Sandelia AI Employees — MIA & LEO

AI Automation

Sandelia AI Employees — MIA & LEO

Action-taking AI agents embedded inside a Belgian B2B invoicing ERP

A Belgian ERP With a Bold Vision — AI Employees, Not AI Chatbots

Sandelia is a Belgian B2B invoicing and ERP platform built around the PEPPOL network, which became mandatory for Belgian B2B invoicing in January 2026. The founder's vision goes beyond invoicing: to become the first European ERP with a full internal team of AI "employees" — agents that don't just answer questions, but actually operate the software on behalf of the business owner.

Project image
Preview
How do you put a real accountant and a real admin assistant inside a SaaS ERP — without the risk, hallucinations, or compliance gaps that come with bolt-on chatbots?
  • Small business owners on Sandelia had to learn every module themselves — invoicing, VAT, KPIs, reconciliations — across 7 languages and 10 business sectors
  • Existing "AI employee" platforms (Sintra, Limova) aren't connected to an ERP, and accounting assistants like Accountable's are chat-only with no ability to actually take action
  • Belgian GDPR, VAT legal rules, and audit requirements rule out a generic LLM wrapper — every action needs confirmation, logging, and dispute protection
  • Any AI mistake on an invoice or VAT declaration is a legal and financial liability, not a support ticket
Project image
Preview
Meet MIA & LEO — Your AI Admin and AI Accountant, Built Into the ERP

We designed and built a multi-agent AI system deeply integrated with the Sandelia platform — not a chatbot, but two specialized AI employees that take real actions inside the ERP, with enterprise-grade safeguards.

  • MIA — the administrative assistant: creates, edits, and converts invoices, quotes, and credit notes; applies correct VAT rates including EU reverse charge and export rules; reads uploaded PDFs in 7 languages; exports quarterly documents for the accountant
  • LEO — the AI accountant: answers Belgian fiscal and VAT questions, analyzes revenue, margins, and sector KPIs, detects anomalies, generates financial reports in natural language, and simulates the fiscal impact of business decisions
  • Mandatory action confirmation — every write action surfaces an inline popup showing the parsed arguments, with cancel / preview / edit / submit, so nothing happens without the owner's explicit OK
  • Full audit trail — user intent, AI interpretation, user approval, and final action are all logged for dispute protection
  • GDPR-compliant consent layer — users explicitly opt in to data leaving the EU before MIA or LEO become available
Project image
Preview

🤖 Multi-Agent AI Team

A coordinator dispatches requests across specialized agents (MIA, LEO, with HR, Marketing, Legal, and Operations on the roadmap). Inter-agent routing means MIA hands off accounting questions to LEO automatically and returns one clean answer.

Project image
Preview

✅ Action-Taking, Not Just Chatting

Agents call real tools inside the ERP — `CreateInvoice`, `ListUnpaidInvoices`, `FetchKPIs`, `ExportQuarterlyDocs`, `GenerateReport` — with role-based permissions and structured validation so the LLM decides *when* to act, but the platform enforces *what* is allowed.

Project image
Preview

🛡️ Enterprise Safeguards

Mandatory confirmation popups for every write action, read/write permission separation, full audit logging, and explicit disclaimers across the product. Agents advise; humans approve.

Project image
Preview

🔍 Belgian Accounting RAG

Curated knowledge base of Belgian VAT, fiscal, and accounting regulations. The agent tries the internal knowledge base first, falls back to web search, and only then uses general model knowledge — always citing the source, with metadata-weighted recency so regulation changes (e.g., takeaway VAT moving from 6% to 12%) are applied correctly.

Tech Stack
Built with cross-platform technologies for real-time monitoring, secure data handling, and scalable performance across web and desktop applications with multiple user roles and access levels.

AI & Agents

Python, LangGraph for multi-agent orchestration and tool routing, LangSmith for tracing and live prompt management, OpenAI API (GPT-4 / GPT-5 class) for reasoning.

Backend

FastAPI for the async API layer, Celery + Redis for background tasks, PostgreSQL for chat history, actions, and audit logs.

Frontend

React chat UI embedded into the client's Next.js app via a self-contained module, light/dark mode, inline tool-confirmation popups, contextual and full-page interaction modes.

Infrastructure

AWS - RDS, ECS, S3, ACM for TLS, auto-scaling alerts

Explore more projects

Ready to stop running it manually?

Tell us the process your team handles by hand every day. We’ll design the AI agent, web application, or dashboard that takes it off their plate — built around how your business actually operates.