AI Portfolio Modelling
Orèvance analyses market, macro, and volatility data continuously, then converts it into a plain-language recommendation you can act on or decline. The model handles the data volume. You keep the decision.
The friction
Most professionals we speak with in Toronto and Vancouver already have capital to allocate. What they don't have is three hours a week to track sector rotation, rate announcements, and volatility spikes across separate accounts.
That gap doesn't get closed by more dashboards. It gets closed by a system that reads the data continuously and surfaces only what changes your position, so the review takes minutes rather than evenings.
Core engine
Four functions run in parallel, each addressing a distinct part of the allocation problem rather than one generic "AI insight" feed.
01 / Forecasting
Models forward-looking price and volatility scenarios from historical patterns and live market data, updated as new information arrives rather than on a fixed schedule.
02 / Exposure
Monitors concentration, correlation, and volatility drift across your holdings continuously, flagging exposure changes before they compound.
03 / Execution
Adjusts allocations within thresholds you set in advance. Nothing moves outside those bounds without your explicit approval.
04 / Capacity
The same modelling depth applies whether the account holds five figures or a much larger sum. Insight quality scales with data, not with account size.
Methodology
Orèvance runs an ensemble of statistical and machine-learning models rather than a single black-box predictor. Each model votes on a probable range of outcomes for a given asset class, and the outputs are weighted by recent forecast accuracy before being combined.
This does not produce certainty. It produces a ranked, probability-weighted recommendation with a stated confidence band, which is a materially different claim than "the model knows what happens next." No model does, and we don't represent otherwise.
The output is always reviewable before execution. Automated rebalancing only applies within thresholds you define; the initial allocation decision remains yours.
Ingestion normalizes incoming feeds into a common schema. Scoring applies the weighted model ensemble. Risk filter checks the output against your stated tolerance. Output returns one ranked recommendation for your review.
One-click setup
Connect an account with read-only access. No trading permissions are granted at this stage.
Set a risk tolerance and time horizon. This constrains every recommendation the model can return.
Review the first allocation model. Approve, adjust, or close the session — nothing executes without you.
Before you initialize
Account connections use read-only access wherever the provider supports it, and all data in transit and at rest is encrypted. Trading permissions, where required, are scoped narrowly and can be revoked at any time from your connected institution.
Coverage spans public equities, exchange-traded funds, fixed income instruments, and a defined set of digital assets. The model does not currently extend to private placements or illiquid holdings, and we won't represent otherwise.
Fee terms are disclosed in full before you approve any allocation, with no trading costs applied without prior visibility. There is no requirement to commit to execution simply to view your first model output.
Setup takes under 60 seconds. No trading commitment is required to view your first model.