# How to Use the AI Stock Research Assistant

> The AI Assistant answers plain-English questions about stocks using the same data you find on bestshares.com, such as financial statements, analyst ratings, earnings, insider and congressional trades and market news. Type a question or pick an @Agent, read the answer and its sources, then check key numbers on the linked pages. It is research help, not investment advice.

Source: https://bestshares.com/learning-center/ai-chat

Last reviewed: 2026-11-13

## What is the AI Assistant?

The AI Assistant is a chat tool for stock research. You ask a question in plain English, for example "How has Apple's free cash flow changed over the last three years?", and it answers in a few paragraphs, often with a table, using data stored on bestshares.com.

It is useful when the answer is spread across several pages. A single question can combine a company's income statement, its analysts' price targets and its next earnings date, which would otherwise take several clicks. It is also a quick way to have a metric explained in context, such as what a debt-to-equity ratio of 1.5 means for a specific company.

It is **not** an adviser. It does not know your finances, it can be wrong, and it does not tell you what to buy or sell. Treat it like a fast research assistant whose work you still check.

## How to use the AI Assistant

1. **Sign in.** Chatting requires a free account, and each message uses one **credit**. Your balance appears next to the send button as *Credits*.
2. **Type a question** in the box under *Ready to analyze the markets?* Name the company and ticker, the metric and the period, for example "Compare Microsoft (MSFT) and Amazon (AMZN) operating margins over the last 3 years."
3. **Pick an agent (optional).** Type **@** or open the *@Agents* menu to add one of these:
   - *@FundamentalData*: financial statements, cash flow, debt and returns on capital.
   - *@CompareStocks*: side-by-side comparison of two or more companies.
   - *@Analysts*: consensus rating, price targets and estimate changes.
   - *@BullvsBear*: the main arguments for and against a stock.
   - *@StockScreener*: finds stocks that match criteria you describe, such as "dividend yield above 3% and payout ratio below 60%".
4. **Or start from a suggestion.** Four suggestion cards sit below the box. If you have a watchlist, one card uses a stock from it; another often features a large company with earnings coming up.
5. **Read the answer** on the conversation page. Use the copy buttons to copy your prompt or the answer, edit a prompt to ask again, and click a suggested follow-up question to continue.
6. **Come back later.** *Recent Conversations* lists your past chats with when each was last active. Click *Open* to continue one or the trash icon to delete it. Only the five most recent show until you click *Show All*.

## How to read an AI Assistant answer

An answer has three parts worth looking at separately.

**The figures.** Numbers come from the data the assistant looked up: reported financials, quotes, estimates and filings. Look for the **period** next to each number (fiscal year, quarter, trailing twelve months) because mixing periods is the most common source of confusion. A fiscal year is not always a calendar year: Apple's ends in late September, Microsoft's on June 30.

**The interpretation.** Phrases such as "strong balance sheet" or "rich valuation" are the model's reading of the figures. They are a starting point for your own judgement, not a conclusion. Ask a follow-up such as "compared with what?" if a judgement has no benchmark.

**The sources.** Below many answers is a list of the data the assistant used, linking to the matching bestshares.com pages, followed by *Related* follow-up questions. If a number matters to your decision, open the source and confirm it.

### What each agent is good for

| Agent | Best for | Watch out for |
|---|---|---|
| *@FundamentalData* | Trends in revenue, margins, cash flow, debt | Fiscal vs calendar years |
| *@CompareStocks* | Ranking peers on the same metrics | Companies in different currencies |
| *@Analysts* | Consensus, targets, revisions | Targets are opinions, not forecasts you can bank on |
| *@BullvsBear* | Seeing both sides of a debate | Balanced does not mean equally likely |
| *@StockScreener* | Turning criteria into a list | It returns the top 10 matches, not every match |

Without an agent, the assistant chooses the data it needs from the question. Agents make the answer more focused, and a question that reads like screening criteria is sent to the screener automatically.

### Asking about your watchlists

If your question mentions "my watchlist", "my portfolio" or the name of one of your [watchlists](/watchlists), the assistant can see up to 20 tickers from each of up to five of your lists, so you can ask "Which stocks on my watchlist report earnings next week?"

## Worked example: checking an Apple cash flow answer

Suppose you pick the suggestion *Apple cash flow generation & buybacks*, which asks *@FundamentalData* to analyze Apple's annual free cash flow and share repurchases. The data the assistant draws on is Apple's cash flow statement. On November 11, 2026, [Apple's annual cash flow statement](/stocks/AAPL/financials/cash-flow) showed:

| Fiscal year ended | Free cash flow | Shares repurchased |
|---|---|---|
| Sep 26, 2026 | $136.1B | $86.4B |
| Sep 27, 2025 | $98.8B | $90.7B |
| Sep 28, 2024 | $108.8B | $94.9B |

How to check the answer:

- **Right periods.** A correct answer refers to fiscal years ending in September, not calendar years.
- **Right arithmetic.** Free cash flow is operating cash flow minus capital spending. For fiscal 2026 that is $145.1B minus $9.0B, about $136.1B. Growth from fiscal 2025 is about 38%.
- **Sensible interpretation.** In fiscal 2024 and 2025 Apple spent close to all of its free cash flow on buybacks; in fiscal 2026 buybacks were about 63% of it. An answer saying buybacks "exceeded" free cash flow in 2026 would be wrong, and the table lets you catch that in seconds.

The same habit applies to any answer: find the one or two numbers the conclusion rests on and confirm them on the source page.

## Limitations and common mistakes

- **AI can make mistakes.** Models sometimes state a wrong figure, attach it to the wrong year or misread a ticker. The more an answer matters, the more you should verify.
- **Free accounts get free data.** The assistant sees what the site shows your plan. On a free account it works from the same rows the Pro-only tables show you (usually the 3 most recent), 5 years of financial history and the same stock rating details the site shows you, and it tells you when a full answer needs Pro. It cannot read earnings call transcripts or business segments for free accounts.
- **Only bestshares.com data.** The assistant does not search the open web. It cannot know about news that is not yet in the site's news feed, and it covers companies and data sets the site covers.
- **As fresh as the underlying pages.** Prices, news and filings have the same delays as the pages they come from. Ask for a "latest" figure and check its date.
- **Not personal advice.** The assistant does not know your goals, taxes or time horizon. Use it to understand a company, not to decide a trade.
- **Long conversations.** The assistant only remembers your last 5 questions and its answers in a chat; anything earlier is not taken into account. For a new topic, start a new conversation instead of continuing an old one.
- **Vague prompts give vague answers.** Name the ticker, metric and period. "Is NVDA good?" invites opinion; "How did NVDA's gross margin change over the last 8 quarters?" invites facts.
- **Screener results are a sample.** *@StockScreener* lists up to 10 matches. Use the full [Stock Screener](/stock-screener) to see every match.

## Where the data comes from

- **Original sources:** company financial reports (10-K and 10-Q filings and press releases), exchange market prices, Wall Street analyst ratings and estimates, SEC Form 4 insider filings, congressional trade disclosures, company and market news, and economic calendars.
- **How fresh it is:** the same as the page each figure comes from. Quotes refresh throughout the trading day, and financial statements update once a day. Credits reset on the first day of each month.
- **What bestshares.com adds:** the assistant selects the relevant data for your question, writes the summary and comparison, and lists the pages it used.

## Frequently asked questions

### Can the AI Assistant tell me which stocks to buy?

No. It summarizes and compares data, explains metrics and lays out bull and bear arguments, but it does not give personal buy or sell recommendations. Every answer is research help that you should check and weigh against your own goals and risk tolerance.

### Where does the AI Assistant get its information?

It reads the data published on bestshares.com, including company financial statements, stock quotes, analyst ratings and price targets, earnings results and estimates, SEC insider filings, congressional trade disclosures, market news and economic calendars. It does not browse the wider internet, so its answers are only as current as that data. On a free account it uses only the data the site shows free accounts.

### How many questions can I ask?

Each message uses one credit. Free accounts receive 10 credits a month and Pro accounts receive 1,000. Balances reset on the first day of each month, and your current balance is shown next to the send button.

### Can the AI Assistant make mistakes?

Yes. Language models can misread a question, mix up periods or state a wrong number with confidence. The page itself warns that AI can make mistakes. Check important figures on the stock's own pages before you rely on them.

### What are @Agents?

Agents are shortcuts that point the assistant at one kind of task, such as @FundamentalData for financial statements or @StockScreener for finding stocks that match criteria. Type @ in the message box or use the @Agents menu to insert one.

Open the tool: https://bestshares.com/chat
