How It Works | Verisian Clinical Trial Automation Platform

SDTM & ADaM programming

With Verisian

AI agents generate unambiguous, complete specifications, validated by experts

AI agents generate analyses and results directly from specifications

Code, data, and results continuously validated with traceability and AI, overseen by experts

Code, data, and results always in a validated, submission-ready state

Continuous conformance with CDISC CORE

New data failures are automatically detected and fixed in minutes

All impacts for any change are transparent with traceability

Without Verisian

“Good enough” specifications copied from previous studies

Analyses manually programmed from ambiguous, incomplete specifications

Validation based on expensive double programming or risk-based manual review

Validation only at set milestones, too costly to run more often

Pinnacle21 conformance runs only during submission timelines and milestones

New data breaks existing code silently

Changes in SAP and study design have unclear effects across code, data, and results

Database lock

With Verisian

Achieve database lock days after last patient visit

Confidently unblind on continuously valid code, data, and results

CDISC conformance already validated before database lock

Conformance issues already resolved by expert-guided AI agents, before
the critical path begins

Data loss and implementation errors resolved as they happen, before the critical path begins

Without Verisian

Multiple soft-lock cycles after the end of your final pivotal trial

Unblinding signed off slowly, on code and data with high risk of unseen issues

Expensive CDISC conformance checks on XPT files delaying time-to-market timelines

P21 issues decoded by hand, one at a time, out of thousands, then manually fixed

Data loss and implementation errors noticed during final validation, affecting submission timelines

Final study results

With Verisian

Fully validated final SDTM, ADaM, and TLFs generated right after lock, ready for sign-off

Earlier clinical review and decision making on continuously valid, submission-ready results

Analyses stay validated pre-lock and pre-LPLV, so final results arrive without re-analysis loops

Without Verisian

Weeks of final results re-programming after database lock

Clinical review only at the end, forcing last-minute changes

Results not trusted until the final locked run

Submission documents

With Verisian

Guaranteed consistency between code, data, the SAP, specifications, and submission documentation

Full SDTM and ADaM Define-XMLs generated in minutes by AI agents grounded in traceability

SDRGs and ADRGs generated in minutes by AI agents grounded in traceability

Patient narratives composed by AI agents directly from study data, reviewed and approved by experts

Key CSR sections generated by AI with guaranteed consistency, experts own the clinical interpretation

Without Verisian

Inconsistencies between code, data, and submission documentation, leading to health authority questions and potential rejection or resubmission

Define-XML built and manually validated during submission timelines

Reviewer's guides built by hand during submission timelines

Patient narratives written by hand

CSRs drafted entirely by hand

Customer Story

How Bayer Uses Verisian AI to Automate Submission Document Generation

Integrated analyses (ISS/ISEs)

With Verisian

Pooling starts immediately on continuously valid input studies

Input study specifications are unambiguous and complete

Pool planning as an AI-assisted workflow, in full context of input study
code, data, and specifications

AI agents adjust pre-validated code re-used from the input studies

Continuous code, data, and results validation automated with traceability and AI, guarded by experts

Patient-level differences automatically detected and reported

Valid ISS/ISE Reviewer’s Guides and Define-XML generated in minutes

Without Verisian

Pooling waits for input study DBLs and final results

Input study specifications incomplete and ambiguous

Pool planning from incomplete, ambiguous specifications, with data issues found during implementation

Thousands of lines of code written from scratch, then double programmed for validation

Double programming, risk-based validation on the critical path delaying time-to-market

Patient-data differences spotted by chance (if at all) across multiple studies

ISS/ISE Reviewer’s Guides and Define-XMLs assembled manually during submission timelines

Regulatory review

With Verisian

Augmented submissions give reviewers full transparency into code, data and results, speeding up review

Fewer health authority questions, as reviewers resolve them directly using Verisian traceability

Health authority questions answered in minutes, with every question and answer fully traceable

Additional analyses created and validated in minutes using AI agents under expert guidance

Without Verisian

Reviewers working from static documents without full traceability

High volume of health authority questions due to analysis complexity

Large number of health authority questions take days or weeks to answer

Requests for additional analyses take days or weeks to answer

Thank You