KnitKnot

# Work an issue and ship a playbook

Check a tracked issue, use its linked plan, record the page you shipped, and read the next benchmark.

## At a glance

  1. 1. **Choose an issue**: In Intelligence → Issues, filter the list and open the evidence behind one row.
  2. 2. **Review the plan**: Open a linked playbook and check its proposed work against the captured answers.
  3. 3. **Record and measure**: Mark the playbook done with its exact URL, then return after a later benchmark.

Use [Intelligence → Issues](https://app.knitknot.ai/intelligence/issues) when you need to decide which observed problem to work on. The list keeps issues across benchmark periods, including ones that have faded. A high rank tells you where to start reading; the evidence tells you whether the proposed work fits.

## Check the issue

  1. 1. Search the issue list, or filter its **Type** or **Competitor** column. **Reach** counts observations in the unit shown beside it; **Trend** needs at least two readings. **Demand** is a paid-search estimate of buyer interest, not expected revenue.
  2. 2. Open a row. On **Overview**, read the issue description, supporting observations, and any affected pages. If the **Issue over time** chart appears, select a point to narrow the evidence to that benchmark period. Select it again to clear the period filter.
  3. 3. Open an underlying answer or page before accepting an explanation. A repeated wrong claim, a missing mention, and a page access failure call for different work. The issue can point to the affected material without proving which page caused an AI answer.
  4. 4. Open **Linked playbooks** to see proposed and completed work tied to this issue. An active playbook also appears as **Next step** near the top of the issue.

For a disputed sentence, use [Investigate a claim](/docs/investigate-a-claim/) to check the exact wording and approved fact first.

## Use the playbook

Open a linked plan, or go to [Content → Playbooks](https://app.knitknot.ai/content/playbooks) and search for it. The list has **Active** and **Done** views, with search and filters. If there are no playbooks yet, run a benchmark and inventory crawl before expecting a plan.

On the playbook detail, read **Plan** for the proposed action, target page or suggested path, buyer terms, proof points, and linked issue evidence. A create plan may suggest a new page; a fix plan may concern access or technical health rather than copy. Follow the evidence link back to the issue when the recommendation seems too broad. The plan is a brief for your team to check and implement, not a claim that the work has shipped.

Open **Impact** before shipping to see what will be measured. The view remains available afterward for the later observations.

## Record what you shipped

After the work is live, choose **Mark done** on the playbook. For a new page, enter its published URL. For a page you revised or repaired, check the prefilled URL and replace it if the work went live elsewhere. The URL must start with `http://` or `https://`. It anchors later page checks and measurement.

An **Earn** playbook concerns outreach to a third party. **Log outreach** records that you sent a pitch or requested a correction. Its evidence link is optional. This is your account of the outreach, not confirmation that the third party changed anything.

Marking a page playbook done starts its measurement window. KnitKnot checks the page and its citations at a later benchmark run, not on a separate daily timer. The **Impact** view can then show answer, citation, and linked-issue movement when comparable data exists. If the page is not cited or the issue remains, read the newer answers before choosing another change. Movement after shipping is an observation, not proof that your edit caused it.

Deleting a playbook permanently removes that plan. Its issues stay tracked and can be covered by a future playbook.

Docs navigation
Docs › Guides

Work an issue and ship a playbook

Check a tracked issue, use its linked plan, record the page you shipped, and read the next benchmark.

Updated

At a glance

  1. Choose an issue

    In Intelligence → Issues, filter the list and open the evidence behind one row.

  2. Review the plan

    Open a linked playbook and check its proposed work against the captured answers.

  3. Record and measure

    Mark the playbook done with its exact URL, then return after a later benchmark.

Use Intelligence → Issues when you need to decide which observed problem to work on. The list keeps issues across benchmark periods, including ones that have faded. A high rank tells you where to start reading; the evidence tells you whether the proposed work fits.

Check the issue

  1. Search the issue list, or filter its Type or Competitor column. Reach counts observations in the unit shown beside it; Trend needs at least two readings. Demand is a paid-search estimate of buyer interest, not expected revenue.
  2. Open a row. On Overview, read the issue description, supporting observations, and any affected pages. If the Issue over time chart appears, select a point to narrow the evidence to that benchmark period. Select it again to clear the period filter.
  3. Open an underlying answer or page before accepting an explanation. A repeated wrong claim, a missing mention, and a page access failure call for different work. The issue can point to the affected material without proving which page caused an AI answer.
  4. Open Linked playbooks to see proposed and completed work tied to this issue. An active playbook also appears as Next step near the top of the issue.

For a disputed sentence, use Investigate a claim to check the exact wording and approved fact first.

Use the playbook

Open a linked plan, or go to Content → Playbooks and search for it. The list has Active and Done views, with search and filters. If there are no playbooks yet, run a benchmark and inventory crawl before expecting a plan.

On the playbook detail, read Plan for the proposed action, target page or suggested path, buyer terms, proof points, and linked issue evidence. A create plan may suggest a new page; a fix plan may concern access or technical health rather than copy. Follow the evidence link back to the issue when the recommendation seems too broad. The plan is a brief for your team to check and implement, not a claim that the work has shipped.

Open Impact before shipping to see what will be measured. The view remains available afterward for the later observations.

Record what you shipped

After the work is live, choose Mark done on the playbook. For a new page, enter its published URL. For a page you revised or repaired, check the prefilled URL and replace it if the work went live elsewhere. The URL must start with http:// or https://. It anchors later page checks and measurement.

An Earn playbook concerns outreach to a third party. Log outreach records that you sent a pitch or requested a correction. Its evidence link is optional. This is your account of the outreach, not confirmation that the third party changed anything.

Marking a page playbook done starts its measurement window. KnitKnot checks the page and its citations at a later benchmark run, not on a separate daily timer. The Impact view can then show answer, citation, and linked-issue movement when comparable data exists. If the page is not cited or the issue remains, read the newer answers before choosing another change. Movement after shipping is an observation, not proof that your edit caused it.

Deleting a playbook permanently removes that plan. Its issues stay tracked and can be covered by a future playbook.