FinsightAI is committed to producing accurate, useful, and independent financial analysis. This page describes our editorial standards, methodology, and how we maintain trust with our readers.

Our editorial mission

We help retail investors understand markets through data-driven, transparent analysis. We do not sell investment advice. We do not promote specific assets. We do not earn commissions on any trades you might make.

Our value proposition is simple: take publicly available financial data, run it through proven analytical methods, and present the results in a clear, educational format that helps you become a better-informed investor.

Sources we use

All market data on FinsightAI is sourced from Yahoo Finance, which aggregates feeds from major global exchanges. We use Yahoo Finance because:

  • It provides comprehensive global coverage (stocks, crypto, ETFs, fundamentals)
  • The data is reasonably accurate for educational purposes
  • It's the same source used by many financial websites and research tools

For editorial and explainer content (our blog), we reference primary sources where appropriate: SEC filings, exchange announcements, peer-reviewed financial research, and recognized industry publications.

How our AI analysis works

Every analysis score on FinsightAI is generated by an open, reproducible methodology. Here's exactly what we do:

Technical analysis layer (40% weight for stocks)

  • RSI (14-period): Standard Wilder formula. Values above 70 = overbought, below 30 = oversold.
  • MACD: 12-day EMA minus 26-day EMA, with 9-day signal line. Bullish/bearish crossover detection.
  • Moving averages: 20-day, 50-day, and 200-day simple moving averages for trend identification.
  • Bollinger Bands: 20-period middle band with 2 standard deviation bands for volatility context.
  • Support/Resistance: Calculated as rolling highs and lows over a 30-day window.
  • Golden/Death Cross: 50-day SMA crossing above (golden) or below (death) the 200-day SMA.

Valuation layer (35% weight for stocks)

  • P/E Model: Trailing EPS × sector-average P/E ratio
  • Graham Number: √(22.5 × EPS × Book Value Per Share) — Benjamin Graham's intrinsic value formula
  • Dividend Discount Model: D1 / (r − g), where r = 10% required return and g = 4% assumed growth (for dividend-paying stocks)
  • Composite fair value: Average of applicable models
  • Margin of safety: (Fair value − Current price) / Fair value × 100

Quality layer (25% weight for stocks)

A 0–100 quality score based on:

  • Return on Equity (ROE)
  • Profit margins
  • Debt-to-equity ratio
  • Current ratio (liquidity)
  • Revenue growth
  • Dividend yield

Cryptocurrency weighting

For cryptocurrencies, traditional valuation doesn't apply (no earnings, no book value). The score is reweighted: 80% technical, 20% quality proxy (using market cap and liquidity as proxies for project quality).

This methodology is fixed. We don't override scores manually, hide bearish signals, or boost favored stocks. The same engine analyzes Apple the same way it analyzes any other asset.

Editorial independence

FinsightAI is supported entirely by advertising (primarily Google AdSense) and the operational portfolio of Trend & Brands. We accept no payment from issuers, exchanges, brokerages, or any company we cover.

Specifically:

  • We do not run sponsored content
  • We do not have affiliate relationships with brokerages
  • We do not accept payment to add specific assets to our coverage
  • We do not change analysis scores at the request of any party

Conflicts of interest

The operators of FinsightAI (the team at Trend & Brands) may personally hold positions in some of the assets analyzed on this site. These are individual investment decisions made independently of our editorial work. We do not buy or sell based on internal analysis ahead of publishing.

If we ever publish opinion-style content on a specific company (in our blog), and the writer holds a position in that company, we will disclose it within the article.

Use of AI in content creation

FinsightAI uses computer algorithms to perform mathematical calculations (technical indicators, valuation formulas, quality scoring). These are deterministic mathematical functions, not generative AI.

Where we publish educational articles (in our blog), they are written and reviewed by human editors. We do not publish generative-AI-written articles without human editorial review. Any content that is AI-assisted will be clearly labeled.

Accuracy and corrections

We strive for accuracy, but we recognize that errors happen — in data, in our calculations, or in our editorial content. If you find an error:

  1. Email us at hello@finsightai.online with the URL of the page and a description of the issue
  2. We will investigate within 2 business days
  3. If confirmed, we will correct the error and (for editorial content) add a correction note at the bottom of the article

Limits of our analysis

We want to be transparent about what FinsightAI's analysis cannot do:

  • Predict the future. No model can. Markets are influenced by countless factors — earnings surprises, geopolitical events, central bank decisions, sentiment shifts — that no algorithm fully captures.
  • Account for breaking news. Our scores reflect data as of the last refresh, not real-time news.
  • Read management quality. Numbers can't fully capture the quality of a CEO, the strength of a moat, or the trajectory of innovation.
  • Replace professional advice. A financial advisor knows you. We know the numbers. You need both.

User feedback

Reader feedback is the best way for us to improve. If you spot something wrong, want us to cover an asset, or have a feature suggestion — email us. We read every message.

Contact

Editorial inquiries: hello@finsightai.online