From source to submission. Nothing in between.
KERNEL covers the full cycle: scientific aggregation, referenced drafting, human validation, submission. Every step traced.
Your Scientific Data Warehouse
PubMed, EMA, ANSM, ClinicalTrials.gov and your proprietary data, aggregated, structured, and queryable in seconds.
Step 01 / 05
Aggregation
PubMed, EMA, ANSM, ClinicalTrials.gov, Cochrane, FDA, and your proprietary data: KERNEL continuously connects and collects thousands of documents from dozens of sources.
- 30+ connected sources
PubMed, EMA, FDA, ANSM, Cochrane, ClinicalTrials.gov and more
- Continuous automated collection
New publications detected and indexed in real time
- Proprietary data included
Internal reports integrated alongside public regulatory sources
Step 02 / 05
Structuring
Every document is indexed across dozens of dimensions: compound, indication, publication type, source, clinical phase, date, and regulatory region.
- Multi-dimensional indexing
Compound, indication, phase, source, date, region and more
- Hundreds of categories
Granular, automatic organization with no manual effort
- Faceted navigation
Filter and cross-reference dimensions to find exactly what you need
Step 03 / 05
Research Workspace
Ask your question in natural language. KERNEL creates a research workspace with filtered results, grouped by source, ready to explore.
- Natural language search
Ask your question as you would a subject-matter expert
- Dedicated research workspace
A structured workspace for each research question
- Results grouped by source
PubMed, EMA, FDA, ANSM: each source identified and scored
Step 04 / 05
Prioritization
Within your workspace, results are ranked by evidence level, regulatory weight, and clinical relevance. KERNEL automatically surfaces what matters most.
- Evidence level ranking
Level I (meta-analysis) to Level V (expert opinion), automatically ranked
- Integrated regulatory weight
EMA, FDA, and ANSM guidelines weighted by regulatory authority
- Clinical relevance scoring
Every source evaluated for relevance to your specific question
Step 05 / 05
In-depth Analysis (PICO)
Click any source to access its full analysis. Population, Intervention, Comparison, Outcome: every element extracted automatically.
- Automatic PICO extraction
Population, Intervention, Comparison and Outcome identified
- Evidence-based medicine standards
Every study deconstructed according to EBM standards
- Actionable structured data
Extracted elements feed directly into your documents
Your Scientific Data Warehouse
PubMed, EMA, ANSM, ClinicalTrials.gov and your proprietary data, aggregated, structured, and queryable in seconds.
Step 01 / 05
Aggregation
PubMed, EMA, ANSM, ClinicalTrials.gov, Cochrane, FDA, and your proprietary data: KERNEL continuously connects and collects thousands of documents from dozens of sources.
- 30+ connected sources
PubMed, EMA, FDA, ANSM, Cochrane, ClinicalTrials.gov and more
- Continuous automated collection
New publications detected and indexed in real time
- Proprietary data included
Internal reports integrated alongside public regulatory sources
From first draft to submission, without losing the source.
KERNEL writes the first draft. Your teams verify, annotate, approve. Every sentence stays traceable back to its original source.
Step 01 / 05
Initial Drafting
KERNEL generates a structured first draft from your trial data. The document follows regulatory templates and automatically populates all required sections.
- Structured draft in minutes
No blank page — a complete document from the start
- Built-in regulatory templates
ICH, EMA, FDA: correct format from the first version
- Clinical data injected automatically
Your trial results feed directly into the narrative
Step 02 / 05
Claim Detection
Every scientific claim is identified and flagged as traceable, ready to be sourced and verified.
- Automatic identification
Every scientific claim detected and highlighted
- Zero claims missed
Exhaustive coverage, even in dense technical sections
- Traceability counter
See at a glance how many claims still need sourcing
Step 03 / 05
Linked References
For each claim, KERNEL retrieves the primary source from your data and the scientific literature.
- PubMed, EMA, FDA in real time
Automatic search across regulatory databases
- Direct link to the source
One click to verify any claim
- Relevance scoring
Every source evaluated and ranked automatically
Step 04 / 05
Evidence Audit
KERNEL assesses the strength of your references: regulatory coverage, gap recommendations, and compliance checks.
- Coverage analysis
Every claim backed by a source of sufficient quality
- Smart suggestions
Recommendations for additional supporting references
- Pre-submission compliance
EMA and FDA checks completed before human review
Step 05 / 05
Review & Validation
Reviewers comment and approve directly within the document. Export is ready for submission in one click.
- Collaborative review
Real-time comments and approvals from your team
- Multi-reviewer sign-off
Clear audit trail of who approved what
- Submission-ready export
One click, compliant format, ready to send
From first draft to submission, without losing the source.
KERNEL writes the first draft. Your teams verify, annotate, approve. Every sentence stays traceable back to its original source.
Step 01 / 05
Initial Drafting
KERNEL generates a structured first draft from your trial data. The document follows regulatory templates and automatically populates all required sections.
- Structured draft in minutes
No blank page — a complete document from the start
- Built-in regulatory templates
ICH, EMA, FDA: correct format from the first version
- Clinical data injected automatically
Your trial results feed directly into the narrative
6.2 EVALUATION CRITERIA
The evaluation criteria for this Phase III study were defined in accordance with EMA regulatory requirements and ANSM recommendations for oncology trials.
The efficacy of compound XR-442 was demonstrated in advanced hepatocellular carcinoma with an objective response rate of 34.2%.
The recommended dosage is based on dose-response data in accordance with ICH E4 guidelines.
RECIST v1.1 criteria will be used for tumor response assessment, with blinded centralized review.
Submit before your competitors on the same molecule.
See how teams are going from data to dossier in a fraction of the time.