Decision-making intelligence augmented by AI
Transform massive volumes of data into actionable investment signals. Real-time portfolio optimization and risk management for demanding professionals.
Explore the Dashboard
Transform massive volumes of data into actionable investment signals. Real-time portfolio optimization and risk management for demanding professionals.
Explore the DashboardEach recommendation generated by Swap Avoirense is time-stamped and kept in a searchable register. The community of users can compare the signals emitted with the results actually observed on the market.
| Asset class | Signals generated (30 days) | Match rate | Confidence interval |
|---|---|---|---|
| Actions | 1,240 | 71.4% | ±3.2% |
| Forex | 860 | 68.9% | ±4.1% |
| Raw materials | 410 | 64.7% | ±5.0% |
Illustrative structure of the register. Up-to-date values, segmented by time horizon and model, can be viewed in the dashboard after creating an account.
Proportion of signals whose predicted direction corresponds to the actual movement observed over the indicated period.
Statistical margin associated with each rate, calculated from the sample size and the variance of the results.
Users can export raw signal history for independent cross-checking, without intervention from the Swap Avoirense team.
Three distinct analytical layers work in parallel to produce a final risk-weighted recommendation.
Natural language processing (NLP) applied to financial reports, press releases and news feeds, combined with continuously calculated sentiment analysis.
Simulation of stress scenarios based on probability distributions to estimate a portfolio's exposure to unexpected volatility.
Models adjust their parameters when structural market changes are detected, rather than relying on fixed rules.
The processing pipeline is broken down into four stages, each independently auditable.
Massive collection of raw data via secure APIs connected to markets and news sources.
Cleaning, deduplication and structuring by proprietary AI layers before any statistical calculation.
Generation of risk-weighted recommendations with an explicit confidence score.
Manual execution or integration of signals directly into your existing workflow via API.
The platform adapts to the level of autonomy sought, from institutional allocation to individual monitoring.
Optimization of asset allocation based on risk-weighted recommendations, across multiple asset classes simultaneously.
Access to analysis tools previously reserved for institutional structures, without requiring a dedicated quantitative team.
Validation or questioning of investment theses based on quantified and traceable data rather than market intuitions.
Swap Avoirense was built around a simple principle: an investment recommendation is only valuable if its history can be reviewed after the fact. This is why each signal produced by our models remains viewable, even when it turns out to be incorrect.
The models combine natural language processing, stochastic modeling, and reinforced learning, but none of these are touted as foolproof. The confidence scores displayed reflect the actual uncertainty of the model at a given time.
The following answers detail the internal workings of the platform and the known limitations of the models used.
Stochastic modeling models incorporate shock scenarios to estimate a range of likely losses, but no statistical model can predict an event without historical precedent. In the event of a structural break detected, the confidence scores displayed automatically decrease and an alert signal is entered in the register.
Data comes from market feeds, public financial reports, and news sources aggregated via API. Each source is weighted according to its historical reliability, and discrepancies between sources are reported rather than silently smoothed.
Any signal emitted is timestamped at the time it is generated, before the actual result is known. Users can export this raw history for independent cross-checking, preventing any retroactive changes to displayed performance.
Join the Swap Avoirense community and access full performance logs today.