KnitKnot

# Manage benchmark questions

How to add, star, and archive the prompts your benchmarks run, and why a stable set helps comparison.

Your prompt library holds the buyer questions sent to the engines selected for a benchmark. Active prompts form the measuring set. Keep that set steady when you want to compare two runs closely.

Open [Prompt library](https://app.knitknot.ai/prompts). Use **Topics** to inspect topic coverage and **Prompts** to work with individual questions.

## What's in the library already

Benchmark setup builds the initial library from approved competitors, buyer topics, capabilities, and roles. Search-demand data helps allocate coverage across topics. The prompts themselves remain visible for review before and after a run.

Each prompt belongs to a [subject](/docs/subjects/): your brand or one of its products. Check the subject filter before changing a product's prompts.

## How many questions each topic gets

Your plan sets a prompt allowance. The initial setup divides that allowance across approved buyer topics using available demand signals. [The question library](/docs/benchmark-questions/) explains how prompt counts affect what can be measured.

Prompts you add manually and prompts you star remain identifiable in the library. Starring is a priority marker; only an **active** prompt runs.

If a prompt is archived, its past results remain available. Check the active set after changing your competitors, products, or plan so you know what the next run will measure.

## Add a question

Use **Add prompt** to write a buyer question. For example: “Which tools support SSO for a team with several workspaces?” Check its subject and wording before saving it. An added prompt joins the library and can be included in later runs.

## Star, activate, and archive

Two independent controls shape what a benchmark actually runs:

  • - **Status:** a prompt is **active** or **archived**. Only active prompts run. Archive one to retire it without losing its history.
  • - **Star:** mark a prompt as important. Starring does not activate an archived prompt or change whether it runs.

Archiving is almost always the right move over deleting. Deleting is permanent and drops the question's history; archiving keeps the record and lets you reactivate later.

## Keep the library stable

Adding or archiving prompts changes the set that future runs measure. For a close before-and-after comparison, keep the active prompts and selected engines steady. Other factors can still change an AI answer between runs.

Next: [manage the competitors](/docs/manage-competitors/) your questions benchmark you against, or [run a benchmark](/docs/run-and-schedule-benchmarks/) once the library looks right.

Docs navigation
Docs › Guides

Manage benchmark questions

How to add, star, and archive the prompts your benchmarks run, and why a stable set helps comparison.

Updated

Your prompt library holds the buyer questions sent to the engines selected for a benchmark. Active prompts form the measuring set. Keep that set steady when you want to compare two runs closely.

Open Prompt library. Use Topics to inspect topic coverage and Prompts to work with individual questions.

What’s in the library already

Benchmark setup builds the initial library from approved competitors, buyer topics, capabilities, and roles. Search-demand data helps allocate coverage across topics. The prompts themselves remain visible for review before and after a run.

Each prompt belongs to a subject: your brand or one of its products. Check the subject filter before changing a product’s prompts.

How many questions each topic gets

Your plan sets a prompt allowance. The initial setup divides that allowance across approved buyer topics using available demand signals. The question library explains how prompt counts affect what can be measured.

Prompts you add manually and prompts you star remain identifiable in the library. Starring is a priority marker; only an active prompt runs.

If a prompt is archived, its past results remain available. Check the active set after changing your competitors, products, or plan so you know what the next run will measure.

Add a question

Use Add prompt to write a buyer question. For example: “Which tools support SSO for a team with several workspaces?” Check its subject and wording before saving it. An added prompt joins the library and can be included in later runs.

Star, activate, and archive

Two independent controls shape what a benchmark actually runs:

  • Status: a prompt is active or archived. Only active prompts run. Archive one to retire it without losing its history.
  • Star: mark a prompt as important. Starring does not activate an archived prompt or change whether it runs.

Archiving is almost always the right move over deleting. Deleting is permanent and drops the question’s history; archiving keeps the record and lets you reactivate later.

Keep the library stable

Adding or archiving prompts changes the set that future runs measure. For a close before-and-after comparison, keep the active prompts and selected engines steady. Other factors can still change an AI answer between runs.

Next: manage the competitors your questions benchmark you against, or run a benchmark once the library looks right.