Trends & News

Stop Losing Clicks to AI: Learn How to Start Winning Citations

8 min read

AI search is moving faster than most brands can react. GPT-5 now powers ChatGPT and Copilot. Google’s AI Mode is steering users toward guided follow-ups.

Truth Social is feeding Perplexity answers. Each shift changes who gets cited and clicked. Move now to secure your place in the results before these new patterns lock in and competitors take the lead.

Here's what changed for citations, clicks, and how your pages get chosen.

1. Quick Signals That Shift GEO This Week

This week has packed real shifts in how answers are chosen across Google, Copilot, and Perplexity. If you need the short version to brief stakeholders fast, let's take a close look:

Also read: Ranking Rules Changed (Again): Is Your Site Structured to Win?

2. Deeper Stories Where AI Citations Change Hands

Some moves deserve more than a headline. Model swaps and new paths are changing the sources that are quoted. To plan content that wins across engines, let's take a close look:

GPT-5 Becomes ChatGPT's default

OpenAI made GPT-5 the core model for ChatGPT, which raises the bar for reasoning and code answers users see by default. Brands that already feed clean, cited content into model answers will gain.

Sites with vague claims or stale data will fall from AI results. Expect answer engines to favor sources with clear provenance and fresh updates. Swap thin FAQs for specific Q&A blocks and dated claims for verifiable stats. (OpenAI, TechCrunch)

Copilot's Smart Mode Changes Model Routing

Microsoft turned on GPT-5 across Copilot with Smart mode. The assistant now routes tasks between faster and deeper models. That improves answer quality and changes the sources quoted.

Enterprise teams should test prompts that reference SharePoint and Outlook content. Check whether internal pages appear in summaries and fix outdated text. The upgrade also widens developer access through Azure AI Foundry. Treat Copilot as a new distribution layer for citations. (Microsoft)

Google Adds "People also search for" Inside AI Mode

PASFs now appear under the AI Mode link cards. That nudges users into guided follow-ups without returning to classic results. It compresses the journey and concentrates visibility on entities with strong profiles.

Add concise follow-up questions to your pages to match these branches. Use schema to declare entities and relationships, not just keywords. Track which PASFs trigger on your head terms and mirror that language in headings. (Search Engine Roundtable)

Truth Social Taps Perplexity for AI search

Truth Social launched an AI search that runs on Perplexity. That creates another surface where brand citations can appear. Source controls may limit which domains feed the answers.

Brands should check whether profiles, press pages, and documentation are current. Short, factual statements with dates and sources travel best in these summaries. Monitor for outdated claims and push corrections through owned channels. (TechCrunch)

Bing Webmaster Tools Extends Performance History

Bing's Performance report now offers 24 months of data plus device and country filters. That makes GEO testing more reliable.

You can compare pre-AIO and post-AIO visibility without stitching exports. Use the longer window to isolate model-driven shifts from seasonal noise.

Chart brand query clicks and non-brand impressions by country. Pair that with your LLM-mention tracker to see if content changes moved both lines. (Search Engine Roundtable)

Also read: When Answer Boxes Rule, How Will You Stay Seen?

3. Fresh Stats That Back Your Next Budget Call

Budgets move when numbers speak plainly. These figures show where clicks shift, how answers rank, and which surfaces matter most. Use them in the next deck; let's take a close look:

  • Enterprise GEO CAC averages $716, with a 5.2-month payback window. (First Page Sage)
  • 45% of Gen Z now search via TikTok, bypassing Google altogether for certain queries. (Ahrefs)
  • GEO leads score 9.2% higher on lead quality metrics vs SEO leads. (First Page Sage)
  • 74% of new indexed content is AI-generated, diluting SERP uniqueness. (Ahrefs)
  • In-house-only GEO succeeds only 52% of the time, vs 87% success with expert agency help. (First Page Sage)
  • Pages with AI Overviews see just 8% CTR, down from 15% for traditional SERP listings. (Ahrefs)
  • GEO deals close 30% faster than SEO-driven ones, at 89 days vs 127. (First Page Sage)
  • 26% of searches with AI summaries end the session, showing high info satisfaction but low downstream traffic. (Ahrefs)
  • Higher Ed faces the highest GEO CAC at $1,014, reflecting complex buyer journeys. (First Page Sage)
  • Google visual search is up 65% YoY, shifting SERP formats toward imagery and non-text media. (Ahrefs)

Also read: What Are the 5 Breakthrough Trends Reshaping How Brands Get Found

4. Brand Spotlight Xponent21 and a Fast Climb in AI Answers

  • What changed? Xponent21 published a live case study showing a 4,162% year-over-year lift in organic visibility and new top placements in Perplexity, Google AI Mode, and ChatGPT for AI-search topics. (Xponent21)
  • Evidence: Their post documents a #1 Perplexity result for "how to rank in AI search results" within 20 days, frequent leading citations in Google's generative answers, and ChatGPT references. Screens and a dated timeline back the claims. (Xponent21)
  • Takeaway for marketers: Clusters win. Publish a flagship guide, surround it with FAQs and glossary entries, add explicit Q&A sections, and ship fast. Use schema to express entities and relationships so answer engines can cite you cleanly. Then track LLM mentions alongside Search Console. (Context on AI-Mode link concentration from Semrush's recent study.) (Semrush)

Also read: 10 GEO Trends That Will Supercharge Your AI Visibility Faster

5. Tactic of the Week: Train Your NLU in Copilot Studio

Custom NLU models in Copilot Studio — Microsoft's NLU+ update lets teams train grammar‑based natural‑language understanding models within Copilot Studio.

Users can define custom entities and synonyms to ensure bots interpret queries correctly. GEO practitioners can harness NLU+ to build internal help bots that mirror their brand's voice and product taxonomy.

Then, reuse that training data to refine external search results. This reduces off‑topic answers and creates a consistent experience across channels. (Microsoft Copilot)

Also read: See Where Brands Go Missing And How To Fix It

6. Next Moves To Lock GEO Wins

Models changed, surfaces appeared, and click paths moved. Turn this week's news into progress with short tasks you finish in days. Pick two and start now, let's take a close look:

  • Re-run your brand checks in ChatGPT now that GPT-5 is the default. Track the mention share on your top twenty queries. (OpenAI)
  • Test Copilot's Smart Mode with your terms. Check which internal pages Copilot cites from SharePoint and Outlook, then fix stale text. (Microsoft, Source)
  • Map "People also search for" inside AI Mode for priority queries. Add short follow-up Q&A blocks on matching pages. (Search Engine Roundtable)
  • Tighten sourcing on key pages. Add dates, named experts, and links that LLMs can cite to reduce vague claims. Use GPT-5's system card as your bar. (OpenAI)
  • Ship a one-page press kit and facts page for Perplexity. Confirm how Truth Social's new AI search displays your brand and correct gaps. (TechCrunch, Yahoo Finance)
  • Set daily alerts for Google volatility. Compare winners and losers by entity and query class instead of only the blue-link rank. (Search Engine Roundtable)
  • Audit mobile journeys in the Google app. The AI Mode button is now prominent, so check how your link cards and entities render. (Search Engine Roundtable)
  • Pull 24-month reports in Bing Webmaster Tools. Segment by device and country to separate model shifts from seasonality. (Search Engine Roundtable)
  • Update your prompt templates for internal testing. Use OpenAI's "minimal reasoning" guidance to cut contradictions and speed up checks. (OpenAI Cookbook)
  • Validate "clicks are stable" against your data. Compare CTR on AIO impressions versus classic listings before changing budgets. (Search Engine Roundtable)

In short, models changed, and surfaces moved. The way answers are chosen keeps shifting. Publish clear, dated facts and tight Q&A to stay visible.

To stay updated on GEO, AI SEO, and LLM rankings, read our latest articles.