AI-assisted IND gap analysis
MedNova analyzes your IND against regulatory requirements and your supporting evidence, surfacing potential gaps and reviewer-style questions before submission.
If an algorithm reads your filing first, run one before you send it.
Why now
“Submissions that are disorganized, that bury key data, or that require reviewers to synthesize inputs may fare poorly when the first pass is algorithmic.”
“Elsa's algorithms may detect patterns or anomalies that human reviewers might overlook, leading to new types of information requests or queries that are more data-driven or granular.”
Their advice: anticipate likely questions and supplement your initial filing with responsive data before submission. Agency officials at the 2026 FDLI Annual Meeting were just as direct: expect AI to be applied to your submission and prepare accordingly, per Skadden's meeting report.
FDA officials report AI-assisted review now completes certain scientific review tasks in minutes that once took days. Gaps, weak justifications, and inconsistencies that a team missed surface as amendments and information requests after filing. FDA press announcement, May 2025
Counsel advising sponsors on FDA's AI systems expect less opportunity for iterative clarification with a reviewer. Ambiguity a human might have resolved in a call can arrive instead as a formal information request, on the agency's clock.
Normally-occurring animal study findings. Cross-references that almost line up. A conclusion any reader would infer that the document never actually states. Charles River Laboratories' chief scientific officer for discovery and safety assessment reports the CRO is getting questions on background findings from pharma companies after their regulatory submissions, per Drug Discovery News. Your team stopped noticing these years ago.
The response
An example IND section, reviewed the way MedNova reviews it: each finding mapped to its regulatory basis, the evidence located or not located, and the question a reviewer is likely to ask.
FDA has cited failure to demonstrate container-closure integrity throughout the labeled shelf life in a CGMP warning letter issued after an inspection of a sterile drug manufacturer. FDA warning letter to Lernapharm (Loris) Inc., 2018
Provide data demonstrating container-closure integrity throughout the proposed shelf life and describe the CCIT method, including its qualification/validation and detection capability.
Provide qualified/validated CCIT methodology and supporting data covering the proposed container-closure configuration and shelf life.
Provide the microbial retention study validating your sterilizing filter, including the challenge organism and worst-case conditions.
Provide a microbial retention study and filter validation data under worst-case process conditions.
Did your media fill cover worst-case conditions, including interventions and shift changes?
Run media fills covering worst-case conditions (interventions and shift changes) and provide the reports.
Identify the NOAEL in each species tested, state which species was selected as most relevant, and show the scaling and safety factor used to derive the proposed starting dose.
Provide the NOAEL for each species with supporting toxicokinetic exposure data, and document the human equivalent dose derivation and safety factor applied.
Define dose-limiting toxicity for this protocol, and specify the criteria for escalating, holding, de-escalating, and stopping dosing at both the individual subject and study level.
Add an explicit DLT definition, cohort escalation and de-escalation rules, an observation interval between cohorts, and individual and study-level stopping criteria to the protocol.
Material findings are source-linked and traceable.
"Not found" means not located in the supplied documents, not that the evidence does not exist.
The platform
Upload an IND section. MedNova runs the gap analysis, maps each finding to its regulatory basis, retrieves your supporting evidence, and recommends remediation, all in one traceable loop.
What is missing, inconsistent, unsupported, or insufficiently explained.
The applicable regulation, guidance, or standard, with its citation.
The supporting data in your documents, or what was not located.
A predicted question an agency reviewer may raise.
A recommended action, from a writing fix to additional data.
MedNova does not:
How it works
MedNova separates deterministic checks from AI-assisted regulatory reasoning, and it tells you the difference between a writing fix and a missing-data gap.
Select an IND section and relevant supporting source documents.
Deterministic rules and AI-assisted regulatory reasoning evaluate completeness, consistency, support, and applicable regulatory expectations.
MedNova traces findings to supporting source documents and distinguishes documentation deficiencies from missing supporting data.
MedNova generates potential reviewer questions and recommended remediation, with traceability to the underlying finding and regulatory basis.
Qualified regulatory professionals review and approve material before it is used in a submission.
No material gap identified against the evaluated rule set and documents reviewed.
A potential deficiency appears addressable with existing evidence, clarification, or improved filing language.
The gap cannot reasonably be resolved by rewriting alone and appears to need additional data, analysis, or validation.
“Clear” does not mean FDA will accept or approve the submission.
Data sources
Regulatory teams hunt across a dozen databases to back a single claim. MedNova queries them all in one pass, so the evidence is already waiting at the gap.
registered studies and posted results
approved products with review packages
biomedical citations and abstracts
Three sources cover most of what a reviewer looks for. MedNova queries all three, plus your own documents, FDA guidance, and licensed regulatory and drug-development intelligence, in one pass.
FDA decision letters. On the roadmap: analysis of the Complete Response Letters FDA has begun publishing and of refuse-to-file letters that have become public, to sharpen what a reviewer will ask.
Counts are public figures as of August 2026. Licensed regulatory and drug-development intelligence extends the public sources. Not every gap needs every source; MedNova queries them all and cites only what applies.
Compliance & security
Programs
A scored gap analysis on one module. A low-commitment way to see exactly where a program stands before filing.
Schedule a demoA credible FDA path for pre-seed to Series A teams, scoped to one lead program, with expert review and eCTD-ready output.
Project-based engagements or ongoing retainers for teams with continuous submission and gap-analysis needs.
A short session to walk your team through MedNova and find the right first program.
Louis Lalas, Founder & CEO, Amia.ai