Why RA teams are buying software now
Regulatory affairs has long run on spreadsheets, shared drives and email. Three forces are ending that in 2026.
Regulators went digital. FDA requires eSTAR for 510(k)s (since October 1, 2023) and De Novo requests (since October 1, 2025), submitted through the CDRH Portal (FDA). In Europe, the first four EUDAMED modules became mandatory on May 28, 2026 (HPRA), with device registration, including legacy devices, due by November 28, 2026 (TÜV NORD).
Regulators use AI. FDA launched its generative AI tool Elsa for staff in June 2025 (RAPS) and added agentic AI capabilities in December 2025 (PharmTech). Inconsistencies across a submission are easier to find on the reviewer’s side, so they need to be found first on yours.
Data quality is weak. In Veeva’s 2025 MedTech benchmark, a vendor survey, 56% of companies planned to adopt regulatory information management (RIM), yet only 17% rated their regulatory data quality as excellent (Veeva).
For how RIM software works, how it differs from document management and how to choose, see RIM Software: Regulatory Information and Document Management Explained.
Categories of medical device regulatory software
| Category | What it does | Typical outputs |
|---|---|---|
| Regulatory information management (RIM) | Single source of truth for products, registrations, submissions, commitments and correspondence by market | Registration status, renewal calendar, submission history |
| Submission authoring and publishing | Templates, content reuse and assembly for 510(k), De Novo, PMA and international dossiers | eSTAR-ready content, technical files |
| UDI and registration data management | Manage device identifiers and attributes; push to GUDID and EUDAMED | UDI-DI records, EUDAMED registrations |
| Regulatory intelligence | Track new regulations, guidance, standards and enforcement by market | Alerts, impact assessments |
| Labeling and claims management | Control labels, IFUs and promotional claims against cleared indications | Approved labeling, claims matrix |
| Change assessment | Structured workflows to decide whether a change needs a new submission | Documented regulatory impact assessments |
| AI document assessment | Check submissions, technical files and procedures against requirements, with cited evidence | Gap findings, consistency checks |
Most RA teams need two or three of these, not all seven. A small company might pair a lightweight RIM with AI document assessment; a global manufacturer may need full RIM integrated with PLM and the eQMS.
What to watch: costs and cadence
Software choices should reflect what is changing in fees and timelines. FY2027 FDA fees took effect October 1, 2026, including $28,653 for a 510(k) and $13,785 for annual establishment registration (Federal Register). The draft MDUFA VI agreement, due for reauthorization by September 30, 2027, proposes a $2,000 pre-submission fee (Exponent; RAPS). When each FDA interaction costs more, tools that improve first-time quality pay back faster.
How to choose regulatory software
| Criterion | Questions to ask | Why it matters |
|---|---|---|
| Data model | Does it model products, device families, UDI-DIs, markets, registrations and submissions the way medtech works? | Pharma-first RIM tools can fit devices poorly |
| Regulator connections | Does it support eSTAR content, GUDID and EUDAMED data exchange? | Manual re-entry creates mismatches |
| Integration | Can it connect to PLM, eQMS and labeling systems? | Regulatory data originates elsewhere |
| Change assessment | Does it support documented “new 510(k)?” and MDR significant-change decisions? | FDA inspects change control as a QMS area (FDA CP 7382.850) |
| Regulatory intelligence | Which markets and sources are covered, and how fast? | Guidance and fee changes affect plans |
| Validation and Part 11 | Audit trails, e-signatures, vendor validation package | Required if records support FDA obligations |
| AI transparency | Do AI outputs cite sources? Can reviewers see why? | Trust and verifiability |
| Fit for size | Can a two-person RA team run it? | Over 80% of US device firms have fewer than 50 employees (AdvaMed) |
AI in regulatory software: promise and limits
AI features are now standard in RA tools: drafting submission sections, summarizing guidance, comparing predicates, checking consistency. RA professionals are cautious, with reason. In RegASK’s 2024 report, a vendor survey, “39% of regulatory affairs respondents are using AI,” but of those permitted to use ChatGPT for compliance guidance, “only 22% fully trusted the responses” (RegASK via Business Wire).
The trust problem is solvable when tools are designed for verification:
- Cite, don’t summarize. Every AI statement about a regulation or a document should link to the exact passage.
- Assess against explicit requirements. Checking a document against a named clause list is more reliable than open-ended questions.
- Keep people accountable. AI recommends; RA decides and records the rationale.
- Validate proportionately. If AI outputs feed QMS or regulatory records, FDA’s Computer Software Assurance guidance applies its risk framework to “artificial intelligence/machine learning tools” (FDA).
Implementation roadmap
- Inventory regulatory data. List products, UDI-DIs, registrations, submissions, commitments and where each lives today.
- Fix the urgent deadline first. For many companies in 2026, that is EUDAMED device registration by November 28, 2026.
- Pick a system of record. Decide whether RIM, PLM or the eQMS owns each data element.
- Add document assessment. Use AI with citations to pre-screen submissions and technical files before filing.
- Connect intelligence to action. Route new guidance and standards to impact assessments and change control.
- Validate and train. Apply a risk-based CSA approach and train RA staff on reviewing AI outputs.
Medical device regulatory software checklist
- Regulatory data inventory completed by product and market
- EUDAMED and GUDID data reconciled with labeling
- System of record defined for each regulatory data element
- Submission content reusable and eSTAR-ready
- Change assessments documented and linked to change control
- Regulatory intelligence sources and owners defined
- AI features cite sources and require human approval
- Software validated proportionate to risk; Part 11 controls confirmed
- Vendor qualified under purchasing controls
Pre-check submissions with cited evidence
AskDona assesses submission sections, technical files and procedures against regulatory requirements and cites the passage behind every finding, so RA teams can catch gaps and inconsistencies before a reviewer does.
Run a sample regulatory document assessment on one submission section. AskDona shows the evidence; your RA team decides.
FAQ
What is medical device regulatory software? Software that supports regulatory affairs work: RIM, submission authoring, UDI and registration data, regulatory intelligence, labeling and claims control, change assessment and AI document assessment.
What is RIM? Regulatory information management: a system of record for product registrations, submissions, commitments and correspondence across markets.
Does FDA require specific submission software? FDA requires eSTAR for 510(k) and De Novo submissions and provides the CDRH Portal for sending them (FDA); other tools are optional.
Is AI in regulatory software reliable? Only when outputs cite their sources and qualified people review them. Survey data shows many RA professionals do not fully trust general-purpose AI answers.
References
All pages opened and checked by October 13, 2026.
| Source | Type | Date | Used for |
|---|---|---|---|
| FDA: CDRH Portal | Regulator | Current | eSTAR |
| HPRA: EUDAMED | Regulator (EU) | 2026 | EUDAMED |
| TÜV NORD: EUDAMED | Notified body | 2026 | Registration deadline |
| RAPS: Elsa launch | Professional body | Jun 2025 | FDA AI |
| PharmTech: Agentic AI at FDA | Trade press | Dec 2025 | FDA agentic AI |
| Veeva: MedTech benchmark | Vendor survey | 2025 | RIM adoption, data quality |
| RegASK via Business Wire | Vendor survey | Nov 14, 2024 | AI trust |
| Federal Register: FY2027 user fees | Regulator | 2026 | Fees |
| Exponent: Draft MDUFA VI | Consultancy | Aug 14, 2026 | Pre-sub fee |
| RAPS: MDUFA VI draft | Professional body | Jul 8, 2026 | Reauthorization |
| FDA: CP 7382.850 | Regulator | Feb 2, 2026 | Change control |
| FDA: CSA guidance | Regulator guidance | Feb 3, 2026 | AI tool validation |
| AdvaMed: Job creation | Trade association | 2021 data | Firm size |