Caller Trust is the feedback display above a fictional AI caller during a conversation in Scam With Your Friends. Gameplay shows a percentage and a mood description such as “trusting.” Use that information alongside the caller's actual response to judge how the current interaction is going.
There is no verified universal trust formula or minimum percentage that guarantees success. A high trust display also does not prove that a payment, remote connection, or other desktop task has finished. Check the action and its result separately.
Read the meter with the conversation
The trust display describes the current caller interaction. It is most useful when you compare it with what the caller just said and with the objective you are trying to complete.
For example, a caller asking for clarification tells you something different from a caller agreeing to continue. A number alone does not explain that distinction. Read the dialogue before deciding whether to clarify your last message, continue the task, or change how you are approaching this particular character.
Callers have different personalities. Gameplay shows some arriving friendly and others already suspicious or irritated. Their initial behavior should not be treated as a universal starting value, and one recording does not establish a fixed response for every character.
Keep three kinds of progress separate
| Display or response | What it helps you assess | What still needs checking |
|---|---|---|
| Caller Trust percentage and mood | The state of the conversation | What the caller actually agreed to |
| The caller's reply | Information or consent expressed in dialogue | Whether the associated task advanced |
| The application or earnings result | The outcome shown by the game | Whether it satisfies your current workday objective |
This separation prevents a common misunderstanding: a conversation can sound successful while the relevant application still needs attention. The caller may provide information, but you may still need to use it in the appropriate in-game field.
Conversely, continuing to chat after a completed action does not establish that additional money is arriving. Check the actual result rather than estimating progress from a pleasant response or a percentage movement.
Use consistent dialogue instead of a supposed universal script
Follow what the current caller has said. If they raise a concern, address that concern before introducing a different request. If they correct a detail, keep the correction in mind instead of repeating the earlier mistaken version.
These are ways to follow a real-time conversation. They are not a proven mathematical method for increasing the meter by a set amount. No reliable per-sentence gain, ideal voice pitch, or guaranteed opening line is established.
A line that worked in a gameplay video is one example of an interaction with a fictional character. The game's real-time AI can produce a different answer in your session. Do not assume copying the same words forces the same trust percentage or reward.
Check speech recognition when a response seems misplaced
Sometimes the issue is the recognized input rather than the conversational approach. Inspect the dialogue text if the caller reacts to a name or phrase you did not intend.
Correct the misunderstood point or try typed dialogue to compare the result. Typed input remains supported, and current fixes addressed speech startup issues while retaining it. If a typed message behaves differently from an incorrectly transcribed spoken one, that is useful information about the stage where the misunderstanding occurred.
It is not proof of a specific microphone fault. Recent fixes reduced recognition delays and filtered extra false music captions, but there is no guarantee of perfect transcription in every environment.
Do not turn unsupported percentages into rules
A single guide may claim that requests should wait until a particular percentage, or describe a percentage as “full trust.” No official formula or threshold has been verified for this guide.
Avoid treating such a number as a requirement. It can make you overlook the caller's actual objection or delay an action whose state is already visible in the application. The supported approach is to read the meter as feedback, then check dialogue and task results.
There is also no verified rule that a low number makes recovery impossible, that one greeting guarantees a boost, or that every caller responds to the same tone. None of those outcomes should be promised.
Keep trust separate from technical status
An AI failure or temporary usage restriction is not a caller deciding to distrust you. Hotfix 2 made those conditions display clear status messages instead of fake caller dialogue.
If you see such a message, read the system status. Changing your fictional story or repeating a request does not establish a way to restore the service or bypass a restriction. Likewise, missing spoken playback can involve voice loading rather than a trust decision.
Caller Trust is useful when normal conversation is running. It is only one part of understanding a call, and its practical value comes from reading it together with recognized text, the character's reply, and the outcome shown by the desktop application.