The updated Elma Cortex platform now enables AI agents to process data from scans, photographs, and PDF files using OCR or multimodal VLM. This allows for automated extraction of text and financial data from invoices and waybills directly into business workflows.
The platform now supports audio transcription through customizable ASR providers, allowing companies to select services based on specific quality and data residency requirements. Agents can generate meeting summaries, identify action items, and assign tasks based on the transcribed content.
New guardrails provide input and output filtering for AI interactions, including LLM-based tone checks, regex-based sensitive data detection, and external webhook verification. Users can also define custom masking rules for internal identifiers and document codes.
The search functionality now features incremental indexing for faster synchronization and improved document context handling. AI agents are now integrated into Elma365 Procurement and Service Desk solutions to assist with price analysis, request classification, and support tasks.
