Essential Insights
-
China plans all looked alike: Every AI agent suggested Beijing → Xi’an → Chengdu → Shanghai, fitting the four-hotel limit—but raising concerns that AI could send travelers to the same places.
-
Dots won the itinerary test: It delivered a clear, useful plan (though another task exposed a niche weakness); Muse came close behind and was the easiest—and free—to try.
-
Empathy varied when the user couldn’t sleep: Grok gave the strongest response, while Gemini and Claude missed the emotional context with advice that felt impersonal or irrelevant.
-
Best and worst experiences differed by design: Muse was the most accessible, Claude Duet had the best interface, and Instinct was the most annoying; cost, privacy, and access remain trade-offs.
Six agents, one China route
I gave six AI agents the same challenge: plan a two-week China trip for two, with flights, hotels, food and nightlife. I wanted history and big-city energy, but no more than four hotel stays.
They all chose Beijing, Xi’an, Chengdu and Shanghai. The route makes practical sense: imperial landmarks, the Terracotta Army, Sichuan cuisine and a lively finish. It also raises a question. If AI tools keep steering travelers to the same places, they could create a new kind of Lonely Planet effect.
Muse earned the highest itinerary score, 5 out of 5. Its tables made the plan easy to scan. Dots scored 4, but stretched its draft to 24 pages. Grok also scored 4 and offered useful tips. Gemini Spark, Instinct and Claude Duet each scored 3.
Helpfulness matters after hours
I also tested how the agents handled a late-night moment when I could not sleep. Grok gave the strongest response, scoring 5 out of 5. It offered to set a reminder to check on my mother. Dots scored 4 and suggested turning anxious thoughts into a short checklist.
Some answers missed the moment. Gemini Spark pointed to my empty morning calendar, as if a clear schedule could explain away sleeplessness. Claude Duet focused on old work and lunch tasks. Both scored 2. A useful agent needs more than task management; it needs to respond with care and good judgment.
Dots wins, with caveats
OpenAI Dots performed best overall in these tests. Its drafts felt polished, and its late-night suggestion was practical. Still, it did not lead every individual task. Grok handled the sleep test better, while Muse offered the easiest setup and costs nothing.
Access and trust matter, too. Dots sits behind an expensive paywall, so many users may want to wait for a free or lower-cost option. Muse is accessible, but Meta’s data history may give some people pause. Claude Duet has a clean, dedicated app, though it requires a wait-list sign-up and a paid plan. Instinct’s messenger-only design left me feeling too far from the work it did.
For now, Dots is the strongest all-around pick from this one-day test—not a universal winner. The right agent depends on your budget, privacy comfort and preferred way to work.
Discover More Technology Insights
Stay informed on the revolutionary breakthroughs in Quantum Computing research.
Explore past and present digital transformations on the Internet Archive.
AITechV1
