AI Investment Intelligence Engine

The reasoning layer behind every insight.

A large reasoning model instructed for investment education. It reasons through bull cases, bear cases, and risks — and is forbidden from inventing prices, earnings, or specific figures. This page explains exactly how it thinks.

Coverage

Built to reason across every corner of the market.

The AI is instructed on how investors actually make decisions — asset by asset, position by position, always weighing risk against context, always educational, never speculative.

The AI understands

Individual stocksETFsMutual fundsIndex fundsBonds & fixed incomeCrypto assetsREITsCommoditiesInternational equitySector exposureCurrency exposureInterest rate sensitivityInflation impactPortfolio allocationRisk concentrationMarket regimesInvestment goalsTime horizonBehavioral biasesBusiness fundamentals
Guardrails

What the model will — and will not — do.

Every response from InvestmentLens is generated with a strict system instruction. The rules below are always active.

It will

  • Explain concepts — from beta to duration to moats — in language a smart beginner can follow.
  • Provide bull and bear cases for any investment side-by-side, so you have both perspectives.
  • Reason about your portfolio — diversification, concentration, sector tilt, currency exposure.
  • Compare investments across business model, growth, risk, and typical investor fit.
  • Flag uncertainty whenever a specific number would matter and it does not know it.
  • Point you to primary sources — SEC filings, IR pages, exchange sites — for exact figures.

It will not

  • Fabricate prices, earnings, revenue, or ratios. If it does not know, it will say so.
  • Predict future prices or guarantee any return.
  • Give personalized advice. It does not know your full financial situation.
  • Answer "should I buy X" as a yes/no. It will reframe to "here is how to think about X".
  • Encourage market timing, options gambling, or leverage without heavy risk framing.
  • Sell your data. Nothing you type is monetized, shared, or used for ads.
Interface

A chat that remembers, without pretending to think.

The AI Assistant page inside your dashboard is a full-featured chat surface:

  • Streaming responses — you see the model reason in real time.
  • Per-conversation memory — every session remembers what came before.
  • Markdown rendering — headings, tables, code blocks all preserved.
  • Pin, rename, delete, search — conversations are permanent until you decide otherwise.
  • Copy any message or regenerate the last answer if it missed the mark.
  • Suggested prompts to spark exploration when you don't know where to start.
Example conversation
Y

Is my portfolio too tech-heavy at 42%?

AI

Depends on your goals. 42% technology is a strong overweight vs. broad-market indices, which typically sit near 25-30% depending on the benchmark. If your time horizon is long and you accept higher volatility, it can be defensible. Consider studying how correlated the top 3 tech positions are, and how the whole book behaves in a rising-rate environment — technology historically de-rates when rates spike.

Under the hood

Fast inference from a large reasoning model.

⚙

70B parameters

A large model with strong reasoning quality, tuned to remain grounded on real concepts.

⚡

Streaming

Tokens appear as they arrive. No waiting on a five-second wall of text.

🎓

Instructed for teaching

Every prompt is prefaced with an education-first system message that discourages speculation.

🛡

No fabrication

The model is explicitly told to defer to primary sources whenever precise numbers matter.

Talk to the AI in your workspace.

Ask about a company, an ETF, a portfolio, or a concept. It will explain, not sell.

Create account