Glossary

Each entry gives the word Praised uses and the words other tools use for the same thing. Search works with either.

Question#

A question is a buyer question Praised asks each engine, worded the way a buyer would type it. Other tools call these prompts. Each distinct question is the unit a reading samples.

Discovery and branded questions#

A discovery question does not name the tracked brand; a branded question does. Headline visibility counts discovery questions only, and branded questions are read as knowledge about the brand.

Engine#

An engine is an AI tool or answer surface Praised asks, such as ChatGPT, Gemini, Perplexity or Google AI Overviews. People also say AI assistant, AI model, chatbot or LLM. The engine catalogue names the models behind each surface.

Named rate#

Named rate is the share of answers that name the tracked brand. Other tools call this mention rate or brand mention rate.

Recommendation rate#

Recommendation rate is the share of answers that recommend the tracked brand. An answer can name a brand without recommending the brand.

Share of voice#

Share of voice is the tracked brand’s part of all brand mentions, counting each question once. Share of voice by position also weights each mention by where the mention sat in the answer.

Visibility index#

The visibility index is one score combining share of voice, named rate and recommendation rate under published weights. Act on the part that moved, not on the score alone.

Owned, competitor and independent#

Owned, competitor and independent are the three kinds of cited page. An owned page is one the brand controls, a competitor page belongs to a competitor, and an independent page belongs to another publisher.

Agent credits#

Agent credits are the allowance the Agent spends, set per plan each month. A task costs what the task measurably cost, which can be a fraction of an Agent credit. The plan table lists the monthly Agent credits of each plan.

Reading#

A reading is a measurement summary for a selected window and scope. It includes the figures, their evidence and notes explaining what the data does not support.

Coverage#

Coverage describes which expected answers were collected. It is part of the context for interpreting a figure, not another name for the named rate.

Evidence#

Evidence is the underlying answer, citation, figure or question that a measurement refers to. It lets you inspect the result at its source.

Scope#

Scope identifies the selected subset behind a view, such as a question or an engine. Figures must be compared over compatible scopes.

Sentiment#

Sentiment is the tone of an answer about a named brand. It is distinct from whether the answer recommended that brand.

Estimated range#

An estimated range is the uncertainty shown beside a figure, with the answer count behind the figure when known. Other tools call this a confidence interval or error bar.

Average rank#

Average rank is where the brand sits on average in answers that ranked it. Lower is better, and only answers that placed the brand count. Other tools call this position or average position.

Work item#

A work item is a finding filed as work: a site fix, an owned page or an earned placement. The check that will judge the work is declared when the item is filed.

Check#

A check is how Praised judges shipped work. A shipped page fix is re-checked at once; a fix judged by a later reading settles once the window after its ship day is final.

Verdict#

A verdict is what the check found about shipped work, such as Verified. Only the check sets a verdict.

Programme#

The programme is the record of work on a product: what shipped, what moved, what did not and what waits for its check.