Q3 Digital Marketing Report: Google Goes AI-First, LinkedIn Rewards Real Voices, and Agents Go Mainstream

Let's Get Digital walks you through the most important developments from the past three months. In Q3, AI moved from the edge of the marketing conversation to the center of it. Google made AI Mode the front door to search. ChatGPT ads crossed a billion-dollar run rate, with financial services brands among the fastest movers. And LinkedIn made it clear that your people, not your logo, drive both reach and AI visibility. Below, we break down what changed across AEO, social, go-to-market, and marketing operations, what it means for B2B brands, and where we think you should focus heading into 2027 planning.

Let's get into it.

🤖 AEO (Answer Engine Optimization)

In Q3, Google shifted from viewing AI as a search feature to making it the primary access point.

• Google Is Funneling Searchers Into AI Mode: The Google homepage now shows an AI Mode button where Search used to be, and AI Overviews now push users into AI Mode instead of out to web pages. With more than 1 billion monthly users, every follow-up question in AI Mode is another chance for your brand to be cited or replaced by a competitor.

• Search Console Finally Shows AI Visibility: Google's Generative AI performance report rolled out worldwide on August 31, isolating how often your pages appear in AI Overviews and AI Mode for the first time. It counts impressions, not clicks, so treat it as a visibility baseline rather than a traffic report.

• Cloudflare Flipped the Default On AI Crawlers: Since September 15, Cloudflare blocks AI Training and Agent bots by default on ad-supported pages for new and free-tier sites. Because Googlebot crawls for both search and AI training, blocking Training can block Googlebot too, a real risk now that daily AI agent requests have grown more than 1,700% in a year.

Top Takeaways for Marketers:

• Pull your Search Console Generative AI report and set a Q3 baseline now. We think page-level impressions are the clearest early read on which content AI already trusts.
• Ask your web team to audit CDN and bot settings this month. One default can make your site invisible to AI agents, and sometimes to Google.
• Start writing for the follow-up question. AI Mode is a conversation, so content that anticipates the second and third question earns more citations.

💬 Social Media

LinkedIn is rewarding real voices and quietly demoting anything that sounds automated.

• Comments Are Now Ranked Per Viewer: LinkedIn now ranks comments by relevance to each user, considering professional interests, connections, and engagement history. As a result, users are spending more time in the comments than ever, up 18% year over year. And with about 30% of comments from April to June being fully AI-generated, the focus has shifted to usefulness rather than just being the first to comment.

• Your People Are Your AI Visibility Strategy: LinkedIn's August playbook highlights Meltwater research showing that 75% of AI citations come from individual member profiles, not company pages. LinkedIn ranks as the second most-cited source for AI answers, after YouTube, and is also second in Financial Services, indicating that executive posts shape AI’s depiction of your company.

• New Paid Levers For Long Sales Cycles: LinkedIn's September release adds buyer group targeting, a better fit for decisions made by committee. It also adds a 180-day attribution window, so measurement finally matches a B2B buying cycle that can stretch most of a year.

• Event Content Gets A Second Life: LinkedIn can now turn recorded events into highlight clips and chapters that viewers can jump between. Every webinar now includes a built-in way to create short-form content without additional production.

Top Takeaways for Marketers:

• Build an executive and SME posting program. It's become the most efficient lever for both LinkedIn reach and AI citations.
• Drop engagement pods and "comment first" tactics. Coach your team to leave fewer, sharper comments that add real expertise.
• Test buyer group targeting, and stretch attribution windows to match your actual sales cycle.
• Cut every webinar into highlight clips before the recording goes stale.

💸 Go To Market

The money is following AI answers, and regulated categories like financial services are moving faster than anyone expected.

• Google Is Testing Ads Inside AI Answers: In mid-September, Google was spotted testing sponsored text links woven directly into AI Mode responses with a "Sponsored" label. If it rolls out, paid search creative will need to read like a helpful answer, not a tagline.

• ChatGPT Ads Hit $1B: OpenAI says ChatGPT Ads reached a $1 billion annualized run rate in under 200 days, and Sensor Tower found financial services grew from 2% to 13% of U.S. spend between April and August, with four of the top ten spenders. Keep it in perspective: agencies say they aren't seeing performance yet, and actual booked revenue is closer to $330 million.

• ChatGPT Is Testing Sponsored Agents: OpenAI began testing Sponsored Agents with select U.S. advertisers in mid-September. It's an early sign that ads inside AI will move beyond links toward brands acting on a user's behalf.

• Safari 27 Cut The Trade Desk Off On iPhones: Apple's newest Safari blocks the domain The Trade Desk uses to serve ads, along with identity domains tied to UID2, ID5, LiveRamp, and Permutive, while Google's bids reportedly went through. With Safari at roughly 15% of global browser traffic and no fix date from Apple, this is a real reach gap for programmatic buyers.

Top Takeaways for Marketers:

• Write answer-first creative now. Sponsored links inside AI answers will reward specificity over slogans.
• If you're in financial services, test ChatGPT ads this quarter with tight measurement in place. Your category peers already are.
• Don't mistake run rate for results. Set up incrementality tests before scaling any new AI ad channel.
• Ask your DSP for an iOS Safari impact readout, and check whether your identity partner is on Apple's blocked list.

🚧 MarOps

The Q3 lesson was blunt: agents are only as smart as the data underneath them.

• HubSpot Bet Its Platform On "Growth Context." At UNBOUND, the renamed INBOUND conference, HubSpot built every announcement around one idea: AI agents are only as useful as the customer data they work from. Every new agent reads a shared context layer before it acts, so gaps in your CRM become gaps in your agents.

• Agentic Adoption Is Steep, And So Is The Failure Rate: Gartner found only 17% of organizations have deployed AI agents, but 42% expect to within the next 12 months. At the same time, Gartner expects more than 40% of agentic projects to be canceled by the end of 2027 due to rising costs, unclear value, or weak risk controls.

• Marketing Needs A Seat At The Data Table: Gartner warns that many CMOs know an enterprise AI data fabric is coming but have no plan to join it. Its advice is to productize customer data and set clear data contracts, so marketing stays in the loop on how that data feeds AI decisions.

Top Takeaways for Marketers:

• Run a data completeness audit before turning on any new agent. Clean CRM data is the difference between a useful agent and an expensive mistake.
• Pick one workflow and define success and shutdown criteria up front. It's the simplest way to avoid joining Gartner's 40%.
• Make AI crawler and agent access a cross-functional decision across marketing, IT, and legal, not a setting someone toggles once.

🛠️ Tool Time

New tools generating attention as we head into Q4:

• Search Console Generative AI Report: Google's free, first-party view of how often your pages appear in AI Overviews and AI Mode, broken down by page, country, and device. Start here before paying for any third-party AI visibility tool.

• ChatGPT Carousel Ads: OpenAI's new multi-product ad format shows several offerings in a single placement. Sensor Tower estimates it already accounted for about 23% of U.S. desktop ChatGPT ads in late August.

• HubSpot Agent Builder: HubSpot's tool for building custom agents, such as scoring leads against your ICP or finding patterns in closed-lost deals. The new external triggers make it worth a second look if you tested it earlier.

• Cloudflare AI Bot Policies and BotBase: Cloudflare replaced its old on/off switch with three bot categories (Search, Training, and Agent) plus BotBase, a database of verified crawlers. If your site runs on Cloudflare, this is where you control your AI visibility.

DaaS a wrap 👏

The Q3 signal is clear: AI is no longer a channel you test on the side. It now shapes how buyers find you, which sources they trust, and where your ad dollars compete. The brands pulling ahead aren't chasing every new feature. They're getting the fundamentals right with clean data, credible voices, and content built to be cited, and then measuring what actually moves pipeline.

At The Ricciardi Group, we help B2B brands build the strategies and systems to stay ahead, from AEO audits to agentic workflow design to LinkedIn programs that actually accelerate pipeline.

If you're ready to pressure-test your approach or dig deeper into any of the above, our strategists are ready to collaborate.

‍