Categories

Every assistant is scored on the same 14 tests, so a travel bot and a general assistant are judged on identical ground. Each category is its own ranking; open one for the exact task and who leads it.

Core

The seven jobs a personal assistant has to do.

1Carrying out an online taskTest: Book a hotel stay. Completes a real browser or web workflow end to end, not just advice.Untested 2Recommendation qualityTest: Pick a restaurant with constraints. Relevance, taste and constraint-following when it suggests options.Untested 3Purchasing a productTest: Reorder a product on Amazon. Finds, compares and completes (or correctly stages) a purchase.Untested 4Responding to emailsTest: Reply to a scheduling email. Drafts or sends replies in your voice, with the right context and recipients.Untested 5Proactive behaviorTest: Flight day, unprompted. Acts or nudges usefully without being asked every step.Untested 6Running a routineTest: Daily digest for a week. Reliable scheduled and recurring work.Untested 7Third-party integrationsTest: Three tools, one request. Uses connected tools like calendar, inbox and Notion correctly.Untested

Endorsed

Seven more that people asked for, and that separate a good assistant from a great one.

8MemoryTest: Remember a preference a week later. Remembers your preferences, people and past requests across sessions.Untested 9PersonalityTest: A normal week of texts. Has a voice worth talking to; reads like a contact, not a form.Untested 10Phone callsTest: Call a business and get an answer. Makes real phone calls on your behalf and reports back.Untested 11Multiplayer / groupsTest: Plan dinner in a group chat. Works in group chats with several people at once.Untested 12Chained tasksTest: Flight check-in chain. Strings several steps across tools into one job, like a flight check-in.Untested 13Proactive restraintTest: Know when not to. Acts before you knew you needed it, and knows when not to.Untested 14Content creation / gamesTest: Make something for the group. Makes images, video or games on request.Untested

How scoring works

Each category has one published task, run with the same wording for every assistant and scored 1–10 against written anchors. A score exists only after a logged run with a date and evidence; nothing is scored from marketing claims. Core and Endorsed means exclude N/A.

A tested score1 – 10

Not tested yet—

Doesn't apply to this productN/A

Benchmark versionv0.1