Attentia Welcome Agent
From emails and excel to central onboarding platform
Attentia is a leading Belgian HR services provider, supporting over 15,000 companies with payroll, well-being, and HR administration. Onboarding new clients is a critical and complex process: for each new company, HR data spanning dozens of variables must be collected, validated, and entered into Attentia’s core systems before a single payslip can run.
That process relied heavily on manual work: experienced operators deciphering inconsistent Excel files, validating data against business rules, and re-entering everything into multiple back-end systems. Together with Growing Pai, Attentia set out to fundamentally change that.
The challenge & solution
Every new client onboarding at Attentia required an operator to manually process a large Excel file containing HR master data: employee records, salary components, contract types, and dozens of validation rules specific to Belgian social legislation. With hundreds of onboardings processed each year, this meant thousands of hours spent on repetitive, high-stakes data work. A single error could delay the first payroll run.
We built an AI-powered onboarding assistant that guides operators through the entire process: intelligently reading and interpreting the incoming Excel data, validating it against Attentia’s business rules, flagging anomalies for human review, and automating the handoff to back-end systems. What previously took days of manual effort can now be completed in a fraction of the time, with fewer errors and full auditability.
Key challenges we solved
No two client files look the same. Companies deliver their HR data in a wide variety of formats, column orderings, and naming conventions. Rather than forcing clients to use a rigid template, the system intelligently maps incoming data to Attentia’s internal data model, regardless of how the file was structured.
Belgian payroll is governed by a dense set of legal rules, collective labour agreements, and company-specific configurations. The assistant validates every record against these rules in real time, surfaces exceptions with clear explanations, and ensures operators can act on issues immediately rather than discovering them at month-end.
The system is built with the operator in mind. Rather than a black box that outputs a result, it offers a full-featured operator interface where the AI supports every step of the onboarding workflow: reviewing extracted data, acting on validation results, handling exceptions, and triggering downstream actions. Operators stay in control throughout the process, with AI doing the heavy lifting at every stage.
Technology approach
The solution is built as a LangGraph-based agentic workflow running on Azure, combining structured data processing with large language model reasoning. The agent moves through a series of defined stages: data ingestion, identification, validation, transformation, and final approval, with human-in-the-loop checkpoints at each critical step.
A custom Angular front-end gives operators a clean, step-by-step interface to review AI output, handle exceptions, and trigger the final export to Attentia’s core HR systems. All processing happens within Attentia’s Azure environment, ensuring full data sovereignty and compliance with GDPR requirements.
By separating the validation logic from the AI layer, the system remains fully auditable and maintainable: business rules can be updated independently of the AI model, and every decision is logged with a clear rationale for compliance purposes.
Still spending days on manual onboardings?
If your operations team is manually processing client data because it’s “too complex” or “too inconsistent” to automate, we can help.
Growing Pai has proven expertise in building AI solutions that work with real-world data, real constraints, and real business impact.
Let’s explore what this could mean for your organisation: via mail hello@growingpai.com or phone +32 475 54 22 16.
