I Let AI Agents Plan My WNBA Trip

Personal AI agents are evolving from text chatbots into autonomous assistants that research, plan, and purchase. To put these emerging capabilities to the test, tech testers pitted OpenAI's Dots against Meta's Muse. The challenge centered around planning a complete trip to Atlanta for Game 3 of the WNBA Finals supporting the Golden State Valkyries, along with planning a themed brunch and tracking down merchandise.

Both agents showed remarkable multistep problem-solving. Each accurately detected that the game date had moved up 24 hours and warned that primary public tickets were not live, preventing accidental purchases of marked-up resale inventory. However, pricing access differs significantly: Meta Muse offers free baseline access with $20 and $100 tiers, while OpenAI Dots requires a ChatGPT Pro subscription starting at $100 per month.

While both models showed sharp situational awareness, small glitches proved that autonomous delegation is still risky. Muse created a visually striking 10-page brunch menu featuring ube waffles and lavender-honey butter, but had basic counting errors in its grocery tally. Dots showed deeper contextual awareness by warning against a personal school enrollment conflict, but required human intervention to authorize any transactions.

The transition from AI assistant to autonomous purchaser represents the next big frontier for major tech labs. While research, scheduling, and comparison tasks are now remarkably competent, full checkout permissions remain restricted by design. For now, users gain significant time savings in travel logistics, provided they maintain manual sign-off on the final payment.