Google Search Central Live Canada 2026: My Takeaways

On April 21, 2026, I spent half of the day in a room with the Google team at the first-ever Search Central Live event in Canada.

I’ve been doing SEO since 2019, and I’ve been to plenty of marketing events, but this one felt different. You don’t usually get to sit that close to the people who actually build Google Search… and ask them questions directly.

I didn’t go in just to listen. I had an agenda.

I came with a short list of questions, and all of them were about AI SEO. It’s not surprising since this is the most confusing part of our work right now, and honestly, the one I’m trying to figure out myself.

Here’s what I wanted to understand:

Is there actually anything we can do to influence how our content gets picked for AI Overviews? Or once the content is “good enough,” is it out of our hands?
Are we going to get better visibility into AI search visibility inside Google Search Console (like a separate report for AI Overviews and AI Mode)?
And with everyone plugging GSC data into LLMs… is Google planning to support that workflow officially and launch an MCP connector?

I didn’t get perfectly clear answers to all of these.

But I picked up enough from the talks and Q&A to start connecting the dots and form an opinion on where things are heading.

Here’s what I took away from the event, and what I think every site owner should understand right now.

How Search Actually Works

Quick takeaways for site owners:

If your page isn’t indexed, nothing else matters. No rankings, no traffic, no AI mentions. You’re not even in the game yet.
“Crawled, not indexed” is often not a technical problem. It is a quality signal from Google telling you the content isn’t strong enough.
Volume does not equal visibility. Publishing more content (especially with AI) won’t fix it. In fact, it often makes things worse in 2026.

What I liked about Martin Splitt’s talk is how clearly he explained what’s happening behind the scenes.

Google basically has two separate processes:

It’s constantly crawling and indexing pages in the background
Then, when someone searches, it pulls results from that index

If your page never makes it into the index, it will never show up in search.

And I think that’s where a lot of site owners get stuck. They focus on rankings, keywords, optimization… when the real problem is that Google never found their content worth indexing in the first place.

One thing that really stuck with me was how direct he was about content quality. He basically said: even if Google can fully crawl and understand your page, it still won’t index it if the content isn’t good enough.

That’s the part a lot of people don’t want to hear.

I keep seeing SEOs dive into technical fixes, checking robots.txt, resubmitting URLs, tweaking settings, when the real issue is much simpler: the content just isn’t good enough.

And honestly, that makes sense. It’s easier than ever to publish content now. AI lowered the barrier even more. So, Google has to be stricter about what actually gets indexed.

Therefore, if your pages are stuck on “Crawled, not indexed,” start by asking a harder question: Is this page actually worth showing in search results? Because most of the time, that’s where the problem is.

Here’s another thing Martin pointed out that many site owners overlook:

Google doesn’t just decide once whether to index your page; it keeps re-evaluating it.

If your page is basically a duplicate (even with a different URL), it might be crawled but never indexed.
If it does get indexed, it can still drop out later if no one clicks it or better content shows up.
Rankings aren’t stable either, and Google is constantly testing new pages to see what users prefer.

He also shared a stat that really puts things into perspective: in 2023 alone, Google ran 700,000+ quality tests. Most of these changes are invisible to us, but they’re happening all the time.

Your content isn’t competing to get visibility once. It’s actually competing in a system that’s constantly experimenting and changing.

AI in Google Search

Quick takeaways for site owners:

SEO and “AI SEO” are not separate things. AI search is built on traditional fundamentals, not a completely new system.
The new thing is non-commodity content. “Top 7 tips for first-time homebuyers” is not going to cut it anymore. Original perspective, real experience, and depth are what AI systems surface.
Traffic from AI Overviews and AI Mode is more engaged. Users arrive with more context, so they spend more time on your site. Stop measuring just raw traffic and start measuring value per visit.
You don’t need to chase every AI query variation. Google already understands context.
Third-party AEO and GEO tools do not have access to Google’s internal metrics. Their scores and recommendations are guesses. Use them if helpful, but think critically.

The following talk was from Danny Sullivan, Director of Google Search, on “AI in Google Search.” This was the session everyone was waiting for. We all wanted clarity on where Google Search is heading and what we, as site owners, can actually do to improve visibility in AI-driven results.

Danny actually answered the most burning question that many SEOs ask on LinkedIn daily: Is SEO and AI SEO different?

I’m glad he addressed it because it cuts through a lot of the noise.

His answer also aligns with what Neil Patel shared with me when I asked him the same question during the Full Circle Conference in Toronto on April 8th. When industry leaders are saying the same thing, it’s worth paying attention.

Danny broke down how AI search actually works in a way that makes things much clearer:

AI models are trained on general knowledge, which comes from patterns across massive amounts of content. But when Google generates AI responses, it doesn’t rely on that alone. It pulls in fresh data from search results.

A single query gets expanded into many related searches, hundreds of pages are reviewed, and Google builds an answer based on patterns across those results.

This changes how we should think about search. It’s no longer just about ranking for one keyword. It’s about being part of a wider pool of content that Google sees as useful, consistent, and reliable.

If your content is generic, like “10 tips” that already exist everywhere, AI can easily replicate it without needing to send traffic to your site. Danny referred to this as “commodity content.” It’s not useless, but it’s not what will help you stand out anymore.

What does stand out is “non-commodity content.” That means content based on real experience, original insights, or depth that can’t be easily summarized. He gave examples of sites that test products hands-on or publish detailed, experience-led breakdowns. That kind of content is harder to replace and more likely to be referenced in AI systems.

He also pushed back on many common assumptions:

You don’t need to rewrite content into an “AI-friendly” structure.
You don’t need to break everything into tiny fragments for machines to understand.
You don’t need perfect HTML or overly engineered structure.

Honestly, this was both reassuring and frustrating to hear. Reassuring because the work I have been doing on behindrankings.com for years, topical authority, first-person case studies, and real data, is exactly the kind of non-commodity content Danny said AI systems reward. Frustrating because it means a lot of the “new” AI SEO advice floating around right now is just repackaged fundamentals with a bigger consulting invoice attached.

My biggest takeaway from this talk was that nothing fundamentally changed, but the bar for quality definitely did.

Telling Stories with Google Trends

Quick takeaways for site owners:

Keywords show intent, but trends reveal human behavior and context.
Google Trends is a real-time signal of curiosity, not just a keyword tool.
Strong content comes from filling “information gaps,” not repeating what already exists.
AI tools in Search can help map questions, but your unique perspective is what makes content stand out.

Annanya Raghavan, Trends Analyst at Google, had a simple core message: keywords show what people want in a moment, but trends show who people are and what they care about over time.

She described Google Trends as a “cultural compass,” because it reflects real-time human curiosity across billions of searches. It’s immediate, large-scale, and surprisingly honest since people search without filters or performance bias.

Instead of treating Trends as a list of rising keywords, she encouraged us to use it to understand timing, emotion, and context.

Her key point was that the best content doesn’t just react to demand, it fills narrative gaps. Using Trends plus AI tools in Search, creators can see what questions are being asked, what’s already covered, and where the missing perspective is.

The process she outlined is following:

Identify breakout interest
Validate relevance
Understand what’s already ranking
Add your unique angle
Publish content that actually adds something new to the conversation

In her words, trends give you the “what,” but storytelling and context give you the “why.”

What's New in Google Search Console

Quick takeaways for site owners:

Query Groups is live in Search Console. It uses AI to cluster thousands of misspellings and variations into themed groups, which matters a lot as queries keep getting longer in AI Mode.
AI-powered filters let you describe what you want in plain language and automatically generate the right Search Console views.
The Trends API is now in closed alpha. It offers consistently scaled search interest, five years of history, and custom time resolution, which the web interface does not give you.

Daniel Waisberg is a Search Advocate at Google and also part of the Search Console engineering team.

He started his talk by saying that Google Trends, Search Console, and Analytics all show you search performance, but from different angles. The challenge has always been turning those separate views into something meaningful.

That’s where the new updates come in.

One of the biggest changes is Query Groups in Search Console. Instead of looking at thousands of individual search terms, you now see them clustered into themes. This helps solve a long-standing problem where the same intent is split across hundreds of variations, misspellings, or languages. AI is used to group these together based on meaning, not just keywords.

He also highlighted AI-powered configuration inside Search Console. Instead of manually setting filters, you can now describe what you want in plain language, and Google will build the report for you using Gemini.

On the Trends side, the new API introduces more consistent data scaling and better control over resolution, making it easier to compare search interest over time without rebuilding datasets every week.

I believe their overall direction is to have less manual digging, more structured insights, and faster ways to understand what search data is actually telling you.

Structured Data, Quality, and AI

Quick takeaways for site owners:

Google uses structured data for accuracy, cost efficiency, and focus when interpreting pages.
AI models may hallucinate properties and nest data incorrectly, especially for complex schemas like shopping. If you care about accuracy, do not rely on Google’s models to infer your structured data from page content.

Ryan Levering from the Google Engineering team talked about whether structured data is still needed in an AI-driven search world.

Ryan cleared up a myth I hear constantly on LinkedIn that structured data is becoming irrelevant because AI can just read your page. He explained why that is wrong. Models are expensive to run on every page. They can also lose focus when there is a lot of content on a page, and they cannot see data that is not visible on the rendered HTML file.

AI can extract information from pages, but it is not consistently precise. It can miss details, misread structure, or even pull incorrect values, especially in complex areas like shopping data. Structured data removes that guesswork by explicitly telling Google what each element means.

Running large AI models on every page is expensive, so structured data remains a faster and more scalable way to feed accurate information into Google’s systems.

Q&A with the Google Team

The closing part of the event was a Q&A session where the Google team answered questions submitted by participants throughout the day. I made quick notes on every response, which you’ll find below.

The first question was: “Do you run any audits to detect AI-generated content?” The answer was clear—no.

Google does not label or evaluate content based on whether AI was used. Instead, they focus on whether content is helpful, original, and valuable to users. The emphasis is on quality, not production method.

The second question focused on “Tips for rendering JavaScript for Google.” The response reinforced that Google primarily processes rendered HTML, not frameworks or code structures.

If content is visible in rendered HTML via Search Console’s URL inspection tool, it is generally fine for indexing. The key issue is not JavaScript itself, but whether it prevents content from being accessible after rendering.

We also heard a question about “How important is publishing consistency?” The answer was publish when it makes sense for your audience, not because of an assumed Google preference. Consistency should serve users, not algorithms.

One of the more technical questions was about hreflang tags. The answer: they are important only when there is a clear multilingual or regional use case. If the content is identical across versions without real localization needs, hreflang is not necessary.

There was also a question on “Does blocking Google-Extended hurt your AI search visibility?” The short answer was — no. Google-Extended only blocks Gemini training. It does not affect your ability to appear in AI Overviews or AI Mode.

During the event, I also had a chance to talk to the Google team, so I asked them a few questions.

First, I asked whether Google plans to launch a Search Console MCP connector for LLMs. The response was “maybe,” which tells me it is probably being explored.

I also asked about AI search visibility reporting similar to Bing’s. The team explained that they are actively exploring options, but reporting is complex because AI responses are dynamic and personalized for each user. This makes standard metrics difficult to define. I also suggested adding reporting on cited pages in AI Overviews, which seemed like a valuable direction for future development.

Overall, I got an impression after this event that Google is still evolving in how it communicates the AI search updates to the public.

While many of the AI features are still in exploration mode, there’s a clear signal that feedback from the SEO community is actively shaping what comes next. It will be interesting to see how these discussions translate into future Search Console updates and AI visibility tools.