diff --git a/es/optimize/feedback.mdx b/es/optimize/feedback.mdx index 7a1b6f3ce..0fd6c5756 100644 --- a/es/optimize/feedback.mdx +++ b/es/optimize/feedback.mdx @@ -170,4 +170,10 @@ Revisa tus datos de feedback para: - **Identificar contenido exitoso**: Las páginas con más feedback positivo muestran qué funciona bien en tu documentación. - **Priorizar mejoras**: Las páginas con más feedback negativo indican qué contenido podría necesitar atención. -- **Tomar medidas**: Realiza actualizaciones en la documentación basadas en el feedback directo de los usuarios. \ No newline at end of file +- **Tomar medidas**: Realiza actualizaciones en la documentación basadas en el feedback directo de los usuarios. + +
+ ## Automatizar las actualizaciones de docs a partir del feedback +
+ +Activa el [workflow **Feedback docs updates**](/workflows/enable#feedback-docs-updates) para que el agente redacte mejoras de la documentación a partir del feedback contextual reciente en una programación recurrente. El agente clasifica las páginas según el feedback negativo y detallado, resume los temas y abre una pull request con actualizaciones específicas. \ No newline at end of file diff --git a/fr/optimize/feedback.mdx b/fr/optimize/feedback.mdx index 382acab47..3a730ff51 100644 --- a/fr/optimize/feedback.mdx +++ b/fr/optimize/feedback.mdx @@ -171,4 +171,10 @@ Analysez vos données de feedback pour : - **Identifier les contenus les plus efficaces** : les pages qui reçoivent le plus de retours positifs montrent ce qui fonctionne bien dans votre documentation. - **Prioriser les améliorations** : les pages qui reçoivent le plus de retours négatifs indiquent quels contenus pourraient nécessiter une attention particulière. -- **Passer à l'action** : mettez à jour votre documentation en vous appuyant sur les retours directs des utilisateurs. \ No newline at end of file +- **Passer à l'action** : mettez à jour votre documentation en vous appuyant sur les retours directs des utilisateurs. + +
+ ## Automatiser les mises à jour de docs à partir des retours +
+ +Activez le [workflow **Feedback docs updates**](/workflows/enable#feedback-docs-updates) pour que l'agent rédige des améliorations de la documentation à partir des retours contextuels récents, selon une planification récurrente. L'agent classe les pages en fonction des retours négatifs et détaillés, résume les thèmes et ouvre une pull request avec des mises à jour ciblées. \ No newline at end of file diff --git a/optimize/feedback.mdx b/optimize/feedback.mdx index 53f150464..f501979b5 100644 --- a/optimize/feedback.mdx +++ b/optimize/feedback.mdx @@ -155,4 +155,8 @@ Review your feedback data to: - **Identify successful content**: Pages with the most positive feedback show what works well in your documentation. - **Prioritize improvements**: Pages with the most negative feedback indicate what content might need attention. -- **Take action**: Make documentation updates based on direct user feedback. \ No newline at end of file +- **Take action**: Make documentation updates based on direct user feedback. + +## Automate doc updates from feedback + +Enable the [**Feedback docs updates** workflow](/workflows/enable#feedback-docs-updates) to have the agent draft documentation improvements from recent contextual feedback on a recurring schedule. The agent ranks pages by negative and detailed feedback, summarizes themes, and opens a pull request with targeted updates. \ No newline at end of file diff --git a/workflows/enable.mdx b/workflows/enable.mdx index 3fe0eb7f5..4f4537baa 100644 --- a/workflows/enable.mdx +++ b/workflows/enable.mdx @@ -74,4 +74,32 @@ You can change the target languages at any time by opening the workflow's settin After saving the workflow, click **Translate now** to run it on demand using the currently saved target languages. The workflow must be enabled to trigger a manual run. Manual runs do not affect the workflow's regular trigger. +## Feedback docs updates + +The **Feedback docs updates** workflow turns recent [contextual feedback](/optimize/feedback) on your published docs into targeted documentation improvements. The agent reviews the pages with the most negative or detailed feedback in the trigger window, drafts updates that address the themes it finds, and opens a pull request with the changes. + +Use this workflow to close the loop between reader feedback and your docs without manually triaging every comment. + +### Prerequisites + +- The **Contextual feedback** add-on is enabled on the [Add-ons](https://dashboard.mintlify.com/products/addons) page. +- Your project has received contextual feedback in the trigger window. If no qualifying feedback exists, the workflow runs but does not open a pull request. + +### How it runs + +- **Trigger**: Scheduled. Defaults to weekly on Monday at 09:00 UTC. Edit the schedule from the workflow's settings. +- **Signal**: Pages are ranked by the number of "not helpful" votes, written comments, and total feedback in the window. The agent receives a redacted excerpt of representative comments along with links back to each item in the feedback dashboard. +- **Output**: A single pull request that updates the highest-signal pages. Each run summary lists the pages the agent changed and the feedback items it cited as evidence. + +Feedback that mentions email addresses, API keys, or other secrets is automatically redacted before the agent sees it. + +### Configure the workflow + +When you enable **Feedback docs updates**, you can configure: + +- **Instructions**: Optional guidance appended to the workflow's base prompt. Use this to set tone, scope, or pages to avoid. +- **Update mode**: Open a pull request for review, or automatically merge changes. See [Update mode](#update-mode). + +This workflow does not accept context repositories or notifications. It always runs on the connected docs repository. +