Compare your watchlist
Scan selected stocks or crypto over a period you choose. Compare price moves, investigate the news behind them, and see which developments relate to your research question.
Premiss agents
AI agents that
trade for you.
Give your agent a trading brief. It researches the market, makes decisions, and places trades through your connected account. You set the markets and position limits, and choose whether it acts automatically or asks for your approval.
“Research my crypto watchlist and trade when the evidence supports it. Stay within my position limits and explain each decision.”
- Markets
- Bitcoin · Ethereum · Solana
- Schedule
- Every morning at 09:00
- Action
- Automatic trading
- Account
- Your connected trading account
Runs on your schedule. Records each trading decision.
Market research
Price moves, filings,
and the context behind them.
Follow a company, compare a watchlist, or revisit a thesis. The agent connects market observations with the sources you need to assess them.
Read the source material
Review company filings and announcements alongside market data. Follow source links, check publication dates, and ask the agent to distinguish a reported fact from its interpretation.
Challenge a trading idea
Ask what contradicts your thesis, which assumptions need testing, and what would change the decision. Save findings and corrections for the next review.
Scheduled tasks
From a question
to a recurring task.
Give one agent several assignments: a morning market scan, a filing review, and a weekly thesis check. Each keeps its own schedule and report destination.
- 01
Describe the assignment
Name the markets, the question, and the sources that matter. Ask for a market scan, a filing review, or a second opinion on your trading thesis.
- 02
Save the schedule and permissions
Choose when the task runs, where reports go, and whether the agent can suggest or take a trading action. Start with research or paper trading when evaluating a new approach.
- 03
Read the report and refine the brief
Review what changed since the last run. Follow the sources, correct an assumption, or narrow the question for the next check.
Trading permissions
Decide what the
agent can do.
Keep a task focused on research, review each proposed trade, or authorize automatic trading with defined limits.
Reports and feedback
Receive findings with source links and open questions. Follow up in chat and keep the trading decision with you.
Trade proposals
Read the proposed position, its rationale, and the evidence that would invalidate it. Decide whether to approve the trade.
Automatic trading
Set the permitted markets and position limits. The agent sends trades to your connected service, where you manage order execution.
Research coverage and trading support are different choices. You can investigate stocks and crypto; order execution depends on the markets and instruments supported by your connected destination. Explore integrations.
Example briefs
Four assignments
to start from.
Adapt the markets, sources, and schedule to the question you want answered.
A morning crypto market scan
Compare the latest completed daily candles for Bitcoin, Ethereum, and Solana. Summarize the largest moves, check the relevant news, and link to the sources. Flag anything you could not verify.
- When
- Daily at 09:00 · your timezone
- Output
- Research report · chat or Discord
A stock research assistant
Review NVIDIA’s latest company filings. Compare the revenue outlook and stated risks with the previous filing. Explain what changed and cite the relevant sections.
- When
- Every day · selected company
- Output
- Filing comparison · saved findings
Feedback on a trading thesis
Revisit my Bitcoin trend thesis each week. List the evidence for and against it, identify assumptions that no longer hold, and suggest a test for each proposed change.
- When
- Every week · saved research brief
- Output
- Thesis review · suggested next steps
An agent with trading permission
Research my selected crypto markets and propose positions with a clear rationale. Paper trade the approach first, then use my connected destination within the trading limits I authorize.
- When
- Your chosen research frequency
- Output
- Trade decisions · rationale and record
Questions
Research, schedules,
and trading.
What does a Premiss AI agent do?
It uses market data, filings, web research, and saved findings to investigate a question. It can return a report, propose a trade for your review, or trade through a connection you authorize. Each task has its own instructions and schedule.
How do I set up an automated market scan?
Describe the markets, question, and sources to check. Save the task with a frequency, timezone, and report destination. You can run it on demand or let it recur, then edit or pause it when your brief changes.
Can I use an agent for research without letting it trade?
Yes. Research tasks can end with a briefing, a saved finding, or feedback on a thesis. You only need a trading connection when you want the agent to place trades.
Can an AI trading agent make trades on its own?
Yes. Enable automatic trading, connect a supported destination, and define the markets, position limits, and conditions for action. The receiving service executes the orders using its account settings. You can pause automation or require your review before a trade.
What is the difference between an AI agent and a trading bot?
A strategy bot evaluates defined rules against market data. An AI agent can investigate new evidence and use it to form a trading decision. They can reach different conclusions from the same market. Use a strategy bot when you want fixed, testable rules, or an agent when the task calls for ongoing research and judgment.
Does scheduled research continue when I close the browser?
Yes. Saved tasks run in the background. The agent keeps its assignment and recorded findings between runs, then sends the next report to your selected destinations. You can return later to review the work.
Where can I receive the agent’s reports?
Read reports in the agent’s chat, collect them in the Desk, your shared research space, or send them to a connected Discord channel. Each task can report to a different place.
How should I evaluate an agent’s trading decisions?
Start with a clear objective and a paper trading period. Compare the agent with a simple benchmark and your existing strategy over the same dates and markets. Review drawdowns, turnover, fees, and the evidence behind each decision. A convincing explanation or a profitable historical period does not establish future performance.
Your first assignment
Give your agent
a trading brief.
Choose the markets, connect an account, and set its limits.