diff --git a/.github/workflows/blog-check.yml b/.github/workflows/blog-check.yml
index 13915444c9..48c5a1bb31 100644
--- a/.github/workflows/blog-check.yml
+++ b/.github/workflows/blog-check.yml
@@ -39,7 +39,7 @@ jobs:
 
       - name: Install dependencies
         if: steps.changed-files.outputs.has_files == 'true'
-        run: npm install ai @ai-sdk/anthropic zod
+        run: npm install --legacy-peer-deps ai @ai-sdk/anthropic zod
 
       - name: Run grammar check
         if: steps.changed-files.outputs.has_files == 'true'
diff --git a/apps/web/content/articles/organize-meeting-notes.mdx b/apps/web/content/articles/organize-meeting-notes.mdx
new file mode 100644
index 0000000000..96fe255e46
--- /dev/null
+++ b/apps/web/content/articles/organize-meeting-notes.mdx
@@ -0,0 +1,40 @@
+---
+meta_title: "How to Organize Meeting Notes So You Can Actually Find Them Later"
+display_title: "7 Ways to Organize Meeting Notes"
+meta_description: "Most people capture meeting notes fine. Finding them three weeks later is the real problem. Here are 7 systems that work, and how to choose between them."
+author:
+  - "Harshika"
+featured: false
+category: "Guides"
+date: "2026-02-22"
+---
+
+You send meeting notes. Three weeks later, nobody can find them. You schedule another meeting to redecide what you already decided.
+
+Teams pick note systems for capture features and skip retrieval design. How you search determines whether you'll use the tool after the first week.
+
+## Automated Capture and Transcription
+
+AI meeting assistants record and transcribe without requiring manual note-taking. Otter and Fireflies both search across projects and dates, though transcripts live on their servers and export options vary by plan. Char saves markdown locally on the device. Most teams underestimate how often meetings end without anyone taking notes at all, which is the core problem automated transcription solves.
+
+## Structured Retrieval
+
+Notion stores notes as database rows filterable by project, date, or attendee, while Obsidian links documents so relationships between meetings become visible over time. Logseq enforces regular review through daily notes with built-in task management. Markdown imports from tools like Char or Otter feed into any of these.
+
+A spreadsheet with one row per meeting (date, attendees, project, decisions, action items, notes link) lets you filter by project or find every conversation involving a specific person. Decision tracking across conversations and reconstructing past rationale become straightforward with a centralized index.
+
+CRMs like Salesforce, HubSpot, Attio, and Folk display complete conversation history when opening a contact. Most teams write notes elsewhere and import them because CRM text editors are clunky for real-time use. Sales teams, recruiters, and account managers use CRMs for this because customer conversations need a central location.
+
+Wiki-style tools like Confluence, Notion, and GitHub Wiki require a designated content owner who reviews and corrects on a schedule. Teams that launch wikis without assigning ownership tend to abandon them within six months as contradictions accumulate and nobody takes responsibility for accuracy.
+
+## Search Method and Workflow
+
+Pick tools by how you search: dates, people, or topics. Tools chosen for features alone get abandoned when the search method conflicts with how the team actually thinks about past conversations.
+
+Email-based note storage works if your workflow centers on your inbox. Notes sit with other correspondence, and Gmail or Outlook search finds them by person, date, or subject. Multiple meetings per day clutter the inbox and notes get buried.
+
+GTD requires writing raw notes during meetings, then processing them into action items and reference material. Most teams skip processing. Raw notes pile up and the weekly review never happens.
+
+## File Format and Portability
+
+Char saves markdown locally, and those files open in Obsidian, Notion, VS Code, or any editor without conversion. Most teams change their note-taking setup every couple of years. Proprietary formats make migrations painful.