Ad Ops Quick Start

Get Google Ad Manager reporting, revenue reconciliation, and yield monitoring running in your AI workflow. Three pre-built tools, one Google connection, five minutes.

Prerequisites

• A Kyew account (sign up here)
• Kyew connected to your AI client (installation guide)
• A Google account with GAM access

Step 1: Connect Google

Connect your Google account with Ad Manager access:

tool(action="conn_connect", provider="google", bundles=["admanager"])

Click the authorization URL, grant access, and you're connected. Your credentials are encrypted and auto-refreshed — you'll never paste a token again.

Step 2: Install Your First Tool

Browse the tool catalog and install the GAM Report tool:

tool(action="catalog")

Install it:

tool(action="install", template_id="tmpl_gam_report")

The tool is immediately active and pre-wired to your Google connection.

Step 3: Run a Report

List your GAM networks:

gam-report(action="networks")

Run a revenue report for the last 7 days:

gam-report(action="report",
  network_code="YOUR_NETWORK_CODE",
  dimensions=["DATE", "AD_UNIT_NAME"],
  metrics=["AD_SERVER_IMPRESSIONS", "REVENUE", "AVERAGE_ECPM"])

You'll get structured data with impressions, revenue, and eCPM broken down by date and ad unit — plus an HTML email version you can schedule.

Step 4: Add Revenue Reconciliation

Install the revenue reconciler for comparing GAM against SSP data:

tool(action="install", template_id="tmpl_revenue_reconciler")

Configure it with your network:

revenue-reconciler(action="configure",
  network_code="YOUR_NETWORK_CODE",
  threshold_pct=5)

Pull GAM data

revenue-reconciler(action="pull_gam")

Import SSP data

revenue-reconciler(action="import_ssp",
  ssp_name="index_exchange",
  ssp_data=[
    { ad_unit: "/homepage/leaderboard", impressions: 50000, revenue: 175.50, date: "2026-04-01" },
    { ad_unit: "/article/sidebar", impressions: 30000, revenue: 45.00, date: "2026-04-01" }
  ])

Run reconciliation

revenue-reconciler(action="reconcile")

The tool compares GAM and SSP data, flags discrepancies above your threshold, and uses AI to analyze likely root causes (reporting lag, timezone mismatches, IVT filtering, etc.).

Step 5: Add Yield Monitoring

Install the floor price monitor:

tool(action="install", template_id="tmpl_floor_price_monitor")

Configure and take a snapshot:

floor-price-monitor(action="configure", network_code="YOUR_NETWORK_CODE")
floor-price-monitor(action="snapshot")

View trends and get AI-powered floor price recommendations:

floor-price-monitor(action="analyze")

The AI analyzes your eCPM and fill rate data to recommend where to raise floors (high fill + low eCPM = money on the table) and where to lower them (low fill = rejecting demand).

Step 6: Schedule Daily Reports

Run any tool on a schedule with email delivery:

tool(action="schedule",
  tool_id="YOUR_TOOL_ID",
  schedule={
    hours: [8],
    timezone: "America/New_York",
    email_summary: true,
    args: { action: "report", network_code: "YOUR_NETWORK_CODE" }
  })

Your team gets a formatted HTML email every morning at 8am.

All Ad Ops Tools

What's Next

• Connect more services — add SSP connections, GitHub, WordPress
• Build a custom tool — describe what you need, AI writes the code
• Schedule automation — daily briefings, weekly digests, alert monitoring
• Bring your own AI — add Anthropic or OpenAI keys for Claude and GPT in your tools