# Google Ads + AI glossary

Every term you meet when an AI starts managing your ads — defined in plain language, with what it means for automation.

- [Ad Strength: What the Label Measures (Not Quality)](https://adcopilot.cloud/glossary/ad-strength): What Ad Strength measures in Google Ads, why it is guidance rather than a performance score, and how to audit the label across every ad with an AI agent.
- [Ads Advisor: Google's Built-In AI Agent, Defined](https://adcopilot.cloud/glossary/google-ads-advisor): Definition of Google's Ads Advisor, what it does inside the Ads interface, the structural incentive question, and how it compares with external AI agents.
- [Ads in AI Overviews and AI Mode: The New Inventory](https://adcopilot.cloud/glossary/ads-in-ai-overviews): Definition of ads in Google's AI Overviews and AI Mode, eligibility through existing campaign types, and a clearly labelled ledger of confirmed versus unknown.
- [Agentic AI in Advertising: Beyond the Chat Window](https://adcopilot.cloud/glossary/agentic-ai-advertising): Definition of agentic AI in advertising, how the tool-use loop works in an ad account, where sane operators sit on the autonomy spectrum, and a hype filter.
- [AI Hallucination in PPC: When the Numbers Are Made Up](https://adcopilot.cloud/glossary/ai-hallucination-ppc): Definition of AI hallucination in PPC terms, the invented-metric failure mode, how tool grounding shrinks it, and the receipts habit that catches the rest.
- [AI Max for Search: Google's Query-Matching Overhaul](https://adcopilot.cloud/glossary/ai-max-for-search): Definition of AI Max for Search, the three features inside it, the controls and reporting that remain, and how AI helps you decide and monitor the opt-in.
- [Asset Group: PMax's Answer to the Ad Group](https://adcopilot.cloud/glossary/asset-group): Definition of the Performance Max asset group, the control it trades away compared with ad groups, and how to structure and review asset groups with an AI agent.
- [Auction Insights: Watching Competitors Without Spyware](https://adcopilot.cloud/glossary/auction-insights): Definition of the Google Ads auction insights report: first-party competitive data from shared auctions, its six metrics, and its limits as intelligence.
- [Audience Signals: Suggestions, Not Targeting](https://adcopilot.cloud/glossary/audience-signal): Definition of Performance Max audience signals, why they are starting hints rather than targeting, and how an AI agent helps build and review them.
- [Audit Trail: Proving What Your AI Did and Didn't Do](https://adcopilot.cloud/glossary/audit-trail): Definition of an audit trail for AI tools, what a complete trail records, how it complements Google's change history, and what to ask any vendor about logging.
- [Budget Pacing: The Two Clocks in Every Account](https://adcopilot.cloud/glossary/budget-pacing): Definition of budget pacing in Google Ads, the daily-to-monthly maths and overdelivery rules, run-rate projection, and the mid-month check an AI agent can run.
- [Change History: The Account's Black-Box Recorder](https://adcopilot.cloud/glossary/change-history): Definition of Google Ads change history, how source attribution and reversal work, and why AI-made changes land in the log under the authorising user's own name.
- [Conversion Actions: The Data Contract Bidding Runs On](https://adcopilot.cloud/glossary/conversion-action): Definition of Google Ads conversion actions, the settings that change everything, the classic misconfigurations, and how an AI agent audits the whole set at once.
- [Enhanced Conversions: Recovering What Cookies Lost](https://adcopilot.cloud/glossary/enhanced-conversions): Definition of enhanced conversions in Google Ads, how hashed first-party matching works, the web and leads variants, and how AI helps diagnose and verify the gap.
- [GAQL: Google Ads Query Language, Now Written by AI](https://adcopilot.cloud/glossary/gaql-google-ads-query-language): Definition of GAQL, the Google Ads Query Language: anatomy of a query, the differences from SQL, and the verification habits that matter when AI writes it for you.
- [Human-in-the-Loop: The Pattern, Precisely Defined](https://adcopilot.cloud/glossary/human-in-the-loop): Precise definition of human-in-the-loop, the three loop positions, checkpoint design by blast radius, and what good HITL looks like in an MCP client.
- [Impression Share: Your Market Share in the Auction](https://adcopilot.cloud/glossary/impression-share): Definition of impression share: impressions received over estimated eligible impressions, with the lost-to-budget and lost-to-rank diagnostics that explain invisibility.
- [Keyword Match Types in the AI Matching Era](https://adcopilot.cloud/glossary/keyword-match-types): Definition of Google Ads keyword match types as intent interpretations: what exact, phrase and broad really permit today, and how to patrol the boundary with search terms.
- [Manager Account (MCC): One Login, Many Accounts](https://adcopilot.cloud/glossary/manager-account-mcc): Definition of a Google Ads manager account (MCC): the account type that links many accounts under one login, and why it doubles as the AI scope boundary.
- [MCP Client: The AI App That Calls the Tools](https://adcopilot.cloud/glossary/mcp-client): Definition of an MCP client: the AI application that connects to MCP servers and calls tools, and how clients differ on approvals, transports and tool limits.
- [MCP Server: The Tool List Is the Capability List](https://adcopilot.cloud/glossary/mcp-server): Definition of an MCP server: the program that exposes tools an AI can call, why its tool list is its capability list, and how to evaluate one in five minutes.
- [Model Context Protocol (MCP), Explained for Marketers](https://adcopilot.cloud/glossary/model-context-protocol): Definition of the Model Context Protocol for marketing teams: the open standard that lets any AI client call tools on any connected system, including Google Ads.
- [N-Gram Analysis: Finding Patterns in Search Terms](https://adcopilot.cloud/glossary/n-gram-analysis): Definition of n-gram analysis for PPC: aggregating search term performance by word fragment to turn scattered queries into evidence, now runnable conversationally.
- [OAuth for Google Ads: What That Consent Screen Grants](https://adcopilot.cloud/glossary/oauth-for-google-ads): Definition of OAuth for Google Ads: delegated, revocable API access without password sharing — what the adwords scope covers and what it cannot touch.
- [Prompt Injection: When Your Data Attacks Your Agent](https://adcopilot.cloud/glossary/prompt-injection): Definition of direct and indirect prompt injection, the ads-specific vector through account data, a realistic risk assessment, and the mitigations that matter.
- [Quality Score: The Diagnostic Google Gives You Free](https://adcopilot.cloud/glossary/quality-score): Definition of Google Ads Quality Score: a 1–10 repackaging of three component diagnostics, why the number misleads, and how to work component-level at account scale.
- [Refresh Token: How AI Tools Stay Signed In](https://adcopilot.cloud/glossary/refresh-token): Definition of a refresh token: the long-lived OAuth credential that mints short-lived access tokens, and why its storage and encryption are the key vendor questions.
- [Remote MCP Server vs Local: Which and Why](https://adcopilot.cloud/glossary/remote-mcp-server): Definition of a remote MCP server versus a local one: same protocol at a different address, and why credential custody is the real deciding question for ad accounts.
- [Responsive Search Ads: 15 Headlines, One Machine](https://adcopilot.cloud/glossary/responsive-search-ad): Definition of a responsive search ad (RSA): a bag of headline and description parts Google assembles per auction, and how to write and audit the parts.
- [Search Terms Report: The Most Underread Ads Report](https://adcopilot.cloud/glossary/search-terms-report): Definition of the Google Ads search terms report: the queries you actually paid for versus the keywords you bid on, and how AI made reading it at scale practical.
- [Seasonality Adjustments: Briefing the Bid Algorithm](https://adcopilot.cloud/glossary/seasonality-adjustment): Definition of Smart Bidding seasonality adjustments, the short-event use case they are built for, the misuses that damage learning, and how AI derives the number.
- [Shared Budgets: One Wallet, Many Campaigns](https://adcopilot.cloud/glossary/shared-budget): Definition of shared budgets in Google Ads, the right and wrong campaign groupings, the difference from portfolio bid strategies, and the AI allocation review.
- [Smart Bidding: What Google Automates, What It Can't](https://adcopilot.cloud/glossary/smart-bidding): Definition of Smart Bidding: Google's auction-time automated bidding, the signals it uses, its dependency on your conversion data, and how it differs from an AI agent.
- [Target CPA: What It Is and How to Set It Honestly](https://adcopilot.cloud/glossary/target-cpa): Definition of Target CPA: the Smart Bidding strategy that holds an average cost per conversion, and the economics homework that decides whether it works.
- [Target ROAS: Definition, Math, and Sane Targets](https://adcopilot.cloud/glossary/target-roas): Definition of Target ROAS: the Smart Bidding strategy that holds an average return on ad spend, the margin maths for setting it, and the volume trade every target implies.
- [Tool Calling: How an LLM Actually Changes Your Bids](https://adcopilot.cloud/glossary/tool-calling): Definition of tool calling in LLMs, the request-execute-return loop, why the separation between model and execution makes AI access to ad accounts controllable.

---

This is the markdown copy of https://adcopilot.cloud/glossary. The index of the site's pages is https://adcopilot.cloud/llms.txt.
