-sponsor handoff conventions will shape every future regulatory document.
2026
Lilly Buys 4E: Why Bolt-On Integration Is Becoming a Writing-Organisation Function
Eli Lilly's acquisition of 4E Therapeutics adds another non-opioid pain platform to its 2026 deal run. Against a broader biopharma backdrop of larger partnerships and rising China-sourced licensing, the durable problem is not deal execution — it is dossier integration: inherited IBs, agency history, CMC source documents, safety language, and clinical narratives have to be absorbed without creating long-term fragmentation.
2026
Parabilis's $670M IPO: What Parallel-Indication Writing Looks Like at Capital Scale
Parabilis Medicines raised $670M in a record-setting biotech IPO, giving it capital to advance zolucatetide toward Phase 3 in desmoid tumors while moving more aggressively into additional indications. For the writing organisation, that changes the operating model: shared molecule content, synchronised safety language, product-level CMC, and indication-specific clinical narratives now have to move in parallel.
2026
uniQure's AMT-130 Reversal: What a Whiplashed BLA Asks of the Writing Organisation
FDA has reopened an accelerated-approval path for uniQure's AMT-130 gene therapy for Huntington's disease, after previously saying the existing Phase 1/2 data were insufficient and recommending another trial. The BLA now has to present one coherent argument across prior FDA alignment, later FDA skepticism, and renewed FDA openness — while grounding the case in the matured three-year data, external-control logic, and confirmatory-trial plan.
2026
Build, Partner, or Purpose-Built: The Pharma AI Question Is Framed Wrong
Pharma leaders are increasingly asking whether to build AI in-house or partner with vendors. That framing is too narrow. The real question is architectural — does the system constrain sources, surface inference, preserve writer judgment, and produce provenance by default? A purpose-built regulatory-writing platform can be internal, vendor-led, or hybrid. What matters is what the architecture asks the writer to verify after generation.
2026
GSK, Nuvalent, and the NDA Integration Problem Inside a Live PDUFA Window
GSK's $10.6B acquisition of Nuvalent brings two late-stage oncology NDAs already under FDA review: zidesamtinib, with a September 18, 2026 target action date, and neladalkib, with a November 27, 2026 target action date. The strategic logic is lung cancer. The writing problem is continuity — agency history, IR responses, labeling, and launch-facing documents now have to survive an ownership transition inside a live PDUFA window.
2026
Biogen's Salanersen Breakthrough: What Licensed-Asset Writing Inheritance Looks Like in Phase 3
FDA granted Breakthrough Therapy Designation to Biogen's salanersen/BIIB115 for SMA on June 4, 2026. The once-yearly antisense oligonucleotide was discovered by Ionis and licensed to Biogen, which is now running three global Phase 3 studies. The writing-org story is how a licensee turns inherited early-development architecture into a coherent registrational dossier.
2026
When Is a Confirmatory Trial 'Underway'? FDA's Accelerated Approval Standard Is Now a Writing Surface
FDA's draft guidance explains when a confirmatory trial may be considered 'underway' for accelerated approval: a credible target completion date, sponsor progress and plans that provide sufficient assurance of timely completion, and initiated enrollment. For sponsors, the implication is practical — the confirmatory-trial plan now has to be written as a defensible, traceable submission argument.
2026
Lilly, Boehringer, and the Module 3 Problem Behind Manufacturing-Geography Shifts
Lilly is reducing its planned Alzey investment by roughly half, and Boehringer Ingelheim has paused or stopped €900M of German infrastructure spend. The trigger is healthcare-cost reform and industrial-policy pressure. The regulatory-writing implication is conditional but important: if manufacturing, testing, release, or supply-chain dependencies change, Module 3 and variation narratives have to move quickly.
2026
SCOTUS Backs Hikma on Skinny Labels: What the Ruling Asks of Labeling Teams
On June 4, 2026, the Supreme Court unanimously sided with Hikma in its skinny-label dispute with Amarin. The ruling protects the carved-out-label pathway, but it also sharpens the writing lesson: the label, public statements, medical-information responses, and promotional materials must not create a record of affirmative encouragement for the patented use.
2026
From a One-Line Indication to a Cited Protocol Synopsis
Asthra's seed-then-expand authoring flow turns a bare indication into a complete, standards-aligned protocol synopsis — drafted end-to-end, grounded in registered trial evidence, and surfaced as cited decisions the writer can steer.
2026
What Is Regulatory Writing Automation? A 2026 Definition for Life Sciences
Regulatory writing automation is the use of AI to draft, cite, and QC regulated documents like CSRs, PSURs, DSURs, CERs, and CMC modules. Definition, examples, and what to look for in 2026.
2026
When 'Zero Citations' Is the Wrong Alarm
Our grounding metric read near-zero on the sections we'd grounded most carefully. We spent a week 'fixing' the drafting before we realized the number was lying — not about the drafts, about itself. The lesson: a confident, wrong metric is worse than no metric.
2026
Brenus Pharma's STC-1010 IND Acceptance: Multi-Region First-In-Class Writing Before the Guidance Exists
FDA accepted Brenus Pharma's IND for STC-1010 — a first-in-class allogeneic in vivo immunotherapy built on Stimulated Ghost Cells technology — for MSS metastatic colorectal cancer. The program already holds ANSM and AFMPS authorisations in Europe. Multi-region, novel-modality, pre-guidance writing is its own architecture problem.
2026
BridgeBio's BBP-418 Priority Review: What a First-In-Disease NDA Asks of the Writing Organisation
BridgeBio's NDA for oral BBP-418 in LGMD2I/R9 was accepted with Priority Review on May 27, 2026 — target action date November 27. If approved, BBP-418 would be the first treatment for any form of limb-girdle muscular dystrophy. The dossier carries four regulatory designations across two agencies. The writing problem is the compression of a full submission program into a six-month review.
2026
Replimune's Third BLA: The Writing Playbook When the Agency Stands By Its Rejections
Replimune is preparing a third BLA submission for RP1 plus nivolumab in advanced melanoma after two CRLs in 2025 and April 2026. Unlike Outlook's wet AMD program, there is no appeal reversal here — the FDA has maintained its substantial-evidence objection to the IGNYTE trial design across cycles. The third-cycle writing problem is different when the regulatory posture has not changed.
2026
Every Run, Every Quality Flag, in One Report — Asthra's End-of-Run QC
When the agent finishes drafting, Asthra runs a structured QC pass — cross-reference integrity, statistical-claim sourcing, citation density, gap detection — and ships a quality_report.json alongside the draft. Here is what it checks and why.
2026
Tables and Figures in the Chat — Agent-Driven Python You Can Audit
Adverse-event line listings, sales-exposure data, lab parameter shifts — the data behind a CSR or PSUR runs into thousands of rows. Asthra's data analysis agent handles the analysis inside the chat with deterministic Python, every step audit-logged. No manual filtering, no LLM-fabricated numbers.
2026
Walk-Away Drafting: How to Generate a CSR Overnight
Initiate the run, step away, and come back to the studio fully briefed. Asthra lands a first draft + bibliography + hyperlinks + QC report in 1–2 hours; the writer reviews and refines in the studio with the full agent history in view — nothing is lost between the run and your review.
2026
What's in a Regulatory-Grade Run Bundle
Every Asthra run ships four artifacts side by side — the .docx, traceability.json, quality_report.json, and run_manifest.json. Each one answers a specific question a regulator or auditor would ask, and together they make the deliverable portable, reproducible, and inspector-ready.
2026
Camizestrant and SERENA-6: Same Data, Different Regulatory Questions
EMA's CHMP recommended approval of AstraZeneca's camizestrant combination after SERENA-6 showed a 56% reduction in risk of progression or death. FDA's ODAC voted 6-3 against a favorable benefit-risk assessment for broadly the same proposed use, focusing on the pre-progression ESR1-triggered switch design. The FDA has now extended its review. The lesson for global submission teams is not that the data changed by region. It is that the regulatory question did.
2026
Lilly's Three Vaccine Deals: The Regulatory Writing Integration Problem Behind the $3.8B Bet
Eli Lilly announced agreements to acquire Curevo, LimmaTech Biologics, and Vaccine Company for up to approximately $3.83 billion. The strategic logic is infectious-disease prevention. The regulatory-writing problem is integration: three vaccine portfolios, three document histories, three sets of agency assumptions, and one acquirer-side writing organisation that has to make them coherent.
2026
Outlook's Fourth FDA Cycle: The Writing Playbook After Three CRLs and a Granted Appeal
Outlook Therapeutics expects to resubmit LYTENAVA/ONS-5010 for wet AMD in June 2026 after FDA granted its appeal and concluded that substantial evidence of effectiveness has been established. After three CRLs across manufacturing, inspectional, and efficacy issues, the fourth-cycle BLA is a case study in cross-cycle traceability and evidence reframing.
2026
BMS Signed Anthropic. The AI-in-Pharma Bet Is Real — and the Writing Room Is Where It Lands First.
BMS announced an Anthropic R&D partnership the same week a CRO report said AI keeps repeating the same trial-design mistakes. Both are true. The question isn't whether AI belongs in pharma — it's which layer of the workflow pays off first.
2026
One Molecule, Many Indications: Retatrutide and the Writing-Org Problem Sponsors Aren't Sized For
Retatrutide's Phase 3 obesity readout is the headline. The operating story is the program shape — one molecule running pivotal trials across obesity, T2D, OSA, MASH, cardiovascular outcomes and knee osteoarthritis, with the GLP class expanding into more indications behind it. That is not a hiring problem. It is a writing-architecture problem.
2026
From Prompt Engineering to Standards Engineering
The first wave of AI in regulatory writing was the chat box. We built a versioned skill registry instead — and it is becoming the system-of-record for how regulatory standards apply to a draft.
2026
Hyperlinks Belong in the Draft, Not the Publishing Phase
Cross-references in regulatory submissions usually get wired up weeks after the draft is locked. We moved that work into the draft itself — anchors at write time, hyperlinks before publishing.
2026
Literature Search Belongs Inside the Regulatory Draft
PubMed, ClinicalTrials.gov, and bioRxiv are now reachable from inside Asthra's chat — without the browser-tab detour that used to break a writer's flow and let evidence go stale.
2026
Atara, Pierre Fabre, and the Third-Shot Submission: Anatomy of a Post-Rejection Rewrite
FDA has indicated a path to reconsider Atara and Pierre Fabre's twice-rejected tabelecleucel BLA. In a third-shot package, the writing function becomes load-bearing: every claim, analysis, and cross-reference has to map back to the prior CRL.
2026
CDER Loses Its Chief: When the Drug Center's Top Reviewer Walks Mid-Cycle
Endpoints reports CDER chief Tracy Beth Høeg is expected to leave the FDA, days after Commissioner Makary's exit. The drug center is now the third top FDA seat in turnover inside two weeks. Sponsors with active CDER submissions need a reviewer-agnostic plan.
2026
Immgolis and Immgolis Intri: What the New Golimumab Biosimilars Tell Writers About Totality of Evidence
FDA approved Immgolis and Immgolis Intri as the first interchangeable biosimilars to Simponi and Simponi Aria in May 2026. For 351(k) writers, the useful lesson is not the approval count. It is the evidence hierarchy FDA highlighted: analytical similarity first, with PK and immunogenicity supporting the totality-of-evidence narrative.
2026
In Vivo Cell Engineering at ASGCT 2026: The IND-Enabling Writing Problem Behind the Science
ASGCT 2026 showed growing momentum around in vivo cell engineering, from in vivo CAR-T to gene writing and gene regulation platforms. The science is advancing quickly, but the IND-enabling writing challenge is still unresolved: how do sponsors build CMC and nonclinical narratives when the precedent base is thin and borrowed from adjacent modalities?
2026
The Zillow for Abandoned Gene Therapies: What an Acquirer Really Inherits
ASGCT and the Orphan Therapeutics Accelerator launched CGTxchange in May 2026 to help shelved cell and gene therapy programs find new sponsors. For acquirers, the regulatory file is not just diligence material. It is the operating manual for whether the asset can be restarted without losing agency context.
2026
CNPV Pilot at Seven Approvals: Bizengri and the Priority Voucher Pathway
The FDA's Commissioner's National Priority Voucher pilot just landed its seventh approval with Bizengri for NRG1-fusion cholangiocarcinoma. A June 4 public hearing will shape what comes next.
2026
FDA's Drug Repurposing RFI: Why This Is a Regulatory Writing Opportunity
The FDA opened a docket on May 13 asking for drug repurposing candidates — explicitly inviting AI/ML preclinical evidence. Comments are due through June 29. The window matters more than the form.
2026
From IPO to First Annual Report: The Disclosure Stack Newly Public Biotechs Miss
An IPO doubles a biotech's disclosure workload overnight. The regulatory writing function rarely gets re-resourced in time. Here is what gets missed and how to plan against it.
2026
Makary Out: What an Interim FDA Leadership Window Means for Your Active Submission
FDA Commissioner Marty Makary resigned May 12, 2026. Existing guidance stays operative; new guidance pauses. Every sponsor with a submission in flight should re-plan inside 90 days.
2026
Milestone Payments as a Writing-Pace Forcing Function
When a partnership milestone pays, the regulatory writing function inherits a calendar. The partnership announcement is the loud event. The document acceleration that follows is the quiet one.
2026
The $5B Impairment and the Documentation Lifecycle
When a commercial decision retroactively reshapes a product's value, the documentation stack does not auto-adjust. The CSL impairment this week is a case study in why regulatory writing needs to be in the room before the writedown.
2026
Biotech's New Company Model: Chinese Assets and the IND-Bridging Problem
Endpoints identified 18 US biotechs since 2025 launching with in-licensed Chinese clinical-stage assets. The startup model is new. The regulatory writing problem — bridging Chinese trial documentation into a US IND — is not.
2026
CNPV Pilot — Seven Approvals in Twelve Months. What the Voucher Program Tells Regulatory Writers.
Bizengri's NRG1-cholangiocarcinoma approval on May 8 marks the seventh FDA approval under the Commissioner's National Priority Voucher pilot. The pace is the signal. Here's what to read into it before the June 4 public hearing.
2026
Daiichi Sankyo's $950M ADC Write-Down: Decommissioning the Capex Bet
Daiichi recorded a ¥149.4B 'extraordinary loss' on May 8 after scrapping antibody-drug-conjugate manufacturing capacity. The decommissioning problem returns — this time at the capex level, with documentation implications that follow.
2026
Postapproval Pregnancy Safety: The Final Guidance and the PSUR Architecture Problem
FDA's May 8 final guidance on postapproval pregnancy safety studies adds new methodological expectations for pregnancy registries, RWD complements, and case-report studies. The labelling implications land squarely in the PSUR.
2026
The CBER Leadership Vacuum: What an Open Center Director Search Means for Cell and Gene Therapy Submissions
Endpoints reports the FDA has narrowed its search for a CBER leader without a name attached. Sponsors with active cell and gene therapy filings need a tactical read on what shifts during interim leadership periods.
2026
Decommissioning a Program's Documentation Stack: When the Sponsor Terminates
Pfizer ending its Trillium-related programs is the headline. The regulatory writing question is what happens to the documentation, the audit trail, and the institutional memory when a sponsor closes a program down — especially mid-stream.
2026
FDA Elsa 4.0 and HALO: When the Reviewer Has AI, Editing Alone Falls Behind
FDA just shipped Elsa 4.0 with custom agents, document generation, OCR, and secure search — and consolidated 40+ submission portals into HALO. The asymmetry between the reviewer's stack and the writer's stack just widened.
2026
FDA's One-Day Inspectional Assessments: Continuous Readiness Replaces Cycle Prep
The FDA's new one-day pilot ran 46 inspections in two months and is expanding through FY26. The era of preparing for inspection windows is ending — your documentation has to hold up cold.
2026
OND Custom 4.1: FDA's Internal Review Tooling Is Versioning Faster Than Most CSR Pipelines
The Office of New Drugs released version 4.1 of its custom review tooling on May 4, 2026. Two meaningful versions inside a year is the leading indicator. Here's how to read it.
2026
Orphan Drug Designation and the Six-Month CSR Calendar Reset
Orphan Drug Designation is celebrated as good news. For the regulatory writing function, it is a calendar event that compresses CSR, briefing book, and pre-BLA preparation into a much tighter window than most teams plan for.
2026
PDUFA VIII: Reading the FY27 Dear Colleague Letter as a Submission Volume Forecast
The FY27 PDUFA Dear Colleague Letter dropped on May 4. Most teams skim it for fee schedules. The strategic teams read it for what it implies about review capacity and submission throughput in 2027.
2026
Phase 2 Win + Capital Raise = Six Months of Documentation Crunch
A positive Phase 2 readout followed by a financing round looks like a celebration. For the regulatory writing function, it's the start of a 180-day timeline collision that most teams underestimate.
2026
Project Optimus and the ORR Bridge: When Confirmatory Readouts Have to Carry More Than the Primary
Post-Project Optimus, the FDA's tolerance for narrow ORR-based accelerated approvals has tightened. The CSR and BLA narrative for a confirmatory readout now has to do framing work the primary readout used to do alone.
2026
Beyond Frontier Models — Why Regulatory AI Needs a Purpose-Built Stack
Frontier LLMs and generic RAG aren't enough for regulatory writing. A purpose-built stack — document model, retrieval, planning, and verification — is what separates a demo from a draft a regulator will accept.
2026
Why Regulatory Writing Is Broken — And How Agentic AI Fixes It
Regulatory writing teams spend 60–70% of project time on mechanical document assembly while sponsors lose $1.4M per day of delay. Here's why generic AI tools fall short, and what an agentic, document-aware system actually changes.
2026
First PROTAC Approval Rewrites the Regulatory Playbook for Degraders
FDA's approval of vepdegestrant establishes the first regulatory precedent for PROTAC protein degraders, creating a framework that will shape development strategies across the emerging degrader landscape.
2026
FDA's First Non-Antipsychotic Dementia Agitation Approval Changes Treatment Paradigm
FDA approves Auvelity for Alzheimer's agitation, marking a breakthrough in dementia care. We analyze the regulatory pathway, novel endpoints, and what this means for CNS drug development.
2026
Kymera's KT-621 Fast Track: What Degrader Designations Signal for Respiratory Programs
FDA grants Fast Track to Kymera's IRAK4 degrader KT-621 for eosinophilic asthma. We analyze what this designation reveals about regulatory pathways for targeted protein degradation in respiratory disease.
2026
When Partnerships End, Documentation Inheritance Becomes a Regulatory Burden
Big pharma partnership terminations leave smaller biotechs holding jointly-authored documents that need sponsor re-issue. The regulatory writing burden of these terminations is rarely discussed but significant.
2025
Writer-Defined Provenance: Why Source Control Matters in AI-Assisted Regulatory Writing
How writer-defined provenance ensures that every piece of AI-generated regulatory content is traceable to its specific source document, section, table, or image.
2025
CSR Timelines: From One Month to Less Than 10 Days
How AI-assisted drafting is compressing Clinical Study Report timelines from weeks of manual effort to days of focused writing, without compromising quality or traceability.
See it on a real CSR, PSUR, or CER.
Asthra's five-step loop — Plan, Retrieve, Draft, Review & refine, Hand off — running on real regulatory documents, not slides.