How AI Alpha Engine Works

AI Alpha Engine is an automated investment research platform. Data is collected and measured deterministically, models then reason across the whole covered universe to produce global research, and a further reasoning stage applies that research to the holdings you own. The order matters — the market is analysed before your portfolio is — and this page follows it.

What is AI Alpha Engine?

A research platform, not a chatbot pointed at a ticker — and the difference is not the amount of AI involved but where it sits. The foundation is deterministic: market data ingested, momentum and relative strength computed against a global benchmark, each piece of evidence attributed to the company it genuinely concerns and repeats of the same story collapsed. Those steps give the same answers from the same inputs every time.

On top of that foundation, models reason twice, at two different scopes. First across the whole covered universe, with no portfolio in the picture: which drivers and themes lead, and where each company’s investment case stands. That global research is published every market day in the model portfolio, so it is visible without an account. Only then does a second stage read your own holdings against it. Both stages cite the evidence they used, which is why the platform can tell you what supports a conclusion.

Data sources — market prices, news, analyst actions, macro drivers — feed the AI Alpha Engine, which outputs theme leadership, risk signals, and watchlist candidates

How it works

Three steps — and the first one is already running. It does not wait for you to sign up.

Step 1
The engine researches the market
Every market night, across the whole covered universe: market data in, relative strength computed, evidence attributed to the right companies — and then reasoned about, universe-wide, to work out what is leading and where each case stands. All of it happens without you, and it is published every market day in the model portfolio.
Step 2
You add your holdings
Import a holdings export from your broker. Your positions and weights become the lens the existing research is read through — they do not start the research off.
Step 3
You get research about your portfolio
A cockpit summary and a full written report: what is still supported, what is weakening, and where a closer look may be warranted — refreshed every market night, with the evidence cited.

“Every market night” means the night after a trading session — the research is built while the market is closed and is waiting before it opens again. Markets do not trade at weekends, so neither does the engine.

What you receive

Cockpit
A concise summary of the most important signals, risks, and review priorities — at a glance.
Report
Detailed portfolio analysis covering themes, economic drivers, catalysts, and highlighted signals and risk flags.
Drivers
Portfolio positioning by economic driver — which drivers you hold, how each ranks on relative strength versus the benchmark, and how much of the portfolio sits in the current market leader.
Thesis
Validation of whether each holding still supports the investment case in light of recent news, leadership shifts, and analyst actions.
News
Catalysts, analyst upgrades and downgrades, and market events affecting your specific holdings and watched themes.
Holdings
Current holdings and economic-driver exposure for each position.
Analysis
Every completed import is kept as an immutable snapshot. Each one is repriced at today’s prices and exchange rates to show whether keeping those recorded holdings unchanged would have left you with more or less money than your latest recorded portfolio is worth today, and you can open a snapshot to see what each position contributed. It is a hypothetical hold comparison, not actual account performance: it assumes no money entered or left the portfolio, and later trades, dividends, fees and cash flows are not reconstructed.

The analysis in detail

Optional reading. Expand a section to see how the nightly analysis is produced.

The methodology

Five analytical principles underpin every portfolio review — from where capital is flowing to where the evidence is strongest.

Three-tier methodology: foundational layer of Economic Drivers and Themes feeds Leadership Detection, which informs Leadership Shifts and Conviction Signals
Economic Drivers
Markets are moved by a small number of macro forces at any given time — AI infrastructure, energy transition, defence spending, housing cycles. We identify which drivers are gaining dominance and find the leading positions within each.
Themes
Every driver breaks down into the themes that carry it — data-centre power and cooling, grid build-out, defence electronics. Themes are where capital concentrates inside a driver, and we track which of them are leading right now.
Leadership Detection
Relative strength and momentum data updated each market night identify which names are attracting institutional capital. Past winners are never assumed to be future winners — leadership is re-evaluated from scratch each session.
Leadership Shifts
The engine tracks where momentum is building and where it is fading — surfacing tickers gaining strength and those where thesis support may be weakening, so changes show up early rather than late.
Conviction Signals
Signal strength and leadership consistency produce a conviction ranking — identifying where the evidence is strongest for further investigation and where risk flags warrant closer review.
Example: how an economic driver breaks down
Economic Driver
AI Capex
Theme
Data-Center Power & Cooling
Holdings
Vertiv, SPX Technologies
From data to decision support

An end-to-end pipeline that turns market data into portfolio-specific insights and review priorities.

Stage 1
Ingest
Market prices, news, analyst actions, and earnings data are pulled from multiple sources each market night and normalised into a single consistent dataset.
Stage 2
Analyse
Momentum scores, relative strength vs the global benchmark, theme alignment, and catalyst signals are computed across the full universe.
Stage 3
Rank
Every ticker is scored by expected alpha. Conviction is driven by leadership strength, economic driver exposure, and catalyst quality — not sector balance.
Stage 4
Review
Each position is reviewed for signal strength, exposure alignment, and risk flags — surfacing what may deserve closer attention.
Stage 5
Report
An AI-generated report surfaces signals, risk flags, emerging themes, and evidence to support your own portfolio review.
From signals to review priorities

Every market night the engine surfaces a clear set of signals for each position. The AI report translates those signals into review priorities — surfacing what may deserve attention and why, so you can make your own informed decisions.

Market information and AI analysis produce four signals — theme momentum, relative strength, catalyst signal, risk indicator — which map to five review priorities: Supported, Weakening, Improving, Risk flagged, Monitor

Join the community

Discuss portfolio strategy, market leadership, and how to get the most out of AI Alpha Engine with other members.

Join Discord ↗

Choose the access level that fits you

Explore the available plans to see which portfolio analysis, reports and features are included.

View plans

All analysis is generated by AI based on market data. Nothing on this platform constitutes personalised financial advice.

← ← Back to Home

This information is for informational purposes only and should not be considered investment advice. Investing in securities involves risk, including loss of capital. Past performance is not a guarantee of future results. AI-generated analysis may be incomplete or incorrect; verify information before making investment decisions.  ·  Privacy  ·  Cookies  ·  Terms  ·  © 2026 Lighthouse Cat AB. AI Alpha Engine is a product of Lighthouse Cat AB.  ·  English / Svenska