SANDR is the main interface. The user does not begin by choosing a report or finding the right dashboard. They begin by typing, speaking, or asking the same way they would ask a trusted operator.
The conversation is connected to the workspace. SANDR can read venues, tasks, plans, signals, files, reports, proof, and the page context behind the work. When an answer needs detail, it can show the relevant object instead of sending the user searching for it.
Capture the app opening directly to the SANDR conversation screen. Show the cold-start prompt, composer, mic, voice conversation, check-in, attachment, and send controls.
What the user does
A manager can ask, "What needs attention today?" SANDR checks the workspace and brings back the places, decisions, blocked work, and stale items that matter now.
They can ask, "Why is North Cafe at risk?" SANDR can show the venue, the signals behind the score, the open work, and the missing proof.
They can say, "Morning rush was rough again. Drinks backed up, one person was late, and the handover was unclear." SANDR can treat that as a floor note and prepare the next step.
Capture the app opening directly to the SANDR conversation screen. Show the cold-start prompt, composer, mic, voice conversation, check-in, attachment, and send controls.
Show a short grounded answer with source chips for venue, signal, plan, and stale work.
Show SANDR answering a venue-specific question and opening or linking to the venue context.
Show SANDR drafting an important change as a preview with a clear confirmation step before anything is saved.
From question to action
Outcome
The user can move from a natural question to grounded context and a prepared next action without starting in a dashboard.
Walkthrough
We will map how SANDR would open in the conversation, show the right context, prepare the work, and keep confirmation with the user.