EvidCue Concept preview
A companion for live conversations

Live conversation cues. On your phone.

EvidCue is a concept for answer cues on a separate phone during permitted live conversations. Explore how approved wording, source evidence, and reviewed AI drafts could be kept distinct.

THE IDEANot every question needs the same engine. Reach for approved wording first, source evidence next, and a clearly marked AI draft only when the question calls for reasoning.

EvidCueCOMPANION VIEW
What is our approved short answer?
Approved card · local

The prewritten, approved response appears here. No new wording is generated.

Status: approved wording · source shown in product
Route 01 / simplest question
Illustrative view · not a live answer service
One question, three paths

Match the route to the question.

Select a route to see what the companion view would show. The order describes relative complexity and expected response posture, not measured latency.

Illustrative routing only. No numeric speed or accuracy claim; real performance depends on the implementation and device. Read the guide to approved cards, source evidence, and AI drafts.

Different answers, different authority

The label matters as much as the cue.

EvidCue should make it obvious whether you are looking at an approved line, a sourced passage, or generated reasoning. The person speaking remains responsible for what they say.

ApprovedPrewritten wording from a controlled card; never silently rewritten by an AI model.
EvidenceA source passage to inspect, with provenance; not permission to quote it as approved policy.
DraftAn AI suggestion, labelled as such; no certification, approval, or definitive answer implied.
Separate viewA phone stays outside an ordinary computer-window share. It may still be visible to a camera, device mirror, or other capture.
Two audiences, one boundary

For teams and individuals who can use assistance.

For organizations

Explore a proposed way to distinguish controlled answer cards from retrieved source passages and AI drafts during authorized customer, training, or internal conversations. Organization-specific content, governance, and validation would need to be built before deployment.

For individuals

Explore a proposed phone companion for permitted calls or conversations. Check the rules of the setting first; do not use outside assistance where it is prohibited, including restricted assessments or interviews.

Straight answers

Frequently asked questions

What is EvidCue?

EvidCue is a concept for a separate-phone companion that displays answer cues during permitted live conversations. This website is an interactive preview, not a live answering service.

How would EvidCue choose an answer?

The proposed design has three routes: bounded local matching to approved cards, retrieval of relevant source passages, and clearly labelled AI drafts for complex questions. Source evidence is not approved wording; AI drafts require human review.

Is EvidCue available as a production-ready service?

No. This website is a concept preview. An older local prototype demonstrated a phone companion view and bounded matching, but the standalone three-route service and domain-specific validation are not established.

Does a separate phone keep cues out of a screen share?

An ordinary share of a computer window need not include a separate phone, but cameras, mirroring, whole-device capture, and other setups can expose it. Use assistance only where permitted and disclose it where required.

Build status

What this preview shows.

This is a concept website. An older, separate Logan-specific local prototype demonstrated a phone companion view and bounded cue matching. The EvidCue three-route illustration here is not connected to a live answering service or a validated, approved corpus. Audit-grade logs, approval/version metadata, and domain-specific validation remain future work. Explore the routes as an explanation of the design, not as answers to rely on in a real conversation.