AI Data Review
A full run reviewed in seconds - not a full day.
Upload run exports and study documents. A rule engine checks each run against ICH M10 or your SOP; an AI reviewer explains what needs attention - quoting the documents it relies on.
What is checked in every run
Calibration
Standards per level, LLOQ and ULOQ, minimum levels passing.
Quality controls
Per level and overall, dilution QCs, QC count against the protocol.
Blanks & carryover
Analyte and IS interference in blanks after high samples.
Internal standard
IS response window and missing IS, retention-time shifts.
Repeats
Above ULOQ, IS response, failed run - each with reason and action.
ISR
Selection, pairing and evaluation of incurred sample reanalysis.
Data integrity
Duplicates, missing IDs, unparseable or ambiguous values flagged.
Protocol
Calibration and QC levels, range and dilutions against the study set-up.
Run files, documents and scans - together
Several run files at once, every worksheet read. Add the protocol, method, SOPs, investigations, sample manifest, chromatograms and instrument logs - PDF, Word, Excel or scans.
- Scans read on the server, confirmed by the uploader
- Samples missing from the manifest found automatically
- Repeats without a documented reason flagged
- Protocol LLOQ, ULOQ and IS compared with the set-up
An AI reviewer that has to show its sources
The AI reads the relevant passages of your documents and the SOPs and guidelines linked to the study.
- Every quote checked word for word against the document
- Unverifiable statements removed and listed
- Rule-engine verdicts are final - the AI is advisory
- Local AI by default; cloud AI only with QA approval
From upload to approval
- Upload
Run exports and study documents; the system reads every sheet and lists what it used. - Check
The rule engine produces run verdicts and findings in seconds; document cross-checks run automatically. - Review
The scientist dispositions every finding; the AI review points to evidence. - QC check
Second-person verification of each finding; queries go back to the scientist. - QA approval
Review and approval with electronic signature; the study dashboard shows every run.
See AI Data Review on your own run files
A 45-minute walkthrough of method development, AI Data Review and the compliance controls - with your own run files.