Agents

Gilito AI agents are automated monitoring tools that continuously watch your assets and market conditions, evaluate rules you define, and notify you when something important happens. Instead of manually checking every asset every day, agents do the watching for you and alert you only when action may be needed.

What Are Agents?

An agent is an AI-powered automation that runs on a schedule, checks specific conditions against your assets or the broader market, and takes action when those conditions are met — typically by sending you a notification. Think of agents as tireless assistants that monitor the market 24/7 on your behalf.

Each agent is configured with:

• A target — What the agent monitors (a specific asset, a watchlist, a portfolio, or the entire market)
• Conditions — The rules that must be met to trigger the agent (e.g., signal changes, score thresholds, price movements)
• A schedule — How often the agent checks its conditions (e.g., daily after market close, hourly, on specific days)
• Actions — What happens when conditions are met (notifications via email, in-app alerts, or future webhook integrations)

Types of Agents

Gilito provides several types of agents, each designed for a different monitoring use case:

How Agents Work

Behind the scenes, agents follow a straightforward execution cycle:

1. Scheduled trigger — At the configured time (e.g., daily after market close), the agent wakes up and begins its check
2. Data retrieval — The agent fetches the latest data for its target: current signals, Gilito Scores, price changes, portfolio positions, or market-wide metrics
3. Condition evaluation — The agent evaluates each of its configured conditions against the retrieved data. Conditions use a flexible rule system that can check for signal changes, score thresholds, price movements, and more
4. Action execution — If one or more conditions are met, the agent executes its configured actions: sending an email notification, creating an in-app alert, or (in the future) calling a webhook endpoint
5. Logging — Every agent execution is logged, whether or not conditions were met. You can review the execution history to see when agents ran and what they found

Agent Configuration & Customization

Every agent is fully customizable. When creating or editing an agent, you configure the following:

Name & Description

Give your agent a descriptive name so you can easily identify it in your list. For example, "FAANG Buy Signal Alert" or "Portfolio Score Drop Warning." The description field lets you add notes about the agent's purpose for your own reference.

Target Selection

Choose what the agent monitors:

• Specific asset(s) — Monitor one or more individual assets by ticker symbol
• Watchlist — Monitor all assets in one of your watchlists (automatically adapts as you add/remove assets)
• Portfolio — Monitor all holdings in one of your portfolios
• Market-wide — Scan across all assets in a supported index

Condition Rules

Define the conditions that trigger the agent. You can combine multiple conditions with AND/OR logic. Common conditions include:

• Signal changes to Buy, Sell, or Hold
• Gilito Score rises above or drops below a threshold
• Signal strength exceeds a minimum level
• Price changes by more than a specified percentage in a single day
• A new top-ranked strategy appears from a different strategy family

Schedule Configuration

Set how often the agent runs. Options include:

• Daily — Runs once per day after market data is updated (most common)
• Weekly — Runs once per week on a day you choose
• Custom — Advanced scheduling for specific days or intervals

Notification Channels

When an agent's conditions are met, it needs a way to reach you. Gilito supports the following notification channels:

Best Practices for Using Agents

• Start simple — Begin with a single signal monitoring agent on your watchlist before creating complex multi-condition agents
• Use descriptive names — Name agents clearly so you can immediately understand what each one does when reviewing your list or reading a notification
• Avoid notification overload — Be selective with your conditions. An agent that triggers every day for every asset will quickly become noise. Focus on meaningful events
• Review execution history — Periodically check your agents' logs to make sure they are triggering as expected and that the conditions are still relevant to your investment approach
• Combine with watchlists — Target agents at your watchlists rather than individual assets. This way, the agent automatically adapts as you update your watchlist
• Test before relying — After creating a new agent, use the manual trigger feature to verify it behaves as you intend before waiting for the scheduled run