Track Las Vegas flight prices and find return flights
The author asked Muse to check Vegas fares daily and find returns to San Diego; it gave only three options and ignored his preferences, so he found it underwhelming. Task: The user instructed Meta Muse to track outbound JSX flight prices to Las Vegas and find return flights to San Diego within a date window. Muse navigated booking sites, set up daily price checks, and attempted to search for award flights using Amex points via British Airways. The human provided credentials and corrected the agent's logic regarding points transferability, but Muse failed to complete the BA login due to a Captcha obstruction.; tools: Meta Muse. Human role: Source evidence is unavailable for specific human roles; the user provided preferences and credentials but faced issues with the agent's execution.; limitations: Muse lacks persistent memory of user preferences like airline exclusions or points programs, requiring explicit instruction each time. It struggles with complex loyalty program logic (e.g., Amex to BA transfers) and fails when UI elements like black bars obscure Captcha puzzles. Integration with airline sites is unstable, prone to errors and login blocks..
Tools
Published on Agent Work Atlas:
Overview
The author asked Muse to check Vegas fares daily and find returns to San Diego; it gave only three options and ignored his preferences, so he found it underwhelming.
Workflow
The user instructed Meta Muse to track outbound JSX flight prices to Las Vegas and find return flights to San Diego within a date window. Muse navigated booking sites, set up daily price checks, and attempted to search for award flights using Amex points via British Airways. The human provided credentials and corrected the agent's logic regarding points transferability, but Muse failed to complete the BA login due to a Captcha obstruction.
Reported result
Outbound price tracking worked correctly. Return flight searches initially ignored user preferences, offering Frontier and layover options. Muse incorrectly stated Amex points could not be used for the suggested carriers. When prompted to check British Airways, it got stuck on a Captcha due to a UI overlay issue, and subsequent flight search attempts failed with errors.
Human role
Source evidence is unavailable for specific human roles; the user provided preferences and credentials but faced issues with the agent's execution.
Limitations
Muse lacks persistent memory of user preferences like airline exclusions or points programs, requiring explicit instruction each time. It struggles with complex loyalty program logic (e.g., Amex to BA transfers) and fails when UI elements like black bars obscure Captcha puzzles. Integration with airline sites is unstable, prone to errors and login blocks.
Try it
Act as a travel booking assistant. I need to track flight prices for [Route: Origin to Destination] on [Date: Outbound] and find return options for [Date Window: Return]. My constraints are: [Airline Preferences: e.g., avoid Frontier], [Loyalty Program: e.g., Amex Points], and [Max Layovers: e.g., 0]. Step 1: Check current prices for the outbound flight on [Airline: e.g., JSX] and set up a daily notification if prices change. Step 2: Search for return flights matching my constraints. If I have points, check if the airline is a direct partner or if I can transfer points to a partner (e.g., Amex to British Airways) to book. Step 3: If a login or Captcha is required, pause and ask me for credentials or manual intervention. Do not guess. Step 4: Present options in a table with Price, Airline, Layovers, and Points Cost. Before executing, confirm you understand my airline exclusions and points strategy. If you cannot verify points transferability, state 'Unknown' rather than assuming it is impossible. If a site blocks you, report the error immediately.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.