For Marketing & SEO Agencies

It's go-time for us marketers and SEO folks. Google I/O 2025 essentially handed us a new playbook to learn. We need to pivot strategies to keep clients visible in this AI-centric landscape. Here's how to do it:

Focus on Answer Engine Optimization (AEO): Traditional SEO (title tags, backlinks) isn't enough when an AI is synthesizing answers. We need to optimize content so that AI chooses to use it. That means answer-first content – lead with a concise answer or definition in your pages (within the first 50-75 words), then elaborate. Use question-based subheadings (think FAQ style) so the AI can easily grab relevant chunks. The goal is to have the clearest, most authoritative answer to common questions in your domain, formatted in a way that's easy for AI to snippet.

Structured data is your best friend: We've preached schema markup for years, but now it's mission-critical. Adding FAQPage, HowTo, Product, and Speakable schema helps Google understand and trust your content. For instance, marking up a how-to guide could make your steps directly quotable by an AI answering a "How do I…?" query. Schema makes your content machine-friendly, which is exactly what we need for AI Mode.

Track new KPIs: We're used to metrics like clicks, impressions, and conversion rates. Add to that list: Answer Share, Conversation Join-Rate, and Citation Depth. In simple terms:

• Answer Share – the percentage of AI-generated answers in your category that reference your brand/content. Forward-thinking agencies are already benchmarking this, running conversational audits to see how often clients get cited in AI answers. 
• Conversation Join-Rate – how often users who see your brand mentioned in an AI answer follow up (either by clicking through to your site or by prompting the AI for more info on that topic). This one's a bit abstract, but essentially it gauges engagement from AI results. If lots of users click the citation or ask the AI "Tell me more about [Your Company]," that's a high join-rate – a good sign of interest generated.
• Citation Depth – how deep or how many times your content is used in an AI conversation. Did the AI just quote you once and move on, or is it pulling multiple facts from your site across follow-up queries? Deeper citation implies your content has broad, valuable info. We'll measure this by analyzing multi-turn queries. If your blog post on electric cars gets cited for a question about range, and then again when the user asks a follow-up about battery lifespan, that's great citation depth.

These metrics aren't neatly available in Google Search Console (yet), but I won't be surprised if Google gives us something soon. In the meantime, use manual testing and emerging third-party tools (some are already cropping up to monitor AI results) to track this.

Multimodal content strategy: Google's search is becoming multimodal (text, images, video, audio). Agencies need to deliver content in various formats. That means advising clients to have not just blog posts, but infographics, short videos, maybe even podcasts – because AI Mode might show an image carousel or a snippet from a video (imagine an AI answer quoting a how-to step from your YouTube video, or showing a product image from your site in a shopping answer). Pushing for "content repurposing 2.0." is key. Every key piece of content should have a visual and audio counterpart. For example, if a SaaS client publishes a report, we might also create a quick infographic summary and a short Slide AI video. This increases the chances of being picked up in those richer AI results.

Continuous education and experimentation: I think I'll always will be in permanent beta mode with these new features. I am already hands-on with some of Gemini's tools (some of us may or may not have spent all night testing Flow and Deep Research mode!). The content strategies of 2025 will incorporate prompt design: essentially, predicting what a user might ask the AI and ensuring our content is structured to be the answer.

We're not just Search Engine optimizers anymore, we're Answer Engine optimizers. The core principle remains: understand what users need and provide it. But now the "customer" for our content is as much the AI intermediary as the end-user. And agencies that get it right will thrive.