Data integrity
Pass ✓
Stop losing study time to data cleanup, report drafting, and last-minute QC.
Preclinical Studio gives an AI workspace for raw-data processing, scientific writing, QA/QC review, and traceable evidence, AI drafts and checks the repetitive work; scientists keep the judgment, review, and approval.
Ask the Virtual SD, or drop a file…
XM-441 registered
chapter-drafter running
evidence-linker queued
60%Reduction in report production time
9+Virtual labs across preclinical study types
100%Audit-trail coverage on every agent action
0Unsourced claims — every conclusion traced
A continuous study lifecycle
From protocol intake to inspection-ready report
From day one, a Virtual SD and six specialist agents work inside the study — so protocols, data, drafts, QC findings and reviews stay connected throughout, instead of waiting on scarce experts to review scattered files at the end.
AI handles the repeatable work. Experts make the scientific calls. The system preserves the evidence.
BenchTeam · your virtual study team
BenchTeam: One Virtual SD. Six Specialist Agents.
Together, they replace the expensive part-time expert functions every CRO needs but cannot always staff on demand.
One Virtual SD
Virtual Study Director
Coordinates the study workflow, tracks status, manages versions, routes issues across agents, and prepares the package for human sign-off.
-
Routes issues, closes the loopRoutes issues across agents, follows every review to closure, and prepares the study package for human confirmation and delivery.
-
Never replaces the Study DirectorKeeps the work organized so your human SD can focus on scientific judgment, quality responsibility, and final sign-off.
Agent 01 of 06
Data Curator
Structures the dataTurns protocol information, raw files, tables, figures, and experimental data into traceable study assets by cleaning, structuring, converting, and summarizing them.
Agent 02 of 06
Statistical Analyst
Checks the analysisReviews analysis logic, statistical outputs, endpoint consistency, trends, and whether results support the intended scientific conclusions.
Agent 03 of 06
Scientific Writer
Writes the reportTurns experimental data, study context, evidence, and confirmed conclusions into clear, professional, delivery-ready scientific documents.
Agent 04 of 06
QA Reviewer
Reviews before sign-offReviews workflow completeness, evidence chain integrity, claim support, narrative quality, unresolved risks, and delivery readiness.
Agent 05 of 06
QC Specialist
Checks consistencyChecks data, tables, references, numbers, units, terminology, formatting, cross-section consistency, and version alignment.
Agent 06 of 06
Regulatory Specialist
Tracks the rulesTracks relevant OECD, FDA, ICH, CDE, and regional expectations, identifies evidence gaps, and prepares regulatory checklists and audit/XAI packages.
Product scenarios
Built for the moments that actually matter
The moments in a study where the cost is real — and where a Director changes the day. Expand any scenario to see it in the product.
“The experiment wrapped weeks ago. The report still isn’t written.”
Preclinical Studio starts QC and evidence assembly the day study data arrives — so the report isn’t waiting for someone to manually reconcile protocol, raw data, figures and conclusions weeks later.
- QC and evidence checks start the day study data lands — not at write-up
- Protocol, raw data, figures and conclusions reconciled continuously
- No multi-week manual reconciliation before the report can begin
XLSLive
tk_xm778_groups1-4.xlsx
received 09:14 · auto-QC started same day
QC & evidence assembly as data lands — not at write-up
Figures bound
12 / 12
Tables checked
8 / 8
Unit consistency
OK ✓
Evidence gaps
1 flagged
Report readiness
96%
Routine checks → digital specialists
Senior scientists keep judgment, interpretation and sign-off
1Done
Data integrity checks
Ranges, outliers and missing values verified
2Active
Table reconciliation
Cross-checking tables against source datasets
3Active
Figure references
Matching every figure call-out to its asset
4Active
Unit consistency
Normalising units and terminology across sections
“Your best experts are stuck reconciling tables.”
Data checks, table reconciliation, figure references, unit consistency and repeated template work move to digital specialists — while senior scientists focus on judgment, interpretation and sign-off.
- Routine, high-liability checks handled by digital specialists
- Senior scientists keep judgment, interpretation and sign-off
- Less low-joy template work, fewer manual reconciliation errors
“The auditor points at one sentence. Where did this number come from?”
Every claim is linked to its source data, calculation, figure, reviewer action and audit trail — giving QA teams and sponsors a defensible path from conclusion back to evidence. Select a highlighted conclusion to inspect its source.
Excerpt · Conclusions
Under the conditions of this 13-week study, across all dose groups. Body-weight gain was , correlating with minimal hepatocellular changes. Based on the overall profile, .
Each highlighted claim is bound to its evidence — select one to inspect the source.
Source · In-life clinical observations
Daily mortality & clinical records
Mortality conclusion traces to the complete in-life observation log — twice-daily checks across all 60 animals for the full 13-week dosing period.
dataset · in_life_obs.csv · 60 animals
| Group | Dose | N | Mortality |
|---|---|---|---|
| G1 Control | 0 | 15 | 0 / 15 |
| G2 Low | 100 | 15 | 0 / 15 |
| G3 Mid | 300 | 15 | 0 / 15 |
| G4 High | 1000 | 15 | 0 / 15 |
Source · Body-weight dataset + Figure BW-3
Group mean body weight · week 13
The high-dose reduction traces to the body-weight dataset and its derived figure. Group G4 shows a statistically significant decrease versus control.
dataset · body_weight.xlsx — fig BW-3
| Group | Dose | Mean BW (g) | vs Ctrl |
|---|---|---|---|
| G1 Control | 0 | 412 | — |
| G3 Mid | 300 | 404 | -1.9% |
| G4 High | 1000 | 368 | -10.7%* |
Source · Integrated NOAEL determination
NOAEL · 300 mg/kg/day
The NOAEL traces to the integrated review across endpoints — body weight, clinical pathology and histopathology — with the high dose excluded on the flagged body-weight finding.
derivation · noael_review.json · 4 endpoints
| Dose | BW | Clin-path | Call |
|---|---|---|---|
| 100 | OK | OK | NOAEL |
| 300 | OK | OK | NOAEL |
| 1000 | -10.7% | Adaptive | LOAEL |
Draft vs delivery
Beyond an 80% Draft.
Beyond an 80% Draft.
Built for Expert-Signed Delivery.
General-purpose AI is good at drafting fast. Regulated R&D requires something deeper: claims tied to source evidence, deterministic calculations, human review gates, version-controlled workflows, and audit packages ready for export.
01
Every number has a source.
02
Every conclusion has a reviewer.
03
Every delivery has an audit trail.
Explainable AI — try it
Click any highlighted claim in this report excerpt. The panel shows why the model wrote it, the data it rests on, and where a human still decides — the heart of the XAI review, on its own.
Built for GLP
Not adapted for GLP. Built for it.
Every architectural decision was made with GLP inspection readiness as the first constraint — not an afterthought.
Ref: OECD ENV/MC/CHEM(98)17
OECD GLP Principles
Built around Principle 8. Aligned to 1–10.
Study Plan structure, personnel records, amendment procedures, and audit trail requirements are mapped directly to OECD GLP Principles. Not interpreted. Not approximated. Mapped.
GLP Principles 1–10 · Full alignment
Ref: 21 CFR Part 11 — FDA
FDA 21 CFR Part 11
Part 11 isn't a checkbox. It's the architecture.
Closed system controls, granular access permissions, cryptographic audit trail, and e-signature workflows are built into the platform foundation — not added on top of a general tool.
eSig · Closed system · Audit trail
Ref: Directive 2004/10/EC
EU GLP & Data Residency
Your data doesn't leave your region.
US, EU, and APAC deployments available. Data residency is enforced at the infrastructure level — not configured per request. No cross-border transfer without explicit authorization.
US · EU · APAC · Enforced at infra
Voices from the lab
What burns the budget isn't the writing. It's the re-checking.
The experiment finishes — and the real grind begins. Here's what changed once directors, QA leads and sponsors put the report loop on Preclinical Studio.
"The experiment used to end and the report never would — re-reconciling data scattered across a dozen places, over and over. Now it auto-drafts from the confirmed protocol data and every number carries its own evidence. The report lands in days, not weeks — and the payment lands with it."
Unlocked the ~60% of contract value tied to the report
Dr. Julien Mercier
VP, Toxicology Operations · mid-size CRO
"It's high-liability, low-joy work, and the people who can do it are expensive and scarce. The platform drafts the data-woven sections and runs QC before anything leaves the building, so my specialists stop reconciling and start reviewing — and reports clear the three-party check the first time."
6–8 week revision cycle collapsed · capital freed
Dr. Lena Brandt
Director of QA & Medical Writing · GLP facility
"Everyone asks whether a stronger model just replaces this. It won't — a bigger model writes faster, it doesn't make a report defensible. Preclinical Studio's XAI shows why every sentence is trustworthy, and controlled recursive learning turns each real project into a reusable asset: higher QC pass rates, less rework, faster client fit. That compounding is the moat."
Every project compounds — the moat widens with use
Dr. Anders Holm
CTO · CRO group, 12 facilities
Pricing
Choose the scale that fits your lab
All plans include the full study-type catalogue. The difference is activation depth, governance tier and Credits volume.
Starter
$79/ moSave 10%
Evaluate the workflow with non-production studies and demo data.
- Full study-type catalogue visible
- Demo Profile + Archive Mode
- Limited Demo Credits — non-production only
Most popular
Professional
$599/ moSave 10%
Team-level production. Activate your reporting pipeline.
- Up to 5 Active Profiles
- Standard Credits + top-up
- QC + evidence chain + multi-user
Enterprise
$2,999/ moSave 10%
Company-wide standards. Scale activation across departments.
- Unlimited Active Profiles
- Large Credits pool + tiered discounts
- QA Review Pack + cross-project audit
Private Deploy
Custom
Data never leaves your network. Maximum compliance and regulatory fit.
- On-premise / air-gapped deployment
- Deep LIMS / ELN / QMS integration
- SLA + on-site / remote co-build support
All plans include access to the full study-type catalogue. Production capability requires separate activation.
Study-type activation fees are tiered by complexity: Type A $500 · Type B $1,000 · Type C $2,500 · Type D $7,500+
In Silico Workforce
See how one real study becomes audit-ready
We will show how the Virtual SD sets up the study, activates digital specialists, builds the evidence chain and prepares the QA review package.