How to Recover Traffic Lost to AI Search (2026 guide)
AI Overviews now cut clicks by 58%, not the 30% reported in 2025. What the current research shows about recovering that traffic, and what Google says does not work.
Guilherme Hortinha · 18 September 2025

Why traffic is shifting to AI and how to recover it
If you’ve seen your organic traffic drop, you’re not imagining it. AI has changed how people search. According to Pew Research, when Google’s AI Overviews appear, people click traditional links only about half as often as they do without one. Ahrefs re-ran its 300,000-keyword study in February 2026 and found AI Overview presence now correlates with a 58% lower click-through rate for the top-ranking page, up from the 34.5% it measured in April 2025.
At the same time, adoption of generative search is skyrocketing. Google reported 2.5 billion monthly users for AI Overviews at I/O in May 2026, and AI Mode passed 1 billion monthly users in its first year, with queries more than doubling every quarter. adults have now used ChatGPT, double the number from just a year earlier. Google has rolled out AI Overviews across more than 200 countries, and these features now appear in billions of searches each month. The traffic hasn’t disappeared, it’s just flowing through new channels (MarketingCharts). The challenge is making sure your business shows up there.
Three paths to recover traffic from AI
Path | What it does | Speed to impact | Ongoing cost | Compounding power | Best when… | Key risk |
AI SEO optimisation | Positions your brand inside AI answers and Overviews | 2–8 weeks | Medium | High | You can own expertise in your niche | Needs content depth, not just one page |
Diversify demand capture (citations, PR, forums) | Gets you mentioned in sources AI trusts | 2–12 weeks | Medium | High | Your industry has active communities/directories | Requires consistent outreach |
Paid ads (Search, Shopping, Discovery) | Buys immediate visibility | Days | High | Medium | You have high-intent terms with good margins | Rising costs, ad fatigue |
How to choose: Paid ads plug short-term holes, but long-term resilience comes from AI visibility. For most businesses, the smartest play is to invest in AI SEO and off-site citations while layering paid ads only where they deliver a clear return.
What “AI SEO” really means
Classic SEO was about ranking for clicks. AI SEO, also called Generative Engine Optimisation (GEO), is about being understood, selected, and cited by AI systems like ChatGPT, Gemini, and Perplexity.
Generative engines don’t only crawl, but they also synthesise. They look for content that is accessible, structured, authoritative, and corroborated across multiple sources. To win, you need to think less about ranking positions and more about answer inclusion.
The five pillars of AI SEO

1. Accessibility
Make sure AI systems can actually reach your content.
Update robots.txt to allow reputable AI crawlers like GPTBot and OpenAI’s crawler.
In Cloudflare or your firewall, whitelist verified bots while blocking suspicious ones (Cloudflare Radar).
Keep sitemap.xml clean and up-to-date.
Add /llms.txt to point AI models to your most authoritative resources and define usage terms (Managing OpenAI’s web crawlers).
2. Authority
Generative engines reward clarity and coverage.
Publish content that covers every angle of your unique selling proposition.
Support claims with credible citations from research, surveys, and official docs.
Build off-site credibility by being cited in forums, Wikipedia, YouTube and Reddit.
3. Answer design
Free check
See how visible your brand is in AI
Ask AI the questions your customers ask, and find out if you show up.
Write like you’re crafting the perfect AI response.
Start each page with a clear, two-to-three-sentence direct answer.
Use short bullet lists or numbered steps.
Expand with detailed sections under H2/H3 subheadings.
Add structured data like FAQ or HowTo schema to make content easier to parse (Bain & Company insights).
4. Structure
Support content with technical clarity.
Implement schema markup for FAQs, products, reviews, and services.
Use canonical tags to consolidate duplicates.
Label your brand and services cleanly with Organization schema.
5. Feedback loop
Test, measure, and adjust.
Regularly ask ChatGPT, Gemini, and Perplexity the questions you want to rank for.
See what answers and citations appear.
Patch missing angles on your site and strengthen weak spots.
Run AI Visibility Audits to track progress (see how we did it at Indexed Lab).
Real-world proof: Wandr’s story
One of our clients, Wandr, a healthy restaurant in Dubai, faced exactly this challenge. Competitors were siphoning off customers, and Wandr’s organic traffic was shrinking. After implementing Indexed Lab’s AI SEO optimisation strategies, which included cleaning up their sitemaps, adding FAQ schema, creating answer-focused content, and making sure AI crawlers could reach their best pages, the brand began appearing in Gemini and ChatGPT responses for “healthy food JLT Dubai.”
“Working with the Indexed Lab team was eye-opening. We were losing customers to other healthy restaurants and didn't even know it. Since featuring in AI search results, we've doubled our online deliveries in just 3 months.”Anki Ram, CEO of Wandr
This is the power of AI visibility: not just defending clicks, but actively regaining customers through new discovery channels.

A 30-day action plan
Week 1: Accessibility
Audit robots.txt and firewall rules to ensure trusted AI bots are allowed.
Refresh your sitemap. Skip llms.txt: Google's documentation states plainly that Search does not use it, and Ahrefs' server-log study of 137,210 domains found 97% of published llms.txt files received zero requests in May 2026.
Week 2: Answer hubs
Identify your top 5–10 customer questions.
Create answer-style pages with direct responses, steps, and FAQs.
Week 3: Corroboration
Contribute useful, non-promotional answers in industry forums.
Build or update your Wikidata entry or submit to directories.
Publish fair comparison pages (your service vs. alternatives).
Week 4: Feedback loop
Ask AI engines your target questions.
Record which sources are cited.
Close the gaps with new or updated content.
Should you just buy Google Ads?
Buying ads can give you short-term visibility. If your ROI math works, ads are worth running. But the long-term shift is clear: fewer people click links when AI Overviews cause fewer clicks. Ads will always be competitive and rising in cost. Meanwhile, the opportunity to appear inside the AI answer compounds. That’s where businesses are quietly winning back discovery.
What are some counterarguments to optimise for AI SEO?
“AI summaries might link out more in the future.” That’s possible, but current data shows zero-click searches are rising. Optimising now ensures you’re already inside the answer if linking expands.
“I don’t want AI using my content.” A valid concern. The solution is selective: allow reputable bots while blocking abusive scrapers with firewall controls (Managing OpenAI’s web crawlers).
“We already have SEO articles.” Traditional blogs often aren’t AI-friendly. They’re too long, too keyword-stuffed, or too shallow. Refactor them into clean, concise, Answer-Model content sprints.
Expert signals and research
What changed in 2026
Three developments matter more than anything in the original version of this guide.
Google added generative AI performance reports to Search Console in June 2026. They show impressions from AI surfaces but deliberately withhold clicks, click-through rate, position and queries. You can now prove you appeared. You still cannot prove what it earned.
Rankings stopped predicting citations. Ahrefs found only 37.9% of AI-Overview-cited URLs now rank in the top 10, down from around 76% in July 2025. Nearly a third do not rank in the top 100 at all.
Bain's April 2026 analysis of roughly 500 million citations found 89% of citations for unbranded queries come from third-party sources rather than brand-owned content. Recovery is mostly an off-site problem.
One honest caveat worth carrying into any recovery plan: a July 2026 survey of 45 studies concluded that no published technique has yet shown a stable, cross-platform causal effect on discoverability, and that citation overlap between platforms is below 20%. There is no single list to get on.
While AI SEO is a relatively new topic, the body of research on the topic is rapidly developing. We want to mention some good pieces of research for further reading:
Pew Research (2025): Users click links about half as often when AI summaries are present (Pew Research).
Ahrefs & eMarketer (2025): Organic clicks can fall by 30–35% on queries with AI Overviews (Search Engine Land).
Google (2025): AI Overviews are live in 200+ countries: this is the new normal (MarketingCharts).
Academic research on GEO: Visibility has shifted from “rankings” to “answer inclusion” (arXiv).
Bain & Company (2025): AI agents prefer content that is conversational, list-based, and easy to scrape (Ars Technica).
Key AI SEO tips you can apply today
Put the facts you want quoted in visible body text. A controlled 2026 study found moving the same data out of JSON-LD and into visible text improved answer accuracy by roughly 30%.
Put the definition or answer in the first screenful of every article.
Build comparison pages with balanced pros and cons.
Create topic clusters: what, how, pitfalls, alternatives, case studies.
Use FAQ and HowTo schema to give engines clean structure.
Track AI citations, not just web clicks, as a new success metric.

Frequently asked questions
Is SEO dead because of AI?+
No. SEO is evolving. The game is now about being cited in AI answers and appearing in AI Overviews, not just chasing blue links
Should I block AI crawlers?+
Not all of them. Allow reputable crawlers like GPTBot if you want visibility in ChatGPT answers, but block or throttle stealth scrapers that ignore your rules
Can ads replace AI SEO?+
Ads can help, but they don’t compound. AI visibility does. The best strategy combines both, weighted toward long-term gains (PPC Land).
[ Keep exploring ]
Terms in this article

Co-founder & Head of AI Innovation, Index Lab
Guilherme co-founded Index Lab, an AEO/GEO agency that makes brands the answer AI gives across ChatGPT, Gemini and Perplexity, taking clients from zero AI visibility to top recommendations.
[ Related reading ]
See where your brand stands in AI.
Free visibility check →