BiVelio AI Capabilities
Artificial intelligence in BiVelio is not an add-on: it is a cross-cutting layer that powers every module of the platform. From automatic message classification to data extraction from documents, AI works to reduce repetitive tasks and amplify your team’s capacity.
AI Pillars
Autonomous agents
Operate in the background, reacting to system events without human intervention. They subscribe to specific events (new message, case created, SLA at risk) and execute actions automatically.
Invocable agents
Executed on demand from workflows or API calls. They process input data, apply intelligent logic and return structured results to the flow that invoked them.
OCR and document extraction
Processing of scanned or photographed documents: the system detects text, extracts it and, optionally, an AI agent interprets the content to obtain structured data.
Fernando — Executive assistant
Fernando is the executive-level AI agent that operates as a personal assistant for the leadership team. It checks the calendar, prepares briefings and manages coordination tasks.
Agent Marketplace
BiVelio includes a marketplace with 12 pre-built agent templates, ready to activate with a single click:
- 8 autonomous agents — support, triage, follow-up, compliance, SLA
- 3 marketplace agents — lead qualifier, invoice processor, content moderator
- 2 invocable agents — data extractor, meeting prep
Credits system
AI agents consume credits with each execution. The credits system allows you to control the cost and usage of AI:
| Plan | Included credits |
|---|---|
| Free trial | Evaluation credits for 14 days |
| Starter | 0 credits/month (purchase additional packs) |
| Pro | 250 credits/month included |
| Billing Add-on | +100 credits/month |
Additional packs available: 100 credits (89 EUR) or 500 credits (399 EUR).
BiVelio includes an AI prompt generator that converts brief, conversational instructions into professional, optimized system prompts. This allows any user, without experience in prompt engineering, to configure effective agents.
AI and security
- Agents only access data from the authorized workspace (multi-tenant RLS)
- Each agent execution is recorded in the activity log
- System prompts and responses are stored encrypted
- Data is never shared between workspaces or used to train models