AI-generated search results are reshaping how users find information, and the data now confirms what our team has observed across client campaigns for months: traditional organic traffic is declining, citation sources inside AI tools are volatile, and brands that fail to optimise for generative engine visibility risk losing ground fast. Here’s what we’re acting on right now and what every business owner needs to understand.

Key Takeaways

• AI Overviews are expanding across Google SERPs, and tracking them is now a core part of any serious SEO workflow.
• Reddit’s share of ChatGPT citations collapsed from 3.8% to 0.5% in just four days, exposing the instability of AI source selection.
• Prompt research is emerging as a new discipline — understanding which queries trigger AI answers and where your brand appears (or doesn’t) is essential.
• Google organic traffic is measurably declining for many sites, forcing a shift toward diversified, owned-channel strategies.
• Content optimisation now means writing for readers, traditional search engines, and AI systems simultaneously.

Tracking AI Overviews Is No Longer Optional

Google’s AI Overviews now appear for a growing percentage of queries, pulling content directly into the SERP and often satisfying user intent before a click ever happens. Our team has started monitoring which client keywords trigger these overviews and whether our pages are cited within them. We’ve built this into our standard reporting cycles, a process supported by Semrush’s AI Overview tracking capabilities. If you’re not measuring this, you’re flying blind on roughly a third of your keyword estate.

Reddit’s ChatGPT Citation Crash Reveals Source Volatility

Promptwatch data shows Reddit’s citation share inside ChatGPT responses dropped 86% — from 3.8% to 0.5% — over just four days. OpenAI says nothing changed on their end. Whatever the cause, this sharp decline in Reddit’s ChatGPT citations sends a clear message: relying on any single platform for AI visibility is a losing bet.

For our clients, we treat this as a diversification signal. We ensure brand mentions, expert content, and structured data exist across multiple authoritative domains — not just one community forum — so that when AI models reshuffle their source preferences, our clients’ visibility holds steady.

Prompt Research: The New Keyword Research

Traditional keyword research tells you what people type into Google. Prompt research tells you what people ask AI tools — and whether your brand shows up in the answer. These are fundamentally different questions. We’ve started running prompt audits for key clients, mapping out:

• Which conversational queries mention competitor brands but not ours.
• Which product categories trigger AI recommendations.
• Where content gaps exist that AI models currently fill with competitor data.

This repeatable process is something we now consider foundational, aligned with the prompt research methodology outlined by Semrush. It identifies blind spots that traditional rank tracking simply cannot.

Google Traffic Is Declining — Diversify or Stagnate

The numbers are in. Multiple data sets confirm that Google organic click-through rates are falling as AI answers absorb queries that previously drove site visits. Our response across client accounts has been straightforward: build owned audiences. Email lists, first-party data capture, community channels, and conversion-optimised landing pages all reduce dependency on a single traffic source. We’ve outlined these steps in line with practical guidance on overcoming AI-related traffic loss. Waiting for traffic to recover on its own is not a strategy.

Content Must Now Serve Three Masters

Every page we publish or revise for clients is now evaluated against three criteria: does it satisfy the human reader, does it rank in traditional search, and does it stand a chance of being cited by AI systems? This means tighter structure, clearer entity references, stronger E-E-A-T signals, and concise, quotable answers embedded within longer content. We’ve adopted a comprehensive set of content optimisation tactics designed to boost both SEO and AI visibility as part of our production standards. The days of writing purely for Google’s crawler are behind us.

Generative engine optimisation is not a future concern — it is a present-day operational requirement. Throughout 2025 and beyond, our team will continue refining these workflows as AI Search behaviour evolves. Businesses that act now secure a structural advantage; those that wait will be chasing visibility they’ve already lost.

Frequently Asked Questions

What is generative engine optimisation (GEO)?

GEO is the practice of optimising your content so it gets cited and surfaced by AI-powered search tools like Google’s AI Overviews and ChatGPT. It extends traditional SEO by focusing on structured, authoritative content that AI models prefer to reference in their generated answers.

How do web designers and SEO teams track AI Overviews?

Specialist tools now monitor which keywords trigger AI Overviews and whether specific pages appear within them. Our team integrates this tracking into monthly reporting so clients can see exactly where AI results are replacing traditional blue links.

Why is Google organic traffic declining in 2025?

AI-generated answers at the top of search results satisfy many queries without requiring users to click through to a website. This zero-click behaviour reduces organic traffic even when rankings remain stable.

What is prompt research and why does it matter for SEO?

Prompt research analyses the conversational queries people ask AI tools and identifies where a brand does or doesn’t appear in the responses. It fills a gap that traditional keyword research cannot address, revealing competitive blind spots in AI-generated recommendations.

How do businesses future-proof against AI-related traffic loss?

The most effective approach is diversifying traffic sources by building email lists, investing in first-party data, and strengthening owned channels. Simultaneously optimising content for both search engines and AI citation increases the chances of maintaining visibility regardless of how results are displayed.