Data & AI consultancy
Messy data becomes auditable decisions.
Fraud, risk, catalog and financial data. From diagnosis to an engine running in production — with an audit trail your team can actually defend.
of the catalog resolved without human review.
MARKETPLACE · AI PRODUCT MATCHINGWe reply within 2 business days.
What we do
Fraud and abuse detection
Coordinated patterns across accounts, ranked suspicious cases and an audit trail that holds up under scrutiny.
Catalog intelligence
Matching, enrichment and classification of products at scale, with human review only where it matters.
Financial data operations
Consolidation of multiple sources, dimensional modelling and reliable reporting for whoever decides allocation.
AI agents and automation
Assistants that answer from your real data and hand over to a person at the right moment.
How we work
Diagnosis
We understand the problem, look at the real data and tell you what is feasible and what is not. Fixed scope and price.
Pilot
An engine running on your own data, with a measurable result before any larger commitment.
Operation
In production, versioned and auditable — with your team able to run it without depending on us.
If the diagnosis shows it is not worth it, we say so. That is cheaper than the wrong project.
Case studies
Wash and cross trading detection
Coordinated trades across accounts created regulatory and reputational risk, and review was manual.
- Dedicated append-only data infrastructure with per-run auditing.
- Deterministic, explainable detector: order matching by time, price and quantity across multiple windows.
- Priority Score ranking suspicious pairs by risk.
Result: a prioritised queue of cases with traceable evidence, produced on a recurring basis for the compliance team.
Catalog matching and enrichment
Every seller listed the same product differently. Customers could not find what they wanted and sales were lost.
- Pipeline for data preparation, enrichment, classification and analysis.
- Machine learning combined with generative AI over a similarity base.
Result: the volume requiring human review dropped to 5% of the base.
Data architecture and reporting
Assets held across different banks and managers, each with its own format, with no single portfolio view.
- Dimensional modelling with asset, flow and strategy facts and dimensions.
- ETL of custody statements, cash flow and return calculation.
Result: a single base feeding decision dashboards and client reporting.
AI-powered enquiry assistant
High volume of repetitive enquiries about available units, and leads that go cold while waiting.
- Knowledge base generated from the real inventory of units.
- Agent that qualifies by profile, price and size and recommends what actually exists.
- Panel for the operator to keep inventory up to date.
Result: immediate replies and triage done before the agent joins the conversation.
Data career assessment
People entering or growing in data do not know where they stand or what to study next.
- Backend in TypeScript, Prisma and PostgreSQL, with automated tests.
- Questionnaire, score and recommendation by career track and proficiency level.
Result: a complete product built from the data model to a tested backend.
Who is behind it
Vettore Data is led by Bruno Golfette.
A physicist by training, he has worked with data and artificial intelligence for over a decade, between Brazil and Italy. He has delivered demand and risk models, fraud detection and decision systems for clients in retail, finance and real estate.
He also teaches: courses and talks on data, Python and applied AI.