CONSULTING · PROJECTS · APPLIED AI
AI that works for your business.
We identify where AI makes sense and build a first project with clear objectives. For businesses, SMEs and public organisations in Andorra, Catalonia and Occitania.
01 / USE CASES
Start with a real task.
Explore three starting points. The project is scoped around your team, processes and data.
Today
Emails, PDFs and data that someone copies between applications.
The project we can build
A workflow that classifies, extracts fields and prepares the next action. Exceptions reach the right person before approval.
- 01Document
- 02Extraction
- 03Validation
- 04ERP / CRM
What we will measure
Time per operation, manual corrections and exception rates.
02 / ARCHITECTURE
Your AI. Your environment.
The model is one decision. Where it runs, who can access the data and how it is maintained matter just as much. Compare the options carefully.
Local / private LLM
We assess models and hardware for workloads on local or private infrastructure. Model size, memory, access and maintenance determine feasibility.
Environment control and dedicated sizing.03 / DELIVERY
From the first challenge to a measurable pilot.
- 01
Understand the process
Listen to the team, observe the workflow and establish baseline time and error measures.
- 02
Prioritise a useful case
Define the problem, sources, access, risk and criteria for pilot success.
- 03
Build and evaluate the pilot
Test with authorised data and representative cases. Review results, errors, costs and human control points.
- 04
Integrate and support
Connect the validated workflow, train the team and define monitoring, maintenance and review.
More capability. With control.
Access permissions, authorised data sources, traceability and human oversight are part of the design. We assess risks, provider terms and data protection needs before production.
ANDORRA / CATALUNYA / OCCITÀNIA
Based in Andorra. Working across borders.
From Escaldes-Engordany, we scope projects for organisations in Andorra, Catalonia and Occitania. A local team connecting business needs, language, infrastructure and regional context.
ESCALDES-ENGORDANY · +376 800 750
Clear questions. Useful answers.
How do we start an AI project?+
We start with one process: workload, time, errors and available data. That baseline helps scope a pilot with success criteria before deciding whether to scale.
Can we run an LLM on local machines?+
It is an option to assess. Feasibility depends on the model, memory, workload, latency, security and maintenance needs. Compare local, private, cloud and hybrid approaches before choosing.
What is a RAG assistant for internal documents?+
It retrieves authorised information to prepare contextual answers with references. It requires access controls, quality documents and evaluation. References help verification but do not guarantee a correct answer.
Can it integrate with ERP, CRM or IBM i systems?+
We assess existing APIs, connectors, permissions and data. The project defines what AI can prepare or execute and what requires human approval, with operations recorded.
How do we approach AI in the public sector?+
Use a defined case, authorised data, access controls, traceability and staff oversight. Assess risks and applicable obligations before deployment. An assistant does not replace administrative decisions.