Beams
Beams, a Berlin start-up, built Safety AI to sort and analyse airline safety reports and spot risks early. Comply365 acquired it in August 2025.
Founded in 2018 with 11 to 50 staff, it made CoAnalyst, which processes free-text reports from pilots, cabin crew and engineers, and CoReporter, which helps crews file complete reports. It claims reports are handled up to 20 times faster in more than 60 languages. The products now sit within Comply365, which serves more than 450 customers, and no airline deployment has been independently confirmed.
No verified deployments yet
Beams is profiled in the companies register, but no deployment has been verified against published evidence. Records are added as evidence is reviewed and graded.
Products & Solutions
2 productsCoAnalyst
An AI-powered safety report processing and risk management module that automates hazard identification, event classification, report summarization, and hazard trending within aviation Safety Management Systems (SMS). Integrated with ASQS iQSMS and now marketed under Comply365 following the August 2025 acquisition.
Aviation safety teams face growing volumes of free-text safety reports from pilots, cabin crew, and ground personnel that require manual reading and analysis to identify hazards and trends. This creates bottlenecks in risk identification, delays compliance reporting, and makes proactive risk management difficult at scale.
- Automated hazard and event detection from incoming safety reports
- Auto-classification of reports, hazards, and events
- AI report triaging and quick filtering for high- and low-risk reports
- Report summarization and AI search
9 more features
- Hazard correlation mapping (highlights connected hazards and combined likelihood of occurrence)
- Hazard and event trending across operational data
- Event occurrence prediction and auto root-cause analysis
- Historical data matching for similar past events
- Cross-divisional dashboards and top risk alerts
- AI-generated safety promotions and investigation dashboards
- AI Spotlight feature for high-level top-management briefing summaries
- Regulator event-tagging and taxonomy auto-detection
- Analysis of missing information in reports
Technical specifications (7)
- Infrastructure: AWS (Lambda functions, Amazon RDS PostgreSQL with PG Vector extension, SQS queues)
- Backend: Python, FastAPI
- Embeddings: Azure OpenAI text-embedding-3-large, 2048 dimensions
- Vector search: PostgreSQL PG Vector (IVFFlat indexing), two-stage retrieval with reranking
- Deployment model: cloud-based SaaS; PII removal models optionally deployable on-premise at customer site
- Language support: 60+ languages (translation and multilingual report processing)
- Models: in-house and third-party ML models and LLMs
CoReporter
An AI-powered safety report submission tool for aviation operations staff. Enables report submission via voice or text in over 60 languages, with AI-powered follow-up questions that ensure completeness and accuracy of submitted data before it is sent to safety teams or authorities.
Reduces errors and missing information in safety incident reports submitted by flight crew and operations staff, improving the quality of data available for safety analysis and regulatory reporting.
- Voice and text report submission
- Support for 60+ languages
- AI-powered follow-up questions to improve data completeness
- Minimises reporting errors
1 more feature
- Enhances data accuracy through AI interrogation of missing fields
Technical specifications (3)
- Voice and text input modes
- 60+ language support
- Mobile and web integration
Profile data from the Aerostack companies register; deployment records from the verified register. Evidence grades: T1 first-party confirmation, T2 credible reporting. See Our approach for how records are graded.