Use-case guide
A personal AI assistant should not leave finished work buried in chat. Use receipts to create an approval queue where every item has a source, draft, status, owner, and next action.
A reminder says something needs to happen. A receipt explains what the assistant already did. That difference matters when the assistant checks websites, reads messages, drafts replies, or generates publishable assets.
If the work starts in SMS, iMessage, or a customer thread, preserve the original context. A text message AI assistant can capture the ask, but the approval queue should also show who asked, when, and what the assistant inferred.
Browser evidence and repeated checks belong in the receipt. This is where computer-use cache becomes useful.
The assistant can prepare the response, page, or follow-up, but the queue should make the approval state impossible to miss.
Approved receipts can become messages, briefs, or focused pages through AI agent website building.
The simplest durable workflow is capture, enrich, draft, decide, then log the outcome. Each step should update one receipt card instead of scattering context across tools.
Record the original ask, channel, user, urgency, and expected output.
Attach links, message excerpts, browser state, screenshots, and assumptions.
Prepare the reply, plan, queue note, or generated page from the evidence.
Approve, edit, ask for more context, schedule, publish, or archive.
Every queue card needs source, state, draft, owner, and outcome.
Include the original request and the evidence used to create the draft. Link back to message context, browser context, or files.
Use explicit states: new, researching, drafted, waiting for approval, approved, sent, published, blocked, archived.
Make clear whether the AI assistant, user, operator, or customer is responsible for the next step.
After approval, log what happened. Super is useful when the same workflow spans messages, browser work, and generated outputs.
This guide is based on personal AI assistant workflows, message-native intake, browser-use agents, approval queues, and agent-generated publishing. Relevant Super workflows include Super, text message AI assistance, computer-use cache, and AI agent website building.