PREDITTALenses
Sign In

Who We Are

Technology Company Specialized in Risk Intelligence

We create products that integrate AI, machine learning, and advanced analytics for enterprises

Preditta Advanced Risk Solutions is a technology company that develops cutting-edge products integrating artificial intelligence, machine learning, and advanced analytics. We focus on risk management and smart solutions for financial institutions, investment funds, and enterprises seeking to optimize their operations.

Our solutions combine deep domain expertise with state-of-the-art technology to deliver actionable insights, automate complex decisions, and transform how organizations manage risk in real time.

Inside the Platform

The Four Modules

Expert Rating

A credit grade built on financial statements and peer comparison:

  • Statement Ingestion: Balance sheet and income statement as required input, cash flow and value added as optional support, in PDF, XLSX or XML.
  • Thirty-Four Indicators: Liquidity, profitability, leverage, asset composition, working capital, cash flow and operational leverage, computed from the mapped accounts.
  • Position Against Peers: Each indicator becomes a percentile inside the sector peer group, rather than being judged against an absolute value.
  • Grade From AA to D: The weighted average of the percentiles sets the band, with configurable weights and a simulation workspace for testing scenarios before they apply.

Statistical Rating and Prediction

Probability of default estimated by machine learning:

  • Forward-Looking Outcome: The model is trained to predict material delinquency in the following year, not to describe what already happened.
  • Validation by Company: Cross-validation splits whole companies between training and test, so no company appears on both sides.
  • Discrimination and Calibration: Ranking power and the fit of the predicted probability are published inside the platform, alongside the importance of each variable.
  • Cut-Off Strategy: Analysis of where to place the decision threshold, with the hits and misses of each choice laid out over the assessed portfolio.

Payment Behaviour

Reading receivables history as a risk signal:

  • Portfolio Aging: Distribution of invoiced value by delay band, from paid on time to delinquency beyond ninety days.
  • Collection Period and Weighted Delay: Collection speed by client and by business segment, weighted by invoice value rather than by count alone.
  • Concentration: Weight of the largest clients and portfolio coverage, to size exposure to a handful of names.
  • Effect on the Grade: Payment history enters the statistical model as a variable, with its effect explained rather than buried.

Public Data and Query

A comparison base built on open sources:

  • Listed Companies: Quarterly and annual filings published by CVM, with variables, indicators and documentation for each calculation.
  • Credit Registry: Series from the Central Bank credit information system, broken down by delinquency, region and term.
  • Sector Reports: The distribution of each indicator within the sector, which is precisely what gives a company's percentile its meaning.
  • Natural Language Query: Questions in plain language answered over a fixed catalog, always read-only, with a trail of what was asked and what was refused.

How It Works

From File to Grade, in Five Steps

  1. 01

    Upload

    Statements arrive as PDF, XLSX or XML.

  2. 02

    Mapping

    Accounts are recognised and matched to the standard chart.

  3. 03

    Review

    Extracted values stay visible for checking before any calculation runs.

  4. 04

    Comparison

    Indicators are positioned against the sector peer group.

  5. 05

    Grade

    The assessment is recorded together with the methodology that produced it.

Principle

No Grade Without an Explanation

A risk number that cannot be audited cannot carry a credit decision. That is why every scoring engine in the platform exposes its own calculation to whoever uses the result.

  • Published Methodology

    Each engine has a page describing its calculation, open to whoever uses the grade.

  • Parameters in Plain Sight

    Weights, bands and limit parameters are configurable and shown next to the result.

  • Audited Query

    Every natural language question is recorded, including those refused by the security layer.

Leadership

Founders

Alexandre Ywata

Board Member

Alexandre Xavier Ywata de Carvalho

is a mechanical-aeronautical engineer and specialist in aerial armament engineering from ITA, holds a master's degree in statistics from UnB and a PhD in statistics from Northwestern University, United States. He holds CGA/ANBIMA certification and CVM qualification for portfolio management. He was a professor at the University of British Columbia, Vancouver, quantitative analyst at UBS, Chicago, Director of Regional, Urban and Environmental Policies and Substitute President at Ipea. He was President of Caixa Participacoes, Vice President of Risk and Internal Controls and Vice President of Investment Funds at Caixa Economica Federal, Undersecretary of Economic Law at the Secretariat of Economic Policy (SPE), Secretary of Infrastructure Development (SDI) at the Ministry of Economy, and Special Secretary for Productivity and Competitiveness (SEPEC) at the Ministry of Economy. He was CEO at Banco Digimais. He is a professor of statistics and econometrics in the Master's and PhD program in Economics at IDP. He is the author of the book 'Statistical Methods for Economics and Finance', published by the University of Brasilia. He has almost 30 years of experience in developing risk systems and portfolio optimization.

Alexandre Marinho

Board Member and Strategist

Alexandre Saldanha Marinho

is an experienced technology executive and founder of SIACorp, with over 40 years of experience in delivering innovative credit automation solutions. A recognized leader in the implementation of AI-driven credit decision systems across Latin America, Alexandre has a proven track record of implementing large-scale automation projects in banks and major industrial sectors, including agribusiness, chemical, steel, and pharmaceutical. Before founding SIACorp, Alexandre worked at IBM and PricewaterhouseCoopers, where he developed a solid foundation in enterprise technology, financial systems, and business process optimization. He is the creator of CreditFlow, a leading credit automation platform that transformed credit analysis operations for several multinationals and publicly traded companies in the region. He also led the development of CreditChat, an AI-powered assistant that allows real-time interaction, analysis, and collaboration within credit teams, and CreditVision, a next-generation platform for financial statement automation and advanced credit risk analytics, now enhanced with generative AI capabilities. Alexandre also founded Agrometrika, an agritech credit intelligence platform that integrates climate, agricultural, and financial data (acquired by Aliare, a BTG Pactual portfolio company) and Akrual, a fintech for cloud management of securitized products (CRIs, CRAs, FIDCs, debentures).

Ready to Transform Your Risk Management?

Contact us to schedule a personalized demo and discover how our solutions can optimize your operations.

Access PlatformContact Us