Visual: a translucent surface made of fine lines, with a line chart running across it. Three structures behind it push it into shape. Scrolling carries you through the surface to the other side, where the three become visible: Business Intelligence, Artificial Intelligence and Leadership. The data is illustrative.

Wilson MotaBusiness Intelligence Manager

Beyond the dashboards.

What if the reports I build are only a fraction of what I can deliver?

Step behind the dashboard

I design data platforms, build with AI and lead a BI team.

Wilson Mota

Wilson MotaBusiness Intelligence Manager at CorientSwitzerland

Open the CV

Artificial Intelligence

AI brings speed. We bring judgement.

My AI projects

Started November 2026 · app at version 3

Life HQ

Six AI specialists and one shared folder plan my training, meals and recovery around the week I actually have.

What I decided

  • Designed the structure: one folder as the shared brain, one owner per file.
  • Set the rules: records are only added, never rewritten, and the AI never invents precision.
  • Make the calls. The specialists analyse and propose.

What the AI does

  • Six chats, each with one job: HQ, training, nutrition, progress, the app and recovery.
  • Built the plan, 14 exercise sheets and a phone app with a 3D figure.
  • Rebuilds the Excel report from the raw files, never by hand.

Not there yet

  • Early days: no trend is read before week three.
  • The app has not been tested on a real phone yet.

The app

The phone app on a walking day: today asks only for a walk, 20 to 30 minutes on the treadmill, shown by a small 3D figure.

A walking day. Shown in English here; the app itself is in Portuguese.

Running at home since autumn 2026

Smart home

A Home Assistant system that moves the blinds with the sun, shows the solar power and the energy on a wall display, keeps an eye on the heat pump, and has an AI say each morning what to wear for the weather.

What I decided

  • Set the rules: check first, simulate before changing, back up before every change.
  • An automation never overrides something done by hand.
  • The AI only suggests. Nothing physical moves without my consent.

What the AI does

  • Writes the configuration and the wall display, straight through Home Assistant.
  • Simulated 25 cases with fake sensors before one set of rules went live.
  • Works inside the house too: what to wear, morning and evening, and a plan for the day.

Not there yet

  • The lighting routine is written, but not yet seen live.
  • Knowing which room is in use is still to come.

In numbers

  • 9screens on the wall display
  • 10blinds under its routines
  • 25cases simulated before going live

From the project’s own record.

Working pipeline · product not built

Personal CFO

My whole financial situation, live: where the money goes, what changed and what comes next.

What I decided

  • Set the rule: the AI never calculates. Code does the sums; the AI classifies and explains.
  • Kept a human check on what the AI classifies.
  • Set a commercial version aside, for three reasons that are on record.

What the AI did

  • Profiled the bank exports: formats, duplicates, transfers and gaps.
  • Wrote the pipeline in Python, from raw exports to analytics.
  • Flagged what it did not know instead of filling the gaps.

Not there yet

  • The product itself: phase one is not built.
  • About 29% of spending had no named payee on the statement.
  • The rules were written by hand for these banks. A second user would need new ones.

How it works

  1. Bank exports come in
  2. They are cleaned
  3. They become one ledger
  4. The ledger is analysed

About 2,400 real transactions. Every number traces back to one of them.

In alpha · first alerts not sent yet

Property Deal Radar

Brings together the property investments with the best economic viability, and alerts me when a hot deal appears.

What I decided

  • Set the criteria for a deal that makes economic sense.
  • Keep one list, so the same property is never shown twice.
  • Asked for an alert only when a new hot deal appears.

What the AI is built to do

  • Find listings across portals and remove the duplicates.
  • Run the numbers: yield, cash flow and the equity needed.
  • Rate the risk, and say what information is missing.

Not there yet

  • In alpha: no alert has been sent yet.
  • Not every portal allows automated collection.
  • Rents are estimates, not income.

How it works

  1. Find the listings
  2. Remove the duplicates
  3. Run the numbers
  4. Rate the risk
  5. Alert me

Built for my own investments, and the kind of tool an estate agent would want.

Live on paper trading

AI Trading Bot

AI builds a trading bot in minutes. The hard part is tuning it to behave the way I would.

What I decided

  • Chose the broker: Alpaca, with one paper account per bot.
  • Chose the AI that reads the news: Jev, through OpenRouter.
  • Kept it to stocks, and to paper money.

What the AI did

  • Wrote the test engine, the bots and their tests.
  • A second AI agent redid the research sums and found 3 bugs in the test code.
  • An audit of the bot's code added four safety rules before the pilot.

Not there yet

  • No proven edge: none of the three ideas reached the bar.
  • No real money. The pilot is reviewed on 11 December 2026.

The result so far

How sure can we be that the edge is real? The bar, set before the tests: 95%.

Read the case study

Business Intelligence

Data you can trust. Built to be audited.

Seven years in regulated environments, from analyst to manager.

Today

Business Intelligence Manager

In this position since June 2025

  • A team of six developers
  • The Power BI platform
  • Automation and AI
  • Reporting strategy

2016 – 2019

Business consulting · HOP

  • Led the market research team.
  • Business consulting, before BI.
  • A few highly technical spreadsheets were the first step towards working with data.

09/2019 – 03/2021

BI Analyst · Scores & Levels

  • Developed and maintained 10+ advanced Power BI reports.
  • Developed the on-premises (small-scale) data warehouse.

SQL ServerSSISPower BIExcel

11/2019 – 01/2021

Co-founder · WeValue – Consulting

  • Management and client advisory: 9 of the 10 projects I submitted were approved.
  • Responsible for public contracts.

Excel

03/2021 – 11/2021

BI Consultant · Freelance

  • Developed Power BI reports that delivered management gains for clients.
  • Focused on the Power BI user experience, with positive client feedback.
  • Managed a SQL database and several SSIS ETL processes.

SQL ServerSSISPower BIExcel

12/2021 – 05/2025

Senior BI Analyst · Stonehage Fleming

  • Led the implementation and configuration of Power BI Service in the company.
  • Responsible for the Dremio and dbt architecture, security and governance.
  • Developed several Alteryx processes that support business operations.
  • Developed products with SSIS, Visual Studio, MDM and Mage.ai.

SQL ServerSSISAlteryxMage.aiPower BIExcelDremioBusiness CentralCopilot

06/2025 – today

Business Intelligence Manager · Corient

  • Lead a team of 6 BI developers delivering analytical solutions to the whole group.
  • Own the Power BI platform, ensuring governance, security and scalability.
  • Drive automation and AI adoption to increase business performance.
  • Partner with leadership to shape reporting strategy and ensure compliance.

Stonehage Fleming became part of Corient: one continuous path since 2021.

SQL ServerMage.aiPower BITableauDatabricksCopilotClaude

Projects and certification

Certified · 2026 Databricks Data Analyst verify the credential.

Live · read-only

Claude × Power BI

The AI asks the model, under my permissions. It never touches the data. On its first day it ran every measure once.

What I decided

  • A live connection, not screenshots: the version an interviewer can watch.
  • Read-only, under a named user, with no secret stored anywhere.
  • The data: my own CRM report, masked.

What the AI did

  • Planned the four pieces from Microsoft's documentation.
  • Registered the app and set the tenant setting, each step with my permission.
  • Diagnosed why every query failed at first, and found the fix.

Not there yet

  • Not public: a visitor cannot open it. It is shown live in conversation.
  • It runs on a trial licence that ends on 16 December 2026.
  • Microsoft's endpoint is in preview and may change.

How it works

  1. A tenant
  2. An app
  3. A connector
  4. A model

Four pieces, and none of them is custom code.

Read the case study

Audit done · the AI version is next

My version vs AI version

My own report next to one Claude builds on the same model, so anyone can compare the two. Recorded, so the comparison can be watched.

What I decided

  • Built my version, and keep it untouched as the reference.
  • Set the rule: improvements are measured, not claimed.

What the AI did

  • Audited the model through the live connection: 98 measures, 7 failing, 2 root causes.

Not there yet

  • The AI version has not been built yet.
  • The recording is still to come.

Leadership

From analyst in 2019 to AI-empowered manager.

What colleagues say

“He takes ownership, asks the right questions before building anything and turns complex data into something leadership can act on. I would recommend him without hesitation.”

Carla Silva ReisCEO at Scores & Levels. Wilson reported directly to her

“I praise his dedication and sense of ownership, as he took every task as his own and could always be counted on to see a problem through to the end. … Any team would be lucky to have him.”

Daniela GomezOperation Team Manager at HOP, and Wilson's manager

“Wilson is a hard-working, honest and intelligent person always willing to help the team. … He was able to understand the business needs and helped us gain clarity of strategic objectives…”

José MatosProgram Manager, on the same team at Scores & Levels · On LinkedIn

“From the IT side, what stood out was how well he bridged business and technology. … He was ambitious, yet careful with data, reliable and easy to work with.”

Rodrigo TaveiraHead of IT at Scores & Levels

Contact

Say hello.

wilson.lisboa@wilsonlisboa.com LinkedIn GitHub The CV