Guide

AI investment research assistant

Updated November 4, 2026

An AI investment research assistant answers investing questions written in plain language by running the analysis on real financial data, and it shows you the data behind the answer. Financial AI® (thefinancial.ai) is one: you describe what you want to know, and it returns a one-line conclusion, the data table, a chart and a note on how it read your question, so you can check the work before you act on it.

What an AI investment research assistant is

A research assistant does not just tell you what it remembers from its training. It works out what you mean, gets the series it needs from a curated database, runs financial functions on them, and when nothing ready-made fits, it writes and runs the code for the calculation, such as a custom indicator, a backtest rule or a basket of stocks. You don't have to fill in forms or write code, and you can ask in any language.

Financial AI is built on more than a decade of work by quantitative portfolio managers, analysts, data scientists, traders and engineers, from data processing to backtesting and performance evaluation. It uses leading third-party AI models to understand your request and connects them to its own financial functions, quantitative library and curated data.

What you can ask

Real questions from Financial AI's quick-start cards and sample prompts:

• "What percent of NVDA revenue comes from China comparing to the US?"
• "How has NVIDIA performed compared with the S&P 500 over the past year?"
• "What is your CPI Forecast?"
• "Based on Financial AI's forecasts, do you think AMD will deliver upside surprise?"
• "What are the top 5 outperforming stocks in ARKK ETF for this year?"
• "I have AAPL, NVDA, INTC, CSCO, XOM in my portfolio. Which one poses earnings risk?"
• "Backtest a 50/200-day golden cross on SPY since 2015 and chart it against buy-and-hold."†
• "Show me the performance of a covered call for PLTR stock. Sell the calls 30 days out at the money only if PLTR is within 10% of its 52-week high."†

You can also attach up to 6 files to a message. A brokerage statement is imported as a portfolio, and from any other document, such as a report or a table in a screenshot, Financial AI pulls out the data you ask for.

How the answers are built

1. Data. Financial AI pulls the series it needs from 1,400+ curated sources. Every source is evaluated for relevance, consistency, accuracy and licensing before it is added, and it is reviewed continuously after that. The data is factual, with no panel data.
2. Analysis. It picks sensible defaults for anything you left open, such as the period, the benchmark or the weighting, and then runs the calculation: a comparison, a screen, a forecast lookup, a portfolio breakdown or a backtest. A status line shows the current step, and Activity info lists every step it took.
3. Chart and table. The answer comes back with the data table and the chart. You can download both, as CSV and as chart data. Backtests also give you the trades behind the result.
4. A note on interpretation. Note: I interpreted… says how your question was read and which assumptions were used. If something is off, rephrase it with explicit dates, tickers or a benchmark. Suggested follow-up questions sit underneath.

Most answers take about a minute. A backtest of several strategies can take longer. Financial AI can make mistakes, and your data is never used to train its models. Because each answer shows its interpretation, its steps and its data, you can verify it rather than take it on trust.

Financial AI in numbers

Who it is for

• Individual investors who want to research a company, check a portfolio's real exposure or test an idea without code. You can connect a broker read-only, build a portfolio by hand or import a file. ETFs are looked through to their holdings, so you see your true company, sector and country exposure.
• Active traders who want to test strategies in words: indicator and crossover rules, stop-losses, covered calls, rebalanced mixes, and intraday rules on 1-minute bars, including futures. Any strategy can be saved and rerun later.
• Professionals, such as portfolio managers, analysts, advisers, funds and allocators, who want proprietary forecasts of revenue, adjusted gross income and adjusted EBITDA, macro forecasts such as CPI and recession probability, factor exposures and hedge ideas. The forecast reports are also available in FactSet and Bloomberg; see Forecasts.

How it differs from a general chatbot

A general chatbot answers from memory and the web, and it analyses the data files you upload to it. Financial AI computes from curated financial data, runs real backtests and screens, has its own forecasts, and analyses your actual portfolio, and it shows its work. It also works inside ChatGPT and Claude, so you can bring it into the assistant you already use. For a feature-by-feature view against fiscal.ai, Composer, TrendSpider and a chatbot on its own, see the comparison of AI investment research tools.

What it does not do matters too. Financial AI is not a broker and not an adviser. Broker connections are read-only, and it places no orders, moves no money and sends no trading signals. It provides information and analysis, not personal investment advice.

Start

Questions about data, brokers or privacy are answered in the FAQ. To try it on your own questions, start your free trial. No card is needed.

* Measured over historical point-in-time forecasts. Past accuracy is not indicative of future results.

† Illustrative example. Simulated performance is hypothetical, does not represent actual trading, and is not indicative of future results. See Disclaimers and Disclosures.

Nothing on this page constitutes an offer or investment advice. Past performance is not necessarily a guide to future performance. Independent advice should be sought in all cases.