# askbowtie playbooks > Plaintext for LLMs and crawlers. Canonical HTML: https://askbowtie.com/playbooks/ > Product: askbowtie (askbowtie.com). The brand is always lowercase, one word. "Bowtie" is the persona you ask; the mark is the relational-algebra join operator. ## What playbooks are 19 indexed procedures your AI agent pulls at the point of work, each carrying the trap it avoids. They are shared, versioned memory: the antidote to AI amnesia, the context a fresh agent session does not otherwise have. Your agent stops guessing and follows steps that already work, grounded in the site's real numbers. ## How to call them (MCP) 1. Connect the askbowtie MCP server to your agent once (bearer-token authenticated, standard MCP setup). 2. Call the `playbook` tool with no arguments to get the full index of all 19. 3. Call `playbook({name})` to pull a single procedure and follow it. Or, in plain language to an agent that has the MCP: `pull the askbowtie playbook "" and follow it for your-site.com`. ## The directory ### Investigate (5) Something looks off. Start here so you rule out the artifact before you chase the ghost. #### triage-errors - **Triage errors without chasing ad-block noise** - Use when: get_incidents / the Errors count looks alarming, or before you investigate any error surface. START HERE so you don't burn the audit on third-party noise. - Run it: `pull the askbowtie playbook "triage-errors" and follow it for your-site.com` (or `playbook({name:"triage-errors"})`) #### investigate-a-metric-drop - **Investigate a metric drop (traffic, conversions, signups)** - Use when: Any 'X is down / something looks off' question: START HERE before digging. - Run it: `pull the askbowtie playbook "investigate-a-metric-drop" and follow it for your-site.com` (or `playbook({name:"investigate-a-metric-drop"})`) #### is-this-position-move-real - **Is this GSC position move real?** - Use when: Before reporting ANY search-position change as good or bad news. - Run it: `pull the askbowtie playbook "is-this-position-move-real" and follow it for your-site.com` (or `playbook({name:"is-this-position-move-real"})`) #### judge-a-single-day-or-hour - **Judge a single day / hour (calendar windows)** - Use when: 'How is today going', 'what happened in the last hour', 'compare to yesterday'. - Run it: `pull the askbowtie playbook "judge-a-single-day-or-hour" and follow it for your-site.com` (or `playbook({name:"judge-a-single-day-or-hour"})`) #### diagnose-conversion-undercount - **Diagnose a conversion undercount** - Use when: Owner says 'sales/leads are happening but bowtie shows fewer/none'. - Run it: `pull the askbowtie playbook "diagnose-conversion-undercount" and follow it for your-site.com` (or `playbook({name:"diagnose-conversion-undercount"})`) ### Set up (6) Wire a new signal or a new site correctly the first time. #### set-up-a-funnel-properly - **Set up a funnel (define_flow) properly** - Use when: Asked to create/track a funnel, or you're about to call define_flow. - Run it: `pull the askbowtie playbook "set-up-a-funnel-properly" and follow it for your-site.com` (or `playbook({name:"set-up-a-funnel-properly"})`) #### onboard-a-new-domain - **Onboard a new domain end-to-end** - Use when: A site was just added / 'set up bowtie for X' / data looks empty on a new domain. - Run it: `pull the askbowtie playbook "onboard-a-new-domain" and follow it for your-site.com` (or `playbook({name:"onboard-a-new-domain"})`) #### send-server-side-events - **Send events/conversions from a backend (server-side ingest)** - Use when: Before POSTing to /api/ingest from a backend, Cloudflare Worker, webhook, or any non-browser sender: ad-block-proof conversions, server-confirmed purchases. - Run it: `pull the askbowtie playbook "send-server-side-events" and follow it for your-site.com` (or `playbook({name:"send-server-side-events"})`) #### page-roles - **Analyze pages by ROLE (page tags), not slug-by-slug** - Use when: 'How is my editorial/commercial content doing', 'compare blog vs product pages', or when a site's pages are untagged. - Run it: `pull the askbowtie playbook "page-roles" and follow it for your-site.com` (or `playbook({name:"page-roles"})`) #### page-actions - **Track WHAT visitors do (declared action markers)** - Use when: 'Which plan do people pick', 'how does the download button convert': a click's MEANING, not its position. - Run it: `pull the askbowtie playbook "page-actions" and follow it for your-site.com` (or `playbook({name:"page-actions"})`) #### thorough-analytics-setup - **Instrument a surface COMPLETELY (the analytics-setup orchestrator)** - Use when: 'Make the analytics bulletproof', 'I want to see exactly what users do', 'improve the tracking on our landers', or ANY 'set up / improve analytics' request: the orchestrator above the individual marker playbooks that enumerates the whole capability surface, enforces completeness, and verifies data actually lands. - Run it: `pull the askbowtie playbook "thorough-analytics-setup" and follow it for your-site.com` (or `playbook({name:"thorough-analytics-setup"})`) ### Watch & automate (1) Let bowtie keep watching after the agent session ends. #### automate-a-recurring-check - **Automate a recurring check** - Use when: Before building a cloud routine or external cron: askbowtie almost certainly watches it server-side already. - Run it: `pull the askbowtie playbook "automate-a-recurring-check" and follow it for your-site.com` (or `playbook({name:"automate-a-recurring-check"})`) ### Site health (2) What's broken, and how each page is really used. #### audit-a-site-crawl - **Audit a site's health with a crawl** - Use when: 'Is anything broken on the site', broken links/images, weak titles, noindex, a pre-launch sweep: the DETERMINISTIC check. - Run it: `pull the askbowtie playbook "audit-a-site-crawl" and follow it for your-site.com` (or `playbook({name:"audit-a-site-crawl"})`) #### page-engagement - **Analyze how a page's viewers USE it (section engagement)** - Use when: 'What do people click on this page', 'is the hero/nav/CTA working', 'why does this page get traffic but not convert'. - Run it: `pull the askbowtie playbook "page-engagement" and follow it for your-site.com` (or `playbook({name:"page-engagement"})`) ### Paid (5) Every dollar of ad spend, matched to the payback it bought. #### audit-paid-traffic - **Audit paid traffic (Google Ads spend to real payback)** - Use when: 'Is my ad spend working', 'what's my best/worst keyword', 'where's the wasted spend': the full paid sweep. - Run it: `pull the askbowtie playbook "audit-paid-traffic" and follow it for your-site.com` (or `playbook({name:"audit-paid-traffic"})`) #### audit-campaign - **Audit a Google Ads campaign, the full structured sweep** - Use when: 'Audit campaign X', 'review my campaigns': ends with a prioritized, dollar-ranked action list. - Run it: `pull the askbowtie playbook "audit-campaign" and follow it for your-site.com` (or `playbook({name:"audit-campaign"})`) #### audit-keywords - **Audit keywords, negatives, and search terms** - Use when: 'Find wasted spend', 'what should I add as a negative': waste lives in the SEARCH-TERM report and conversion RATE. - Run it: `pull the askbowtie playbook "audit-keywords" and follow it for your-site.com` (or `playbook({name:"audit-keywords"})`) #### audit-assets - **Audit ad assets (RSA headlines, pinning) + ad-to-lander match** - Use when: 'Are my headlines good', 'why is ad relevance low'. - Run it: `pull the askbowtie playbook "audit-assets" and follow it for your-site.com` (or `playbook({name:"audit-assets"})`) #### keyword-research - **Keyword research grounded in real demand** - Use when: 'Find keywords to target', 'what should I bid on': backed by what already converts, not Keyword-Planner volume. - Run it: `pull the askbowtie playbook "keyword-research" and follow it for your-site.com` (or `playbook({name:"keyword-research"})`) ## Use askbowtie Utilize our playbooks. 1 MCP. Free. Every playbook, every signal, one connection. Start: https://askbowtie.com/start