For as long as digital marketing has existed, the click has been the currency. You paid for it in search, earned it with content and chased it with every campaign. If people clicked, you were winning. If they didn’t, you changed something until they did.
This week, the click lost a lot of its value.
Chartbeat data showed Google Search referrals to publishers falling 40.2% in a year. A court filing quoted an OpenAI engineer saying that “no matter how prominently we show the links, users won’t click.” At the same time, OpenAI launched always-on AI agents called Dots, Shopify opened its checkout to browser agents, Amazon renamed its ad platform Amazon Ads Agent, and Google said it will use a new open standard for hotel booking inside AI Mode.
The single thesis running through this week’s news is that AI agents now sit on both sides of the transaction. They do more of your customers’ researching and buying, and more of your media buying too. Fewer humans are clicking, so brands that measure success purely in clicks are measuring the wrong thing.
Here is what mattered, why it mattered, and what to do with it on Monday.
Board Analysis, The Digital Maze’s weekly marketing briefing, goes deeper on all of this every Friday. Subscribe here.
Adweek reports that OpenAI used its DevDay keynote to launch Dots, always-on assistants that work across connected apps, run several projects at once and learn from a user’s preferences. “Each Dot gets its own cloud computer and browser, and can plug into over 4,000 apps through OpenAI’s ecosystem.” Dots are available to ChatGPT Pro and Business Premium users, and OpenAI is also previewing specialist Dots for business, built on “lessons from early testing inside OpenAI across procurement, invoice processing, email marketing, customer support, and commercial contracting.” OpenAI’s own DevDay recap describes “more than 20 major announcements” and a ChatGPT ecosystem reaching “our collective 1.2B weekly users.”
This is the clearest sign yet that agents are becoming a normal part of work and shopping. Dots compete with Meta’s Muse and Google’s Gemini Spark, so all three of the platforms that matter most to marketers now offer an assistant that acts on someone’s behalf. Adweek notes that “Dots come with default rules for when to act alone and when to ask for approval, and users can add custom rules to allow, gate, or block specific actions.” For marketing teams, that makes governance a leadership decision rather than an IT setting. For customers, it means the first thing to visit your website may well be their agent.
Read more: OpenAI Takes On Meta’s Muse With Dots, an Always-On AI Agent, Adweek, 29 September 2026
Search Engine Journal reports that “Chartbeat reports a 40.2% year-over-year decline for Google Search across its publisher network from July 2025 to July 2026,” up from a 21.9% decline the year before. Google Discover’s decline grew “from 6.6% in the first year to 34.3% in the second,” and Google Search fell from about 9% of all pageviews in July 2024 to 5% by July 2026. The only growth came from channels publishers own or can influence: “Dark social, described in the playbook as links shared via texts, DMs, and private channels, grew from 7.1% to 11.3%.” SEJ notes the network leans towards news sites, and that Chartbeat sells analytics tools to publishers.
Two more stories this week explain why. A publishers’ court brief against OpenAI and Microsoft quotes ChatGPT head Nick Turley saying there was “no good reason to click” once ChatGPT gave an answer. SEJ is careful to add: “It’s one side’s argument, and the court hasn’t ruled on it.” But its conclusion is hard to argue with: “Being cited and being clicked are different measurements.” Advanced Web Ranking’s Q2 data, also covered by SEJ, shows the gap in Google’s own results: “On mobile, position one had a 35.95% CTR without an AI Overview and 8.51% with one.” When an answer is given for you, fewer humans click through to check it.
Read more: Google Search Decline Steepened To 40% YoY For Publishers, Search Engine Journal, 24 September 2026
Digiday reports that at its unBoxed conference, Amazon “renamed its unified platform Amazon Ads Agent,” and launched “Full-Funnel Campaigns, which combines sponsored ads, display, video and streaming TV within a single campaign.” A new format, DVA+, “consolidates Sponsored Display, Sponsored TV and programmatic display, video and audio buying.” The trade-off is control: “Advertisers approve creative, for example, but cannot select individual channels or specific audiences.” When Digiday asked whether buyers would see how winning prices were set, “Amazon did not commit to providing that level of auction transparency.”
This is the media-buying side of the same shift. Digiday sums it up as a model where “advertisers increasingly tell Amazon what they want to achieve, while its technology determines more of the media plan and execution.” That’s the Performance Max and Advantage+ approach, extended to one of the world’s biggest retail media platforms. Meanwhile, new ad surfaces are arriving with little data behind them. Search Engine Journal’s report on Graphite’s ChatGPT ads data puts UK click-through rate at 0.74%, and notes that OpenAI’s help centre says ChatGPT Ads currently “lack performance benchmarks.” Google is also widening Demand Gen, saying advertisers who add Gmail “on average see their image creatives achieve a 40% increase in conversions at the same ROI.”
Read more: At its flagship unBoxed conference, Amazon Ads collapses buying silos as AI takes a bigger role in media decisions, Digiday, 29 September 2026
Search Engine Journal reports that “Shopify expanded its WebMCP tools into checkout on Sept. 28,” letting browser agents update a shopper’s checkout and place the order once the buyer confirms. Shopify’s changelog says the tools “don’t expose a new API or require merchant configuration.” In Shopify’s own test, “WebMCP was successful in all 60 attempts, while browser automation succeeded in 56 out of 60,” and WebMCP’s cost per attempt “was 58% lower at OpenAI’s list prices.” There’s a catch for agents that don’t identify themselves: “Agents should sign their browser requests using Web Bot Auth, which Shopify relies on to recognize agents. Without it, their requests might be deprioritized or blocked by bot detection systems.”
Agent-ready infrastructure is spreading well beyond retail. SEJ also reports that Google plans to use the Universal Commerce Protocol’s new Lodging Booking draft, with Google stating: “Over the coming months, Google will adopt this standard to power its hotel booking feature on AI Mode.” Its FAQ adds that choosing the integration “does not influence how we rank your properties or rates in listings.” And Search Console’s new multimodal filter, which “covers searches that use an image, including Lens and Circle to Search,” shows how discovery is spreading beyond typed queries. As SEJ puts it, “Without queries, you’re left with pages.” Your site now has to be understandable to machines, cameras and agents, as well as people.
Read more: Shopify Extends WebMCP Into Checkout For Browser Agents, Search Engine Journal, 29 September 2026
The Media Leader reports on Kantar’s latest Media Reactions study: “A net 75% of marketers plan to increase investment in AI assistants in 2027, even as just one-third (32%) of consumers say they currently use AI assistants to research brands and products, with only 23% saying they trust their product or service recommendations.” The study is based on over 800 senior marketers and 23,000 consumers across 33 global markets. Kantar’s Gonca Bubani warned: “There’s a lot of money moving towards AI on the assumption people will use it the way the industry expects.” The same study found two-thirds of marketers plan to increase investment in creator content, with strategies moving towards “micro-” and “nano-creators”.
This is the human side of the week’s thesis. Agents may be doing more of the shopping, but people still decide which brands they trust enough to ask for. Marketing Dive’s profile of Thorne shows what that trust is worth. The supplement brand “surpassed $500 million in annual revenue and notched 63% growth in direct-to-consumer sales in 2025” before P&G’s $3.8bn deal, after research taught it that “messages about its scientific rigor were less important to consumers than it thought.” And at a Digiday event, brand-side marketers warned that scaling AI content is eroding standards, with one saying simply: “The quality control is so bad.” The brands that come through this will be the ones people recognise, trust and seek out by name.
Read more: Consumers don’t trust AI assistants for product recommendations. Brands are investing anyway, The Media Leader, 29 September 2026
Digiday reports on new IAB Europe research: “Fifty ad execs across 44 European markets were asked about it this spring and summer, and 58% of them expect agentic ad buying to hit operational use or scale within a year.” The gap by company size is striking. “Most of them (86%) who are at companies with 501 or more staff already run their most advanced agentic systems at a level where people and agents plan, delegate and execute work together. Drop down to companies with 500 staff or fewer and that number falls to less than half, 48%.” On risks, “security is top, according to the data. Privacy is second.”
If agents are going to buy media, the question for marketing leaders is what their own teams will do instead. Marketing Dive reports Gartner findings that “Nearly one-fifth of surveyed marketing leaders had eliminated some functional roles in their organization due to artificial intelligence,” based on a survey of 1,303 senior marketing leaders. Gartner’s advice is that “CMOs should give junior marketers exposure to higher-value work earlier.” The tools keep getting cheaper too. Anthropic says its new Claude Sonnet 5.5 “runs 30%+ faster, and costs up to 30% less for most work.” Cheaper, faster AI makes adoption easier. It doesn’t make the judgement calls any easier.
Read more: 58% of ad execs expect agentic buying to hit scale within a year, according to IAB Europe research, Digiday, 24 September 2026
When clicks get harder to win, the sales you own matter more. Nest Gifts is a boutique gift shop in Leicester selling over 20,000 products online. It was making substantial revenue through third-party sellers, but that came with a considerable amount of commission. The goal was simple: grow direct website sales and keep more of the margin.
Following a full audit covering technical SEO, paid media, content and UX, we launched Shopping campaigns with part of the budget set aside for testing, which revealed some surprising best-selling products. We optimised the shopping feed to improve and add missing attributes, and introduced Facebook ads and remarketing around seasonal peaks like Christmas and Halloween. On the SEO side, we worked closely with the client’s developers on technical fixes and used keyword mapping alongside NLP to optimise key pages, taking category pages that previously didn’t rank onto page one of Google.
Read the full Nest Gifts case study.
Every story this week points the same way. Agents are taking on more of the work on both sides of the market. OpenAI’s Dots act for individuals and businesses. Shopify’s checkout and Google’s hotel booking standard let agents complete purchases. Amazon’s Full-Funnel Campaigns let its AI decide more of your media plan. And the humans in the middle are clicking less, with Google Search referrals to publishers down 40.2% and mobile position one CTR dropping to 8.51% when an AI Overview appears.
What wins in that world isn’t more clicks. It’s being the brand an agent can read, a checkout it can complete, an ad account you can still audit, and a name people trust enough to ask for directly. Kantar’s data is a useful reality check: only 23% of consumers trust AI assistants’ recommendations today. So build for agents, but keep investing in the human trust that decides which brands agents get asked about in the first place.
We work with ambitious brands across SEO, PPC, web development, and content strategy. If this week’s news has raised questions about whether your site is ready for AI agents, how to measure search when fewer people click, or how to keep control as ad platforms automate, we’d like to talk.
Book a call with The Digital Maze.
Board Analysis is The Digital Maze’s weekly marketing briefing, published every Friday. It covers search, paid media, web development, brand, and AI for marketing managers, directors, and leaders who need to stay ahead without wading through every trade publication themselves.
Sign up to The Digital Maze Newsletter