Google Bets Big on Agents, Bing Bets Big on AI Reporting

📬 You like this podcast? You’ll love the newsletter.
Join the weekly 3-2-1 on AI search + marketing: subscribe

Google held Cloud Next '26 in Vegas this week and barely mentioned AI Overviews. They spent three days on agents instead — and that tells you everything about where the next phase of AI search is heading. Meanwhile, at SEO Week in New York, Microsoft Bing previewed the next chapter of their AI Performance dashboard, and it's further ahead than what Google currently gives us inside Search Console.

In this episode, Cassie breaks down:

•  What Cloud Next means for AI search visibility (hint: it's bigger than ChatGPT and Perplexity now)
• What Bing's new Citation Share, Grounding Query Intent, and GEO Recommendations mean for marketers
• Why now — not Q4 — is the moment to optimize for agentic commerce.

Plus quick hits on Google's back-button hijacking penalty, DeepSeek V4, and the agentic commerce trend line connecting Microsoft and OpenAI.

If you want help finding out where your brand currently shows up in AI engines, head to cassieclarkmarketing.com to start an AI search visibility audit.

Let’s connect:

LinkedIn → Cassie Clark | Fractional Content Strategist
Website → https://cassieclarkmarketing.com

Download Freshness, Structure, Authority: The Framework for AI Search Visibility:

Amazon

P.S. Is your brand losing its "Answer Authority"?

Most series A/B and enterprise brands are being "nudged" out of AI search results because of entity gaps and "stale" content. I am opening a limited number of specialized audit slots to help you reclaim your Share of Voice using the FSA Framework (Freshness, Structure, Authority).

Request your 7-Day AI Search Visibility Audit: https://cassieclarkmarketing.com/ai-search-visibility-audit/

Transcript

SPEAKER_00 1:13

Hey, welcome back to Found in AI. I'm Cassie Clark, a fractional content strategist, an AI search optimization expert, and the host of the show where we break down what's actually happening in AI Search, GEO and AEO, so you don't get left behind in this new wave of user search behavior. Today is Thursday, April 30th, 2026, and we have two stories that affect AI Search this week, or at least are worth knowing about if you're in marketing. Google held Cloud Next 26 in Vegas, and then Microsoft showed up at STO Week in New York and previewed what the next chapter of the Being AI Performance dashboard looks like. Let's get into it. Okay, let's start with the biggest event of the week. Google Cloud Next 26 kicked off on April 22nd in Las Vegas. If you have been wondering what Google's actual position is on the agentix shift, wonder no more. They've made it crystal clear. Soundar P Chai opened with a stat that said something like 75% of all new code at Google is now AI generated and then approved by engineers. That's up from something like 50% last fall. He also shared first-party data that says that they are now processing more than 16 billion tokens per minute through direct API calls from customers. That's up from 10 billion tokens per minute last quarter. 16 billion tokens per minute from customers. That's not an AI is the future kind of thing. That is very clearly AI is here, and we are scaling it. Now you're probably wondering, okay, well, why does all of this matter for AI search optimization? This is a podcast about that. Well, you were right, but Google also really hammered on the Gemini Enterprise Agent platform. I feel like that's a long mouthful of a name, but this is Google's end-to-end system for building, governing, and scaling AI engines inside the enterprise. Think of this as Google's answer to Microsoft Co-Pilot, but built around the assumption that agents are now doing the work, not just assisting with it. Now, here's what else got announced, and things that I think that matter, or at least we should pay attention to. First is that Gemini Enterprise itself, Google is positioning Gemini Enterprise as, and I quote, the front door to AI in the workplace. It connects to software like Salesforce, Slack, and Google Workspace with agents that can move across these tools and share context. This is something that I think we really need to think about maybe a little bit deeper, maybe more than I have time for right now, because it means that everything everywhere has shared data. Like everything is sharing data with all of your tools. Now, the second thing, long-running agents in the agent inbox. That was also something else that they talked about. Google launch agents that can work autonomously in secure sandboxes for hours or days with a central AI agent inbox where you monitor and approve your work. The interface for AI is moving from ask a question, get an answer, to delegate a task and review the output later. Some of us are already doing something similar with Claude or Codex, and now Google has just really entered the game. If your content isn't structured for an agent to retrieve, summarize, and cite, and I mean an agent that's running in the background, not a human typing a query, or optimizing for a behavior pattern that is already aging. So you really need to think about this in context of where things are headed. Three, eighth generation TPUs in the AI hypercomputer. Now, this is kind of technical, but there are two new chips: TPU 8T for training, TPU 8i for the interface, the 8i delivers 80% more performance per dollar, which is the chip that powers serving up answers. This is infrastructure, but it matters because it's how Google is keeping AI search responses fast and cheap as the volume scales. And four, and this one is relevant to marketers, workspace intelligence. Google launched a system that gives Gemini real-time context from Gmail, the calendar, chat, your Google Drive without you having to prompt it. So your Gemini, when you log into Workspace, now has memory and context across your work tools. That's a shift and it has some major implications. AI is starting to know more about how you work than you know about how it ranks you. And think about that in terms of your buyers. So, what's the takeaway from Google Cloud Next? Well, Google did not announce a single AI overviews update this week. Not one. They spent three days talking about agents. That tells you where they think the next frontier is. AI search isn't, and I say this, I say this and kind of giggle, it isn't just about citations in the answer panel. For most of us, that's what we're optimizing for at this very moment, and that makes sense, but it's gonna be about whether an agent running on behalf of a buyer can find your brand, evaluate it, and recommend it without that buyer ever opening a search engine. That is the world that the FSA framework was built for. Freshness for agents that are checking timestamps, structure for agents that extract before they reason, and authority for the agents that weigh entity signals more than they weigh domain authority. None of that is changing. It's just in context of who or what is doing the searching and the evaluating. What's changing is the retrieving, and it's becoming not a person that's doing the retrieving. It's gonna be a software running in someone else's cloud. If you've been treating AI search as just chat GPT and perplexity, consider this your nudge to expand the definition and maybe you're thinking about where this is headed. Now, before we move on, and because I've been talking about Agentic Commerce for the last couple episodes, like feel like it's been one of those brain worms that just keeps digging in. There's something really important that I want to call out here. I can't predict the future. I know I have a crystal ball that gets a little bit dusty, but I cannot see the future. I have no idea when agente commerce becomes the mainstream default for buyers. That could be six months, it could be 18 months. I'm not in that business, I'm not the one making the decision. But what I can tell you based on the signals from Google's Cloud Next this week, Microsoft's AI Max rollout, OpenAI's Workspace Agent, Yelp's AI Assistant, Perplexities Computer, all of the things coming from these organizations that are really setting the stage here. AgendaCommerce is where we're headed. These signals are not subtle anymore. They are very loud and very clear. Now, as a fractional content strategist, here's what I would recommend. The brands that optimize well to be the answer now, we still have a few months to test strategies. And I would think about doing that now instead of waiting, because waiting will take a Herculean effort once this becomes mainstream. You want to be the answer now so that when the plug on agenda commerce is pulled, you are already within that decision and recommendation layer. Now, the brands testing AEO and agent readable content today, they're gonna be the ones who already have data when the rest of the market is just figuring out the question. So optimizing for AI search now is gonna put you ahead when agenda commerce is a real thing. I say real thing like it's coming. Okay, let's move into that second story. This one's quick but kind of exciting if you've been following along with the Bing AI performance arc. SEO Week 2026 ran from April 27th through today, April 30th in New York City. It's iPool Ranks annual event, Mike King's Conference, and on April 27th, Krishna Madhaven from Microsoft Bing did a session and previewed what's coming next to the Bing Webmaster Tools. Now, if you remember my February episode, Bing launched the AI Performance Dashboard in their Webmaster Tools. This was exciting on many levels because this was the first major search engine to ship native reporting for AI citations. It's still live and it is still the closest thing that we have for Search Console for AI answers. What Krishna previewed at SEO Week was like the next step of where that's going. There's a couple new features that he demoed, and they include citation share, which measures the distribution of how often your site is cited relative to competitors on a given topic. This is essentially sure voice, but for AI answers. We have the grounding query intent, which shows the purpose behind the grounding query, not just the query text. So instead of just seeing something like what is the FSA framework, you'd see whether that query was informational, commercial, or navigational in intent. There's also semantic topic label, which is being tagging your site's content with the core topics it identifies you as authoritative for. That's entity authority made visible, that's gonna be super helpful. And then we have GEO focus recommendations. So actual prescriptive recommendations on how to improve AI visibility based on Bing's data. Now, if you remember, Bing supports a lot of these AI engines. So this is gonna be very, very helpful. There are a few things to keep in mind though. These features were not or demoed, they were not shipped, so we don't have this in the webmaster tools yet, and I don't know when this will be live because they didn't give a launch date. But this is the direction they're headed, and I'd argue Bing is now further ahead than Google on native AI search reporting. Google still has not given us a way to filter out AI mode out of Search Console. Search Console has been a hot mess with that impressions bug. Bing is shipping a dedicated dashboard for AI citations and previewing topic labels, intent classifications, and GEO recommendations. If you have been ignoring Bing, I say this gently, if you have been ignoring them, because it's only 10% of search market share, reconsider that. Not because Bing's market share is suddenly gonna balloon, it's probably not, but because what Bing is showing you in their tools is what Google is probably tracking internally and just not surfacing yet. Bing's dashboards are kind of a forward-looking window into metrics that will eventually matter everywhere. So use what we have right now while we're waiting for everything else to kind of solidify. Okay, a few more things worth noting, but not like enough to have a whole segment on. There's three things. One, Google wrote out a back button hijacking penalty. I mentioned this because if you inherited an old site or you're running aggressive landing pages from paid agency, you have until June 15th, 2026 to remove that. After that, you're gonna get hit with manual or algorithmic penalties. Um, again, most of you aren't doing that, but just double check to make sure you're on the safe side. Two, deep seek released before in preview. Now, China's Deep Seek launched updated Pro and Flash variants of their B4 model with bigger context windows and better identic performance. They are claiming competitive performance with the US frontier models. Worth tracking even if you're not using Deep Seap directly, because where this matters for AI search is the grounding behavior. Different models retrieve and cite differently, and Deep Seek getting better means more chat GPT class engines for buyers to use. And three, Microsoft and OpenAI this week are both pushing on agenda commerce. I giggle because here it is again. Microsoft launched AI Max for search updates inside Copilot and Bing earlier this month, and OpenAI rolled out workspace agents and chat GPT for business that can autonomously complete tasks across Slack, Gmail, and other tools. Now the trend line is clear here. AI search is bleeding into AI commerce. The buyer journey is collapsing, and discovery, evaluation, and transaction can now happen in a single conversation, which is why I keep talking about this so much. So here's what this week tells me when I zoom out and look at it. Google spent the week talking about agents, Microsoft spent the week previewing the better AI search reporting, and then we have Google Search Console, which I didn't really hit on. It's still in the middle of a hot mess. Now, when all of this is connected, measurement and infrastructure are catching up to behavior. People have been searching with AI for almost two years now. The tools to track that behavior are finally starting to catch up and arrive. And the infrastructure to act on that behavior, identical, is also here. This is the maturity moment that I talked about back in February. It is just someone is pushing the pedal on the gas pedal. If you've been treating AI search visibility as a next quarter problem, we are already behind. The brands that are going to win the next 12 months are the ones who treated this as a now problem six months ago, and the ones who treat it as a now problem today. So even if you didn't start six months ago, you still have some time to catch up where it's going. And I would encourage you to start doing that as soon as possible. If you want help figuring out where your brand actually shows up in AI engines and what's keeping you out of the answers that your buyers are asking, I run AI surge visibility audits for B2B brands. You can find more information on that at Cassie Clark Marketing.com or in the show notes. I'm Cassie Clark. That's it for the week. I will see you on Tuesday. Until then, stay visible.