July 31st-August 6th // Estimated Reading Time: 9 minutes

The most popular AI bootcamp in the marketing industry is back! The AI Accelerator helps marketers level up on all things AI for marketing and advertising so they can make a meaningful impact on their business, fast.

We only have 50% capacity left! Enroll now using promo code EarlyBirdsLearnAI for 20% off.

A debate is brewing over the risks and rewards of serving ads to AI agents🔒
Source: Business Insider
August 5th, 2026

Summary: Ad tech startup Mobian is helping Time run a new kind of ad on its website. Time already built stripped-down, text-only pages for AI crawlers to read instead of its normal site. Now those pages also carry ads, styled as FAQ blocks🔒 aimed at the LLMs. They are labeled “sponsored”. Ally Bank and the Project Management Institute are the first two brands to test the format, hoping to influence how AI tools like ChatGPT and Gemini talk about them. The move has sparked debate over whether this is a legitimate new marketing channel or a new way for bad actors to build their own agent-ad networks, which they could use to push scams or hurt competitors in AI. Ally Bank CMO Andrea Brimmer said the move was easy to justify since AI-driven traffic to its site is up 9x year over year and converts 3.5 times better than traditional search traffic.

Opinion: AI models are sold as a neutral source of answers, and this is a test of that promise. But nobody, including the platforms, knows how much these ads will actually influence what an AI model tells someone about a brand. Once it can be measured, brands will pile in or the AI platforms will just shut this down. Our bet is on the latter.

The case for and against agencies betting on tokens the way they bet on media🔒
Source: Digiday
July 31st, 2026

Summary: Agency holding companies have absorbed the cost of running AI tools for clients for two years but rising token usage and costs are making that unsustainable. Some agencies are now applying their principal media playbook to tokens: buying AI tokens in bulk, pricing the risk, and reselling them to clients with a margin built in. The case for it is that clients get price certainty and agencies can finally recover AI costs they've been carrying. The case against it is that clients can't see what the agency paid or the markup, and agencies could get stuck holding tokens they can’t use. In a Digiday poll of 58 people, 42% said agencies shouldn't do this at all, and only 10% called it a legitimate opportunity.

Opinion: Agencies are always going to look for new margin opportunities. This opportunity is particularly tricky, because the agencies are also fashioning themselves as the AI tech providers. To tell a customer you’re valuable to them and deserve their business because of your AI tool, and then to tell them you’re arbitraging the usage of that tool, is a slippery slope. As always, marketers need to make sure how they’re paying their agency is 100% transparent in order to avoid conflicts of interest. 

Jellyfish launches AI Ads Optimization in Share of Model

What It Does: Jellyfish’s Share of Model tool (which tracks how AI chatbots describe a brand) now turns AI-perception gaps into ad optimizations across Google Performance Max, ChatGPT ads, TikTok, DV360, YouTube and Reddit. It also has an always-on assistant pushing recommendations to SEO, PR, and ecommerce teams in real time. 

Quick Take: Be wary of agency AI tools that “collapse” planning → buying → measurement. They sound nice in theory, but without a human-in-the-loop checkpoint, they could merely be padding agency margins without actually helping the marketer drive better outcomes. 

You can now ask ChatGPT about your Snap ad campaigns

What It Does: Snap will let marketers connect Snap Ads data directly to their AI agents in Claude, ChatGPT, and Gemini. Advertisers can ask natural-language questions about their Snap campaigns, like which ones need attention or which ones are performing well, instead of pulling reports manually. Administrators must approve each AI agent. 

Quick Take: More walled garden platforms letting outside agents in, as opposed to forcing marketers to use their homegrown agents, is a good thing. Could AI agents be the walled gardens’ trojan horse?

Newton Research's new AI layer predicts campaign outcomes before you spend

What It Does: Newton Research's new Unlimited Analytics is an agentic AI intelligence layer that sits across a marketer’s media planning, activation, campaign optimization, and measurement. It includes causal modeling that lets teams ask questions like “what happens if we shift 20% of budget from social to programmatic” and get an answer before committing spend. It also speeds up marketing mix modeling from quarters to days. Horizon Media has seen a 10% to 20% increase in return on ad spend since adopting Unlimited Analytics. 

Quick Take: AI-driven causal modeling and MMM is a huge value unlock for marketers. There’s always a high-level of uncertainty associated with the results anyways, why waste all that time and money trying to get it perfect?

The Setup: Japan Airlines wanted to shift ad spend toward its highest-value routes, but its team needed more information about how AI models like Gemini talked about its brand versus competitors and where it was losing ground in those conversations.

The AI Solution: Japan Airlines' agency Jellyfish built a tool called Share of Model. It tracks how often LLMs mention Japan Airlines versus competitors, what tone they use, and how that compares to the search themes the brand was already targeting. Japan Airlines used those insights on its Google Performance Max campaigns, adding more than 150 new search themes and shifting spend toward its best-performing routes.

The Results:

• 56% increase in conversion rate on PMax campaigns

• 14% improvement in return on ad spend

• 21% drop in cost per acquisition

• 12.5% cut in media spend

Why This Matters: Japan Airlines' website traffic dropped 10% this year as AI search sends fewer people directly to the site, and the visitors who do visit are already deep into research and close to booking, not early-stage browsers. That's the backdrop for Japan Airlines' next move: using Share of Model to optimize ads on ChatGPT, Reddit, TikTok and YouTube. Japan Airlines will be launching its first paid ChatGPT ads this month. If people are researching on AI platforms instead of airlines websites, showing up well in those conversations matters as much as the site itself.

Your Action: 

Before you spend on ChatGPT ads, check what you're already getting for free. Pull your referral data (GA4 or equivalent) segmented by AI domains — chatgpt.com, perplexity.ai, gemini.google.com, claude.ai — and look at what pages those sessions land on and how they convert. The key insight is that the traffic that comes from AI is closer to booking than anything from organic search. If your AI-referred visitors also book at a higher rate, your ad creative for AI platforms should go straight to the CTA and the special offer, because those visitors already did the research and just need to be closed.

The U of Digital AI Edge Newsletter is intended for AI Edge Pro subscribers, but occasional forwarding is okay!

To subscribe visit Uof.Digital/Newsletters/ai-edge or AI Edge Pro, or contact us directly for group subscriptions.

And remember, U of Digital helps teams drive better outcomes through structured education on critical topics in marketing / advertising like AI, mar tech, programmatic, CTV, commerce media, and more. Interested in learning more about how we can supercharge your team? Let’s talk!