H1ST-AI-STAT-004 · Task 3.2
The 100th TLF shouldn't be programmed by hand
When a shell changes on Friday, the last thing you want is to re-program two hundred outputs by Monday. This agent generates the full TLF package directly from ADaM, formatting each output to your approved mock shells and running automated QC against them. The batch-programming grind of building and checking hundreds of outputs collapses into a fast, consistent, reviewable pipeline.
The grind you know
Your submission TLF package can run to hundreds of outputs, each one programmed and QC'd by hand against its mock shell, and every late data cut or shell revision cascades across all of them at once. It's why TLF production is the longest pole in your timeline, and why formatting inconsistencies and rework keep finding their way back to your desk.
Part of the Statistical Programming family
SDTM/ADaM, Define.xml, TLFs & analytics.
When a shell changes on Friday, the last thing you want is to re-program two hundred outputs by Monday. This agent generates the full TLF package directly from ADaM, formatting each output to your approved mock shells and running automated QC against them. The batch-programming grind of building and checking hundreds of outputs collapses into a fast, consistent, reviewable pipeline.
Choose from the template library or load your approved mock shells to define the study's TLF set.
The agent produces every table, listing, and figure from the ADaM datasets, applying study titles and footnotes.
Outputs are QC'd against the mock shells, then rendered as a full package in RTF, PDF, and HTML for review.
Capabilities
Ships with standard safety and efficacy tables, listings, and figures covering demographics, disposition, AEs, labs, and exposure.
Builds each output directly from the corresponding ADaM dataset so numbers trace straight back to analysis data.
Compares generated outputs against approved mock shells and flags formatting, footnote, and population mismatches.
Renders the same output to RTF, PDF, and HTML with consistent titles, footnotes, and pagination.
In practice
Where the agent shows up in the day-to-day of a live trial — the moments the grind usually lives in.
A mock shell is revised late Friday and hundreds of outputs need to reflect it by Monday. The agent regenerates the affected tables directly from ADaM and re-runs mock-shell QC, turning a weekend of re-programming into a single reviewable batch.
A footnote and population count on one AE table no longer match the approved shell. Automated mock-shell QC flags the mismatch in the QC report, so it's fixed before the output reaches your reviewer's desk.
Your team needs the same demographics table in RTF for the CSR, PDF for review, and HTML for the portal. The agent renders all three with consistent titles, footnotes, and pagination from a single generation run.
Proof
Figures shown are pre-launch targets based on internal benchmarks, not guaranteed outcomes.
Pinnacle 21
Validation
Transforms raw and collected clinical data into submission-ready SDTM and ADaM datasets across all standard domains with controlled terminology applied.
Learn moreDrafts study and results narratives and executive summaries grounded in your actual TLFs and datasets, ready for statistician review.
Learn moreAssembles ICH E3-structured Clinical Study Reports that pull results directly from your TLFs and datasets, with full track-changes review support.
Learn morePeace of mind
Every output is generated inside a validated, audit-ready platform, kept under human-in-the-loop control, and mapped to the regulatory and CDISC standards this agent supports.
Walk through it on your own workflow with a clinical-trials expert — no pressure, no obligation, and honest answers, including on the limits.
