A year of focused progress, smarter workflows, and proven AI impact 

✍️ Editor’s Note

This year, our product development efforts in CAPTIS® were guided by a clear goal: help regulatory and medical writing teams work more efficiently without compromising quality or compliance. Rather than building everything at once, we focused on strengthening core workflows, reducing manual effort, and validating innovation through real-world use.

Below is a snapshot of the key themes that shaped CAPTIS® this year, along with a look at a successful internal pilot that demonstrates where we’re headed next.

📣 Key Development Themes in 2025

1. Expanding AI Capabilities: Secure, Practical, and Verifiable 

A major focus this year was significantly strengthening AI capabilities inside the platform without compromising security, traceability, or regulatory confidence. 

Global AI Chat

We introduced Global AI Chat, a secure, AI-chat tool available directly within the system. Users can work directly with uploaded documents to brainstorm faster, extract information, and generate structured content. 

What makes this different is verifiability. Whenever AI responses are generated from source documents, they are linked back to the original evidence, allowing users to quickly validate outputs. 

Key areas where our users use AI within CAPTIS®: 

• State-of-the-art (SOTA) overviews and literature review strategy generation 
• Automated literature data extraction with source-linked evidence 
• Insight generation and custom summaries from scientific articles 
• Extraction and summarization of technical documentation 
• Intelligent querying of internal knowledge bases 
• Preliminary risk mapping using source evidence 

This capability allows teams to offload repetitive cognitive tasks, accelerate early thinking, and focus expert time where it matters most. 

2. Smarter Literature Reviews and Reference Management

Literature review and reference handling remain some of the most time-intensive activities in regulatory writing. This year, we introduced a set of tightly connected enhancements designed to help teams find, assess, extract, and reuse evidence faster - without losing traceability or control. 

Key enhancements include:

• Source-linked AI assistance: AI-generated content suggestions now include direct links back to their source articles, making it easy for users to verify evidence while significantly reducing manual effort during EU MDR and other regulatory literature reviews. 

• Faster summaries with AI-generated tables: AI Chat can now generate structured tables, enabling quicker comparison and summarization of study data.

• Smarter content discovery: 
  • Date filtering for recently added articles fetched through the auto-update feature 
  • Enhanced keyword highlighting to surface relevant data quickly and support faster review decisions 

• Seamless reuse of prior work: Teams can import entire literature reviews completed outside the system—not just article metadata—and continue their reviews seamlessly within the platform. 

• Configurable PRISMA flow diagrams: Automated PRISMA diagrams can now be tailored to reflect project-specific screening and inclusion workflows. 

• Flexible reporting and exports: Literature reports can be exported as Word files for easier formatting and downstream use. 

• Greater transparency of data sources: Support for adding custom literature search databases ensures every evidence source is tracked and documented. 

• Richer data extraction: Users can extract multiple data points or variants from the same article, supporting complex evidence synthesis. 

• Improved reference and PDF management, including:
  

  • Bulk PDF uploads

  • Downloading PDFs with or without annotations

  • Attaching supplementary PDFs to articles

Together, these improvements reduce administrative overhead, improve  evidence  traceability and help teams move more confidently from literature review to submission-ready documentation. 

🔍 Spotlight: Early Results from an Internal AI Pilot

This year, we ran an internal pilot to explore how AI can practically support Clinical Evaluation Report (CER) drafting when embedded directly into real regulatory workflows. 

The results so far have been highly encouraging. 

Using structured technical and clinical source documents, AI was able to generate first-draft CER content for largely factual, data-driven sections. In these early use cases, teams observed drafts that were often close to completion, allowing writers to shift their focus away from baseline drafting and toward clinical interpretation, justification, and refinement. 

What We Learned

• AI performs best when applied to well-structured, low-subjectivity content 

• Grounding outputs in existing technical documentation supports traceability and confidence 

• Even partial first drafts can deliver meaningful time savings in the overall CER authoring process 

These findings reinforce our belief that AI adoption in regulatory writing should be deliberate, evidence-driven, and validated through real use cases before scaling. 

What This Means for the Year Ahead

The outcomes from this year give us confidence in our direction: 

• Continue investing in workflow-first enhancements 

• Expand AI capabilities where they demonstrably save time and improve consistency 

• Keep validating innovation through real use cases before scaling 

Our focus remains the same: help regulatory teams deliver high-quality work faster, with confidence. 

Thank You 

Thank you to our internal teams and customers who provided feedback, challenged assumptions, and helped shape this year’s progress. Your input directly influences what we build next. 

We look forward to sharing more as these themes evolve into the next phase of our product journey.