Four data points dropped this week that marketers running lean teams should care about. Agency fee structures are shifting fast, AI is steering shoppers away from default brand choices, and Amazon just handed its entire ad stack to an autonomous agent.
Agency fee models are collapsing
A WFA and Agency Mania Solutions study of 69 multinationals with a combined $147 billion in marketing spend found that labor-based fees are now the standard model for just 19% of brands, down from 54% in 2011 and 33% in 2022. Fixed-fee or output models have climbed to 33%, and labor-plus-performance arrangements sit at 21%.
Only 20% of brands have already changed commercial terms because of AI, but 61% plan to. The same study found that strong briefing (rated 5.6 out of 6) drives agency performance far more than financial incentives (rated 3.9). If AI is cutting the hours behind your agency’s work, the research says renegotiate around outputs now, and treat the brief as seriously as the fee.
In-house AI creative at mid-market scale
Wyndham Hotels, Opella, and BetMGM are all running in-house AI creative production today without Unilever-level budgets. Wyndham now produces 15 times more assets and has cut concepting and production time by 75%. Opella’s 85-person team has produced more than 20,000 pieces of content, roughly 20 times its pre-AI output.
All three brands still use agencies for large above-the-line campaigns. BetMGM keeps regulated content like betting odds away from AI entirely. The pattern worth copying: define what AI must never generate before you scale, not after.

Chatbots are switching shoppers’ brands
Attentive’s 2026 State of AI in Retail survey of 3,054 US adults found that 68% used a general AI chatbot for at least one shopping task in the past three months. Of those, 62% bought a different product or brand than they normally would based on a chatbot recommendation.
But the purchase rarely happens inside the chat window. 93% of AI-assisted shoppers still checked outside sources before buying, mainly reviews (37%) and price comparisons (36%). And 60% prefer to complete the purchase on the brand’s own site or app, versus 17% who want to buy inside the chatbot. The implication: your product detail pages, review scores, and pricing need to hold up under scrutiny from a shopper who arrives already half-persuaded by an AI recommendation, because that page is where the switch gets confirmed or reversed.
Amazon hands its ad stack to an AI agent
At unBoxed, Amazon merged its DSP and Ads Console into a single platform called Amazon Ads Agent. The new Full-Funnel Campaigns put sponsored ads, display, video, and streaming TV into one AI-run campaign. Advertisers set products, creative, and budget. They cannot choose individual channels or audiences.
DVA+ combines Sponsored Display, Sponsored TV, and programmatic buying starting in late October. Amazon has not committed to disclosing how individual auction prices are set, while it simultaneously fights an FTC lawsuit over ad surcharges. Before handing Amazon the whole funnel, run your own incrementality test alongside its reporting.
⚡ Quick hits
- Contentstack launched Canoe, a free-to-start tool that tracks what ChatGPT, Gemini, Perplexity, and Claude say about your brand.
- ChatGPT ads remain stuck at test budgets, with agencies reporting missing conversions and inconsistent reporting.
- Mixpanel’s Agent Intelligence connects AI agent conversations to what customers actually do next.
- Reddit kills RSS feeds in November and public API access in 2027, citing AI scraping.
- 300 publishers including Condé Nast and Hearst are lobbying Congress for a federal bill that would require stealth AI crawlers to identify themselves.
- 69% of B2B buyers switched vendor because of AI guidance, according to Circle & Square’s new CMO guide.
- Kargo opened agentic ad buying to small teams via Karlo, reporting a 21% ad recall lift for the Ad Council.
