Platform Digest: Google, Meta & ChatGPT Ads, Sept 3–10

Three separate but thematically connected moves landed across the major ad platforms this week. Google started auto-migrating Search campaigns to AI Max, Meta quietly stripped the last meaningful placement control out of ad sets, and ChatGPT Ads opened self-serve access to any Indian brand willing to spend ₹725 a day. The throughline: manual levers are disappearing, and the platforms are betting advertisers won't miss them. Whether that bet pays off depends almost entirely on the quality of signals you feed the machines.

Google Ads: Your Search Campaigns Are Being Upgraded — Whether You Like It or Not

The migration is live. Starting September 1, Google began automatically moving eligible Search campaigns — specifically those using Automatically Created Assets (ACA) and the campaign-level broad match setting — into AI Max, with all qualifying upgrades expected to wrap up by end of September. Dynamic Search Ads follow a separate timeline, their forced migration pushed out to February 2027.

AI Max is a setting layered on top of existing Search campaigns, not a new campaign type. It enables three AI-driven capabilities: search term matching that expands beyond your keyword list to related queries; text customization that generates new ad copy from your existing assets and landing pages; and final URL expansion that can route clicks to whichever page on your site the AI deems most relevant — not necessarily the one you specified. Google's own data claims campaigns running the full AI Max feature suite see an average of 7% more conversions or conversion value at a similar CPA or ROAS compared to using search term matching alone.

That 7% figure deserves scrutiny. It's measured against search term matching in isolation — not against a well-structured traditional campaign with tight keyword lists and deliberate ad copy that someone has spent years refining. For mature accounts, the "improvement" may look considerably less impressive once you read the fine print on the benchmark.

The deeper issue for any account is data quality. AI Max optimizes on your conversion signals, and browser-side tracking routinely misses a substantial share of conversions. Hand the AI an incomplete picture, and that's exactly what it will bid on your behalf. The most useful thing you can do before the upgrade hits your account is audit your conversion tracking end to end — server-side tagging, enhanced conversions, and CRM-connected offline conversion imports should all be in place before AI Max takes the wheel. Check matched search terms once migration begins, and lock down URL exclusions if you don't want final URL expansion sending users to unintended pages.

For financial services advertisers running Google Search — a mutual fund distributor or a term insurance brand, say — final URL expansion warrants particular attention. Regulatory compliance often demands specific landing pages. Let that slip, and AI may route users to pages that fall short of SEBI or IRDAI disclosure requirements. That's a problem no conversion lift stat will fix.

Meta Ads: Placement Exclusions Are Gone. Here's the Honest Read on What Replaces Them.

On August 21, Meta announced the removal of individual ad placement exclusions at the ad-set level. The change is rolling out as a test on Sales and Leads objectives, with no firm completion date — just in-product notices appearing in Ads Manager. Messenger Stories was already cut from August 27, its delivery automatically redistributed elsewhere.

In practice: you can no longer toggle off Audience Network, Facebook Search, in-stream Reels, or any other surface at the campaign level. Meta's AI decides where your ad runs across Facebook, Instagram, Audience Network, and Messenger. Two alternatives remain — value rules, which adjust bids for specific placements but don't remove them from the auction, and account-level placement controls, which apply across the entire ad account and are the stronger of the two remaining tools.

Value rules come with real limits. Only around seven placements are currently eligible for bid adjustments through value rules, and even a bid reduced to minus 90% doesn't fully exclude a placement; it just makes Meta bid less aggressively for it. A placement you've bid down can still receive delivery. For genuine, standing exclusions driven by brand safety or compliance, account-level controls in Advertising Settings are the right place to enforce them — not value rules.

The cynical read is that some placements have historically received limited spend because advertisers tested them and found they didn't convert. Meta's own data suggests Advantage+ Placements can lower cost per result by giving the algorithm maximum flexibility to find cheap impressions. Cheap impressions, though, aren't always the goal. A brand running insurance or BFSI ads has legitimate reasons to control where its creative appears, and those reasons don't evaporate because Meta has disabled the checkbox.

Before the option disappears from your account: document which placements you were excluding and why. Rebuild those exclusions as account-level controls. Pull a placement breakdown report for the last 90 days across any active Sales or Leads campaigns, identify which surfaces are generating your actual conversions, and use that data to set sensible value rules. Do it now — waiting until the rollout lands on your account means scrambling without a baseline.

ChatGPT Ads: India Goes Self-Serve, and the Entry Bar Is Low

This is the most directly India-relevant development of the week. ChatGPT Ads launched in India in two phases: managed ads went live for Free and Go-tier users from August 27, with WPP and Omnicom as the initial agency partners. Then, on September 4, OpenAI opened the self-serve Ads Manager to any Indian business — no agency, no minimum commitment, no gatekeeping. Daily budgets start at ₹725. India is OpenAI's second-largest ChatGPT market globally, with over 100 million weekly users as of February 2026.

The global numbers behind this launch matter. The platform hit $1 billion in annualized revenue run rate in under 200 days since its February 2026 US launch — fast enough to put it on the radar of every major holding company. More than 50 brands participated in India's launch week alone, spanning lifestyle, travel, education, and digital product categories.

How the ads actually work: sponsored placements appear below ChatGPT's response, clearly labelled and visually separated from the answer. Targeting uses contextual matching — the topic of the current conversation, the user's past chat history, and prior ad interactions — rather than keyword bidding in the traditional sense. Advertisers receive only aggregate reporting: impressions, clicks, and similar metrics. Individual chat content stays private. Ads don't appear for users under 18, or for Plus, Pro, Business, and Enterprise subscribers.

For media planners, the structural point worth sitting with is the nature of the intent being captured. Someone asking ChatGPT "which term insurance plan is best for a 35-year-old with two dependents" has already written their own qualification brief. Query volume on ChatGPT will be lower than Google Search. The context attached to each user at the moment of that query is considerably richer. Judging early campaigns on impressions or raw clicks will give you a misleading read — set cost per qualified lead or cost per conversion as your benchmark from day one, not top-of-funnel metrics that flattering CPMs can inflate.

The platform is still new. No published performance benchmarks exist across industries or campaign types, and the early US pilot required six-figure commitments before self-serve existed. India has skipped that barrier entirely. That makes this an unusually cheap moment to learn how the channel performs in your specific category — before CPMs firm up and every competitor has figured out the playbook.

The Week in Summary

Across all three platforms, the same structural shift is accelerating: automation is becoming the default, and manual controls are being treated as technical debt. That's not inherently bad — AI Max's conversion lift is real, and Meta's placement AI does find efficiency. But automation only performs as well as the signals it's trained on. First-party data quality, conversion tracking completeness, and audience signal hygiene are now the actual levers that determine campaign performance, even when they look like back-end plumbing compared to headline ad formats.

The ChatGPT India launch is the piece that deserves the most attention from brands that haven't yet developed a view on it. The channel is live, accessible, and comparatively uncrowded. That window won't stay open indefinitely — and the brands that build category intuition now, while CPMs are low and competition is thin, will have a meaningful head start when this channel matures.

If you're running paid media across Google, Meta, or newer channels like ChatGPT Ads and want a clear-eyed view of how these platform changes specifically affect your account — your conversion tracking setup, your placement exclusion strategy, your signal quality — Transformics does exactly that kind of category-specific audit for BFSI, real estate, ecommerce, and other verticals we work in. Worth a conversation before September's forced migrations land.