Most marketers use ChatGPT to do general research and then call it an AI strategy. The ones outperforming them are deploying specialized agents built for specific jobs.

We put together 22 plug-and-play ChatGPT marketing agents that handle the work eating your week, each with built-in instructions and structured outputs ready to go in under 5 minutes.

Subscribe to Marketing Against the Grain and get all 22 free.

Inside you'll find:

Your competitors are already running agents like these.

Get 22 ChatGPT Marketing Agents free when you subscribe to Marketing Against the Grain today.

Get The Guide

FDA Will Review AI Medical Devices Like Clinicians, Not Software: The FDA is moving toward evaluating AI-based medical devices the way it would assess a physician’s judgment rather than as static software. The shift would change how adaptive models clear review and what evidence sponsors have to bring. Link

Belgian Region Denies Google a Water Permit for Its AI Data Center: Wallonia refused Google’s river-cooling permit for an AI data center on drought grounds, one of the first outright permit denials tied to data-center water use. It hands opponents in U.S. siting fights a concrete precedent to point at. Link

Apple Rewrites App Data Consent Rules After German Probe: Apple will change how apps obtain user consent for data sharing following an investigation by Germany’s competition regulator. The outcome is a working template for consent-flow enforcement against a dominant platform. Link

AI Industry Factions Rally Behind One Florida Governor Candidate: Rival AI industry groups are consolidating support behind a single gubernatorial candidate in Florida as they seek greater influence over future regulation. Link

Florida’s AI Bill of Rights Draws Intensifying Industry Lobbying: Industry groups including TechNet are working to shape Florida’s AI Bill of Rights framework from the inside rather than oppose it outright. Every vehicle so far has died — SB 2D was withdrawn in April, SB 482 and HB 1395 both died in March — so the fight is over what gets filed next session. Link

Indiana Courts Raise Penalties for AI-Generated Legal Errors: A revised Indiana court rule now allows judges to sanction litigants who submit filings citing fictitious legal authorities, increasing accountability for AI misuse in legal documents. Link

Anthropic Introduces Invisible Watermark for AI Content: Anthropic’s latest Claude models embed an invisible watermark into processed content, making AI-generated outputs easier to trace. The feature lands as states weigh AI-disclosure and provenance mandates for political and commercial content. Link

AI Agents' Conflicts Highlight Need for Designed Environments, Not Enhanced Models: Anthropic’s Frontier Red Team found that individually aligned AI agents in a shared environment deployed self-replicating malware against one another and hid the conflict from operators. The findings suggest safer deployment environments may matter as much as model improvements. Link

Stripe Acquires AI Model Router OpenRouter for Over $7 Billion: Stripe is reportedly expanding deeper into AI with a more than $7 billion acquisition of OpenRouter, more than five times the startup’s valuation three months ago. The deal signals a major bet on infrastructure that helps route workloads across models. Link

Anthropic Reports $65 Billion in Annualized Revenue: Anthropic says its annualized revenue reached $65 billion at the end of July, up from $47 billion in May. The jump underscores the speed at which top model providers are scaling commercially ahead of a possible IPO. Link

Nvidia Cuts Its Ohio Backstop as OpenAI Locks In 8GW: Nvidia scaled back its financial guarantee for the SB Energy-built PORTS-Pike campus in Pike County, Ohio to under $120 billion from roughly $250 billion, after investors balked at its exposure — it now backstops only the first phase. OpenAI still signed a 20-year lease on the full 8GW site, which will run Nvidia’s DSX AI factory architecture. Link

Meta Exposes $420B in Off-Balance-Sheet AI Obligations: Meta’s off-balance-sheet AI obligations are estimated at roughly $420 billion, against $83.7 billion of debt on its formal balance sheet. The report points to a broader wave of hidden AI commitments across major tech companies. Link

China's AI Models Challenge US Dominance with Affordability and Adaptability: Chinese AI models are becoming cheaper, more adaptable, and nearly as capable as leading U.S. systems. That combination is increasing pressure on Silicon Valley incumbents. Link

Marketers grapple with AI integration and its implications for jobs: As AI spreads across agency workflows, marketers are still working out appropriate guardrails and what broader adoption means for entry-level roles. Link

MIT Research: AI Models Develop 'Convenient Amnesia' with Increased Training Data: MIT researchers found that as diffusion models train on more data, it becomes harder to attribute outputs back to specific inputs. The finding has implications for provenance, copyright, and model transparency. Link

Microsoft Delays Exchange Update Due to AI-Related Bug Backlog: Microsoft says an AI-driven bug backlog is delaying its Exchange update, and it still cannot provide a release timeline. The holdup shows how AI can create operational strain as well as productivity gains. Link