AI Prepared Logo
All Editions

Friday, July 10, 2026  ·  6 stories

AI Prepared Daily

The morning briefing for marketing & data professionals.

Happy Friday. Here’s what’s moving in AI, data, and martech today.

Your Executive Summary

  • Guideline opens first verified spend data on ChatGPT and Perplexity ads — For the first time you can benchmark what advertisers actually spend on AI platform inventory, and its CPM and CPC economics, against the platforms' own projections before committing test budget.
  • Google begins disclosing which ads were made with AI — If you are scaling AI-generated creative, assume users and competitors can now see which ads are machine-made, and monitor whether labeled creative performs differently from unlabeled.
  • IAB: 75% of buy-side leaders say ad measurement is underperforming — If your measurement stack feels broken, it is an industry-wide condition, and the teams pulling ahead are unifying MMM, incrementality, and attribution rather than picking one.
  • Google Ads' July 1 terms quietly authorize AI to run your campaigns — Audit your account-level automation settings now, because the default legal posture has shifted from you approving AI actions to you having pre-authorized them.
  • BigQuery ML forecasting lands in Connected Sheets, no SQL required — Marketing analysts can produce warehouse-grade forecasts and anomaly alerts on spend or conversion data without waiting on a data science queue, so decide who owns sanity-checking the outputs.
  • Akeneo ships Agentic Ziggy, specialist agents for product data quality — Product data quality is exactly the kind of repetitive, rules-plus-judgment work multi-agent systems handle well, so watch whether this pattern spreads to CDPs and other data-quality tooling.

▸ MarTech/AdTech  ·  MarTech Series

Guideline opens first verified spend data on ChatGPT and Perplexity ads

Guideline extended its Ad Intelligence dataset to capture verified, transaction-level ad spend flowing into AI platforms including ChatGPT and Perplexity. The dataset covers roughly $200 billion in annual media investment across 65 countries and is sourced from actual transactions rather than surveys or platform self-reporting.

The bottom line: For the first time you can benchmark what advertisers actually spend on AI platform inventory, and its CPM and CPC economics, against the platforms' own projections before committing test budget.

Check out the full article →

▸ AI/LLM  ·  TechCrunch

Google begins disclosing which ads were made with AI

Google rolled out a disclosure feature that tells people when an ad they are seeing was generated with AI tools. The label surfaces through ad transparency menus, mirroring the AI-content labeling Meta already applies through its About This Ad panel.

The bottom line: If you are scaling AI-generated creative, assume users and competitors can now see which ads are machine-made, and monitor whether labeled creative performs differently from unlabeled.

Check out the full article →

▸ Data Engineering  ·  Google Cloud

BigQuery ML forecasting lands in Connected Sheets, no SQL required

Google made pre-trained TimesFM forecasting generally available inside Connected Sheets, letting analysts run AI.FORECAST and AI.DETECT_ANOMALIES against BigQuery data directly from a spreadsheet. Forecasts and anomaly detection that previously required a modeling workflow now run as spreadsheet functions.

The bottom line: Marketing analysts can produce warehouse-grade forecasts and anomaly alerts on spend or conversion data without waiting on a data science queue, so decide who owns sanity-checking the outputs.

Check out the full article →

▸ Emerging Tools  ·  The Agile Brand Guide

Akeneo ships Agentic Ziggy, specialist agents for product data quality

Akeneo announced Agentic Ziggy, an orchestration layer inside its Product Cloud that coordinates specialist agents for data modeling, schema mapping, enrichment, and continuous quality checks. It is one of the first mainstream PIM vendors to ship a multi-agent architecture rather than a single assistant.

The bottom line: Product data quality is exactly the kind of repetitive, rules-plus-judgment work multi-agent systems handle well, so watch whether this pattern spreads to CDPs and other data-quality tooling.

Check out the full article →

In Other News

AI Prepared Daily is brought to you by JESTR — natural-language analytics for your data warehouse. Learn more →

Get smarter, faster.

AI Prepared Daily arrives every weekday at 7am — subscribe below.

Browse all editions