Skip to content
SilktideHelp

Reputation screen

Visibility tells you whether assistants name you. Reputation tells you how they talk about you when they do.

Sentiment explains the measurement. This screen is where you read it, brand by brand.

The table only scores you and your primary competitors that have been interviewed: asked probing questions about themselves. That is the one population where the rates can be compared. Recommendations almost never criticize, so a workspace answer and an interview answer are not the same kind of evidence.

Brands nobody has interviewed yet still appear below the table, quoted from answers about the market. You can interview a primary competitor from this screen.

The table

One row per interviewed primary brand, showing how those interview answers broke down across the four states - Praise, Mixed, Neutral, Criticism - as four percentages that add up to 100%.

  • The bar is the shape of the opinion. Two brands can share a praise share and look nothing alike here.
  • The four columns are the same four numbers the bar draws, so the table, the bar, and the quotes behind them can never disagree. Sentiment explains the states.
  • Answers is how many answers the row rests on. A reading built on eight answers is not the same evidence as one built on eight hundred.

Brands with the highest praise share lead the list.

Clicking any brand opens its detail view.

What AI says

Below the table, you and each primary competitor get a short preview of quoted praise and criticism, and a link to the full evidence. Interviewed brands are quoted from their own questions. Brands that have not been interviewed yet are quoted from market answers, and labeled that way.

A quote is evidence wherever it was found. Rates are not: the table never ranks an interview rate next to a workspace rate.

Secondary competitors

If you track any secondary competitors, they get their own table next, quoted from answers about the market rather than interviews. Each row shows what AI says about that brand: praise and criticism on separate lines when both exist.

These readings are comparable with each other, but not with the interview table above: market answers barely criticize anyone. Clicking a row opens the same detail view every brand has.

Watched brands

If you watch any brands, they get the same compressed table under secondary competitors. Watching costs nothing - the mentions are already stored - so this is a standing view of how assistants describe the brands you keep an eye on without competing against them.

Clicking a row opens the same detail view every brand has.

Also being talked about

At the bottom, the brands assistants brought up that you do not track at all. Each row shows what AI says about that name: praise and criticism on separate lines when both exist.

Watch this brand sits under the name and moves it into the watched table above in one click. The actions menu on the right still lets you add it as a competitor, or fold it into a brand you already track. These rows are not links: there is no detail view until you watch or add them.

Progress

Each brand's detail view charts how its four states have moved over time. Switch between daily and weekly points to see whether the mix is changing rather than only what it is right now.

A brand's detail view

The screen the rest of Reputation exists to route to.

Quoted sentences sit in two columns: Praise and Criticism. These are the same walls as the preview on Reputation. An answer that both praised and criticized contributes a quote to each column, with a note on the card that the same answer also argued the other way.

The detail view uses the same date, assistant and brand filters as Reputation. Quotes come from both interview answers and market answers, so a sentence you saw in the table or the preview is searchable here.

Quotes that single a brand out as the best or worst outright are marked Superlative and shown first - they are the sentences most worth reading, though they never change a rate.

Every quote is checked against the stored answer before it is shown, so anything here can be found in the response it links to.

Search on the filter row narrows both columns to matching quotes. Answers that named the brand without judging it have no quote, so they do not appear here. Traits is where that unevaluated content is reported.

When the brand has not been interviewed, the quotes still come from answers about the market, and the screen says so.

What you can do here

  • Open any tracked brand to read what is actually being said about it.
  • Watch an untracked name from the list at the bottom, to keep an eye on it without competing.
  • Interview a brand that does not have a score yet, so it can join the table.
  • Watch the four states over time on each brand's detail view, to see whether a swing is a blip or a trend.

If it is empty

The table is empty until at least one primary brand has been interviewed. The quotes below can still fill from market answers. Widening the date range is the other common fix.

Last updated

Was this page helpful?

Reputation screen | Silktide Help