DIA Global Annual Meeting 2026DIA Global Annual Meeting 2026
Who is on stage at DIA 2026 (June 14–18, 2026, Pennsylvania Convention Center, Philadelphia, PA) talking about AI-powered regulatory authoring, content generation, document authoring/creation, and study design / protocol authoring.
在 DIA 2026（June 14–18, 2026，Pennsylvania Convention Center, Philadelphia, PA）台上谈论 AI 监管撰写、内容生成、文档创建及研究设计 / 方案撰写的各方。
Pharma orgs
药企机构
Presenting on AI authoring / study design
就 AI 撰写 / 研究设计发言的药企
CRO orgs
CRO 机构
Presenting on AI authoring / study design
就 AI 撰写 / 研究设计发言的 CRO
Vendors
技术厂商
AI authoring & study design tool vendors
AI 撰写与研究设计工具厂商
Sessions in scope
范围内场次占比
36 of 138 pharma/CRO sessions
共 138 场药企/CRO 场次中 36 场
Regulatory bodies
监管机构
FDA / EMA / PMDA / MHRA / ICH / …
在 AI 撰写相关场次标题中点名的监管方
Top keywords — Vendors vs. Pharma & CROs vs. Regulators热词对比 — 厂商 vs 药企 & CRO vs 监管
Same dictionary, three corpora, three audiences. Bar length = number of sessions whose title or abstract contains the term. Reveals where each audience's narrative concentrates — and where vocabulary diverges.
同一组关键词，三类语料、三类受众。柱长 = 该词出现于标题或摘要的场次数。可见各类受众的叙事重心与词汇分歧。
Vendors ↗厂商 ↗
13 sessions场次- AI11
- Agentic5
- Protocol5
- Submission4
- Automation3
- Validation3
- Quality3
- Clinical Trials3
- Digital2
- Document Generation2
Pharma & CROs ↗药企与 CRO ↗
138 sessions场次- AI47
- Quality17
- Clinical Trials14
- Digital8
- Protocol7
- Governance6
- Real-World Evidence6
- Cross-functional6
- Submission5
- Medical Writing4
Regulators ↗监管 ↗
8 sessions场次- AI7
- Digital2
- Generative AI1
- Validation1
- ICH M111
- Real-World Evidence1
- RWE1
- Knowledge Management1
Bars are normalized within each column (longest bar = max count in that column), so the chart shows narrative composition, not absolute corpus size. Counts come from a shared keyword dictionary scanned across each session's title + abstract.每栏柱条按本栏内最长柱归一化，反映叙事结构占比，而非语料绝对规模。频次基于同一组关键词在标题 + 摘要上的命中。
Vendors厂商
Selling outcomesStage posture: 'AI works in regulated R&D — here's our proof.' All 10 sessions push commercial AI authoring or AI study-design products, mostly in Innovation Theater format.
登台姿态：「AI 在受监管 R&D 已可用——这是我们的证据」。10 场全部为商用 AI 撰写 / 研究设计产品宣讲，多在 Innovation Theater 时段。
From point tools → agentic, end-to-end platforms从点工具 → 端到端 agentic 平台
The pitch has shifted from 'one document faster' to 'connected submission' / 'agentic platform' running multiple steps autonomously.卖点从「单文档更快」转向「连通式递交」/「agentic 平台」自主跑多步流程。
Generating data, not just documents生成的是数据，不只是文档
Vendors reframe AI authoring as building structured, reusable regulatory data assets — answering pharma's 'AI-ready content' demand.厂商把 AI 撰写重新定义为「构建可复用的结构化监管数据资产」，正面回应药企的「AI-ready 内容」诉求。
Validation & 'regulatory grade' as table stakes验证与「监管级」是入场券
3 of 10 vendor sessions explicitly call out validation, governance, or 'regulatory-grade' AI — recognizing that pharma buyers won't move past pilots without it.10 场中 3 场明确点出验证、治理或「监管级」AI——厂商承认不解决这一关，药企买家就不会走出 pilot。
Protocol & study design as the new AI front door方案与研究设计成为 AI 新的入口
Half of vendor sessions touch protocol authoring or study design — a faster ROI surface than full-submission generation.一半的厂商场次涉及方案撰写或研究设计——相比完整递交生成，这是 ROI 更快的切入面。
Pharma & CROs药企 & CRO
Operationalizing AIStage posture: 'How do we deploy AI under quality, governance, and global rules?' AI dominates (47/138) but is paired with quality (17), clinical trials (14), governance (6) — the focus is integration, not adoption.
登台姿态：「如何在质量、治理、全球规则下落地 AI？」AI 主导（47/138），但与 quality（17）、临床试验（14）、治理（6）并列——重心在落地集成，而非「要不要用」。
Quality, governance & compliance > pure capability质量 / 治理 / 合规 > 单纯能力
23 sessions touch quality / governance / compliance vs. 47 mentioning AI — pharma's narrative is 'how to make AI fit the QMS', not 'what AI can do'.23 场涉及质量/治理/合规，AI 出现 47 场——药企讲的是「如何让 AI 适配 QMS」，而非「AI 能做什么」。
Clinical trials & operations are AI's biggest target临床试验 / 运营是 AI 最大的落地战场
14 sessions name clinical trials directly and many AI-tagged sessions sit in trial execution — sponsors and CROs see ops as the highest-volume, highest-ROI surface.14 场直接命名临床试验，大量 AI 场次都落在试验执行——申办方与 CRO 把运营视为体量最大、ROI 最高的应用面。
AI-ready content & 'digital debt' as the prereqAI-ready 内容与「数字债」是先决条件
Pharma openly states AI value is blocked by unstructured content; sessions push lean authoring, structured content, and metadata maturity.药企公开承认 AI 价值被非结构化内容卡住；场次主推 lean authoring、结构化内容与元数据成熟度。
Cross-functional + workforce — AI as org change跨职能 + 人才——AI 是组织变革
6 cross-functional sessions; medical writing, quality, regulatory writers feature explicitly. Framing: AI redesigns roles, not just tools.6 场跨职能；医学撰写、质量、监管撰写者明确出现。叙事：AI 重新设计岗位，而非只是工具。
Regulators监管方
Setting the railsStage posture: 'AI is here, here's the policy and infrastructure layer.' Only 8 regulator-led sessions but every single one sets framework, expectations, or harmonized rails for AI use.
登台姿态：「AI 已来，这里是政策与基础设施层」。仅 8 场监管方主导，但每一场都在为 AI 应用搭建框架、设定预期或统一规则。
Multi-agency convergence on AI review & decision-making多机构在 AI 审评与决策上的趋同
FDA, EMA, MHRA, PMDA appear together — signaling that AI policy is now a global-harmonization conversation, not a single-agency one.FDA / EMA / MHRA / PMDA 同台出现——AI 政策已是全球协同议题，不再是单一机构的事。
Generative AI gets named — with expectations attached生成式 AI 被点名——并伴随期望
Regulators move from 'AI in general' to specific GenAI use cases, telling sponsors what they expect to see in submissions.监管从泛指「AI」转向具体的 GenAI 用例，明确告诉申办方递交时要看到什么。
Digital protocols & ICH M11 as the structural foundation数字方案与 ICH M11 作为结构化基础
ICH M11, digital protocols, and structured content recur — regulators want sponsors to bring AI-ready, machine-readable submissions, not free-text dossiers.ICH M11、数字方案、结构化内容反复出现——监管希望申办方递交 AI-ready、机器可读的内容，而非纯文本卷宗。
Real-world evidence + AI as the next regulatory framework真实世界证据 + AI 是下一个监管框架
FDA pairs RWE and AI in a single session, signaling combined frameworks rather than separate evidence streams.FDA 把 RWE 与 AI 放进同一场——信号是一体化框架，而非两条独立的证据通路。
Narrative summaries are hand-curated from the actual session titles and abstracts in each corpus — they explain what each audience is doing with the keywords above, not just which words they use.叙事概要由各语料的实际场次标题与摘要人工提炼——回答的是「他们在用这些关键词做什么」，而不仅是「他们用哪些词」。
Top 10 hot topics — AI authoring & study design at DIA 2026DIA 2026 中热度 Top 10 主题 — AI 撰写与研究设计
Mined from DIA 2026 pharma & CRO sessions. Ranked by number of in-scope sessions touching each topic. Hover any card to see the source sessions and the pharma / CRO speakers driving it.
从 DIA 2026 所有药企 / CRO 场次中挖掘。按命中本主题的范围内场次数降序排列。鼠标悬停任意卡片可查看来源场次以及驱动该话题的药企 / CRO 演讲机构。
AI-augmented protocol & study designAI 增强的方案与研究设计
Using AI / standardized data to optimize protocol design, complexity, and patient burden.用 AI 与标准化数据优化方案设计、复杂度与患者负担。
AI in regulatory review & agency decision-making监管审评中的 AI 与机构决策
How FDA / EMA / PMDA themselves use AI in submission review and reg decision-making.FDA / EMA / PMDA 等监管机构在审评与监管决策中使用 AI 的实践。
Adaptive, master & decentralized protocol designs自适应 / 主方案 / 去中心化试验设计
Master protocols, adaptive designs, decentralized trials and digital-protocol enablers.主方案、自适应设计、去中心化试验及其数字化载体。
AI-powered regulatory authoring & submission generationAI 驱动的监管撰写与递交内容生成
End-to-end AI for drafting eCTD-ready dossiers, IND/NDA/MAA submissions, and regulatory cover documents.用 AI 端到端撰写 eCTD/IND/NDA/MAA 等监管递交文档与封面材料。
AI governance, validation & fit-for-purpose risk frameworksAI 治理、验证与风险可信度框架
Risk-based credibility, validation playbooks, and governance backbones for regulated AI.面向受监管 AI 的基于风险的可信度框架、验证手册与治理底座。
Medical writing transformation & writer up-skilling医学撰写的 AI 转型与人才升级
How AI is reshaping medical-writing roles, competencies, and organizational structure.AI 如何重塑医学撰写岗位、能力模型与组织架构。
Agentic AI & copilots across R&D and regulatory贯穿 R&D 与监管的智能体 / Copilot
Multi-step agents and copilots orchestrating cross-functional regulatory and clinical workflows.在监管与临床流程中跨职能编排的多步 agent / copilot。
Generative AI for clinical & regulatory document generation生成式 AI 用于临床与监管文档生成
GenAI / LLMs producing first drafts and final-quality clinical and regulatory content at scale.用生成式 AI / LLM 大规模产出临床、监管文档的初稿乃至成稿。
AI-ready content foundations & data debt remediationAI-ready 内容基础与数据债治理
Cleaning up unstructured trial / regulatory content into AI-consumable, metadata-mature foundations.把非结构化的试验 / 监管内容治理为 AI 可消费、元数据成熟的内容底座。
Structured content authoring (USDM, M11, eCTD v4.0)结构化内容撰写（USDM / M11 / eCTD v4.0）
Single-source authoring on USDM, ICH M11, eCTD v4.0 and other machine-readable structures.基于 USDM / ICH M11 / eCTD v4.0 等机器可读标准的单源撰写。
Source: DIA 2026 pharma & CRO session corpus (live.diaglobal.org). Topics are assigned by keyword detection on each session's title + abstract; AI-required topics additionally require an AI / generative / agentic / automation signal in the text or in DIA Track 03 (Data-Tech-AI). A session can appear under multiple topics.数据来源：DIA 2026 药企 / CRO 场次语料（live.diaglobal.org）。主题由关键词在标题 和摘要上的命中分配；标记为「需 AI 信号」的主题还要求文本中出现 AI / 生成式 / 智能体 / 自动化等词，或属于 DIA Track 03（Data-Tech-AI）。一场可同时归属多个主题。
Key pain points raised on the DIA 2026 stageDIA 2026 台上点出的关键痛点
Pain points stated verbatim in DIA 2026 Innovation Theater and workshop abstracts — i.e. what the pharma sponsors and CROs explicitly chose to call out on stage.
逐字摘自 DIA 2026 创新剧场与工作坊议程摘要 —— 这是药企申办方与 CRO 选择在台上明确点出的痛点。
Pilots don't translate into production deployments
试点无法落地为生产级部署
“Sponsors focus on finding AI that can get words on a page, but they often lose sight of a broader, cross-functional, data strategy that this needs to fit within in order to achieve significant ROI.”
“申办方聚焦于把字写到页面上的 AI，却忽视了想要真正实现 ROI 必须配套的跨职能数据策略。”
Generic GenAI is not accurate, traceable or consistent enough for submission
通用 GenAI 在准确性、可追溯性、一致性上达不到递交要求
“Content generation is easy. Delivering outputs that are accurate, traceable, and consistent enough for regulatory submission across diverse document types is not.”
“生成内容很简单；做到足够准确、可追溯、一致到能用于跨文档类型的监管递交才难。”
AI tools that work for CSR fail when teams move to protocol / CMC / PV
在 CSR 上能用的 AI，一移到 protocol / CMC / 药物警戒就崩
“As teams move beyond CSR into protocol, CMC, and pharmacovigilance workflows, many AI solutions fail to adapt, creating risk around compliance, scalability, and usability.”
“团队从 CSR 扩展到 protocol、CMC、药物警戒时，很多 AI 方案无法适配，带来合规、可扩展性与可用性风险。”
Time-to-first-draft is no longer the only ROI metric that matters
首稿时间不再是唯一的 ROI 指标
“While initial focus was solely on speed and accuracy, as solutions mature it's clear that time to first draft isn't the only factor that should be considered when deploying AI within your team.”
“最初焦点只在速度和准确性，随着方案成熟，团队部署 AI 不能只看首稿时间这一项指标。”
Document-by-document tools fragment validation and prevent cross-document learning
「一份文档一个工具」让验证碎片化、AI 无法跨文档学习
“The first wave of AI in life sciences took a document-by-document approach — one solution for the CSR, another for the IB, another for INDs and regulatory submissions. The result: fragmented tools, duplicated validation efforts, and AI that can't learn across your document ecosystem.”
“生命科学第一波 AI 是「一份文档一个工具」——CSR 一个、IB 一个、IND 与监管递交又是一个。结果是工具碎片化、重复验证，且 AI 无法跨文档生态学习。”
AI adoption stalls without governance and cross-functional alignment
缺乏治理与跨职能协同，AI 落地就停在试点
“Attendees will also learn key lessons from a successful collaboration — highlighting the importance of governance, cross-functional alignment, and working as one integrated team to drive adoption and deliver measurable impact across R&D.”
“参会者将听到一次成功合作的关键经验——治理、跨职能协同与作为一个整合团队的重要性，是推动 AI 在 R&D 全面落地、产生可量化影响的关键。”
Static Word protocols cause manual handoffs and downstream delays
静态 Word 方案带来手动交接与下游延误
“Clinical trials still rely heavily on static Word documents to define and communicate protocol intent, creating manual handoffs, data inconsistencies, and delays across every downstream system.”
“临床试验仍严重依赖静态 Word 来定义与传递方案意图，造成下游每个系统的手动交接、数据不一致与延误。”
Trial start-up is delayed by manual processes, fragmented systems, inconsistent data
试验启动被手动流程、系统碎片、数据不一致拖慢
“Clinical trial start-up is often delayed by manual processes, fragmented systems, and inconsistent data.”
“临床试验启动常被手动流程、系统碎片化和数据不一致所拖慢。”
Submissions can't scale without process standardization
没有流程标准化，递交无法规模化
“Process standardization serves as the critical enabler for automation, supporting faster, more scalable submission execution.”
“流程标准化是自动化的关键前提，支撑递交执行更快、更可规模化。”
Rigid fixed designs and historical correlations no longer fit complex trials
复杂试验中，刚性固定设计与历史相关性已不再适用
“As clinical trials grow increasingly complex, relying on historical correlations and rigid fixed designs is no longer viable.”
“随着临床试验日益复杂，依赖历史相关性与刚性固定设计已不可行。”
Regulators demand explainability while sponsors push to compress trial duration
监管要求可解释性，申办方却要压缩试验时长
“Multimodal AI platforms are currently helping biopharma and CROs navigate regulatory demands for explainability while cutting trial durations by 40%.”
“多模态 AI 平台正在帮助 biopharma 与 CRO 在满足监管可解释性要求的同时把试验时长缩短 40%。”