AI consultancy · Brussels · since 2005

Applied AI, built by researchers

Mentis is a Belgian AI company, spun out of the Université libre de Bruxelles in 2005. We design, build and run machine-learning, NLP and generative-AI systems for organizations that need them to work in production — not just in a demo.

Talk to us
What we do
Neural network visualisation

Services

What we do

From the first feasibility question to a model running in production — one team, end to end.

AI & machine-learning development

Predictive models, NLP, computer vision and generative AI (LLMs, RAG) — trained on your data and evaluated against your metrics.

AI-powered applications

Web, mobile and back-office applications with AI built in: assistants, search, recommendations, automated document handling.

Responsible & secure AI

Data protection, model evaluation, bias checks and human-in-the-loop design from day one, so what we ship can be trusted and audited.

MLOps & DevOps

Pipelines, monitoring, retraining and deployment on your cloud or on-premise — models that keep working after launch.

Custom software

The APIs, integrations and data platforms around the model. We build the whole system, not just the notebook.

AI strategy & roadmap

Where AI pays off in your organization, what data you need, what to build first — a concrete plan, not a slide deck.

Why Mentis

Research roots, production habits

Twenty years of AI, before it was a buzzword

We were building NLP and predictive systems long before large language models — and we know which problems they solve and which they don't.

Born at ULB

Mentis started in 2005 as a spin-off of the Université libre de Bruxelles. Twenty years later, the scientific rigor is still how we work: measure first, then decide.

We ship our own products

KnowledgeRAG (retrieval-augmented generation) and OptimizeFlow (route & field-service optimization) are ours. Running them in production keeps our consulting honest.

Small senior team

Data scientists, ML engineers and software architects who have delivered 40+ projects together. You work with the people who build it.

How we work

Start with a scoped pilot

A few weeks, a clear success metric, your real data. You see results before committing to a full build.

Your data stays yours

We deploy on your cloud or on-premise, work with Belgian and EU data-protection rules in mind, and hand over the code.

Built to be operated

Every model comes with monitoring, documentation and a retraining path — the boring parts that make it last.

Around long after go-live

Support, iteration and knowledge transfer to your team, for as long as you need it.

Use cases

Problems we have solved

A sample of the domains our team has worked in over the past twenty years.

Field-service & route optimization

Scheduling technicians and vehicles under time windows and skills constraints — the work that became our OptimizeFlow product.

Knowledge management with RAG

Assistants that answer from your own documents with citations, powered by KnowledgeRAG. Fewer hallucinations, traceable sources.

Genomics data platform

Storage, search and analysis of large genetic datasets for research and healthcare teams.

Advertising effectiveness

Models that attribute outcomes to campaigns and channels so marketing budgets go where they work.

Bankruptcy prediction

Early-warning scoring of company insolvency risk from financial statements and market signals.

Pricing optimization

Demand-aware price recommendations that balance margin and competitiveness.

Predictive maintenance

Failure prediction from sensor and maintenance-history data, so interventions are planned instead of emergencies.

From the blog

View all posts »

Hands-on tutorials and opinions from our team on NLP, deep learning and where AI is heading.

2005
Founded
20+
Years of AI
40+
Projects delivered

Have a problem AI might solve?

Tell us about it. We will give you an honest read on feasibility, data needs and a first step — usually within a week.

Get in touch