AI Search Is Eating Organic Traffic -- What to Do

Google AI Overviews now answer 30-40% of informational queries directly in the search results. Pages rank fine but clicks are softer because AI answers absorb the queries that used to drive traffic. The brands winning right now stopped optimizing for "what is X" content and doubled down on content with unique data that AI cannot synthesize from training data.

The Traffic Pattern Shift

Semrush lost roughly half its traffic on definitional content ("what is SEO", "what is a backlink") that AI Overviews now answer directly. HubSpot lost nearly all visibility on intro-heavy content that starts with "In this guide, you will learn..." because AI Overviews skip the preamble and answer the question immediately. The content that survives has something AI Overviews cannot generate: verified pricing comparisons, working code, original benchmarks, proprietary data.

What to Do About It

The practical move is to track where you are being cited in AI answers, not just where you rank in traditional SERP. If AI is pulling from your content, you are building authority even if direct clicks drop.

1. Stop writing "what is X" content unless you can add unique data (verified pricing, code examples, original research)
2. Track AI Overview citations for your target keywords daily
3. Structure content answer-first: the first sentence should directly answer the query, no preamble
4. Include data that AI cannot synthesize: live pricing comparisons, working API examples, verified benchmarks

How to Track AI Overview Citations

Use a search API that returns AI Overview data alongside organic results. Query your target keywords daily, check whether your domain appears in the AI Overview citations, and alert when citations are gained or lost.

import requests, os, json
from datetime import date

H = {"x-api-key": os.environ["SCAVIO_API_KEY"]}
DOMAIN = "yourdomain.com"
KEYWORDS = ["best search api 2026", "serp api comparison"]

def check_citations(keywords, domain):
    results = []
    for kw in keywords:
        r = requests.post("https://api.scavio.dev/api/v1/search", headers=H,
            json={"platform": "google", "query": kw,
                  "include_ai_overview": True}, timeout=10).json()
        aio = r.get("ai_overview", {})
        citations = aio.get("citations", [])
        cited = any(domain in c.get("url", "") for c in citations)
        results.append({
            "keyword": kw, "date": str(date.today()),
            "cited": cited, "citation_count": len(citations)
        })
    return results

for r in check_citations(KEYWORDS, DOMAIN):
    print(json.dumps(r))

The Content That Wins in AI Search

Semrush's pages that DO get cited in AI Overviews share a pattern: answer-first format, no introduction, unique data in the first paragraph, and structured content that AI can extract cleanly. The pages that lost visibility had long introductions, generic definitions, and no data that could not be synthesized from training data.

For developer and SaaS companies, the winning content types are: verified pricing comparison tables (with dates), working code snippets with real API calls, side-by-side feature comparisons with specific numbers, and decision frameworks based on concrete tradeoffs rather than generic advice.

Separate AI Traffic Loss from Ranking Loss

Not all traffic drops are AI-related. Before attributing a decline to AI Overviews, check whether your organic position actually changed. If you still rank #1 but clicks dropped, AI is likely absorbing the query. If your position dropped, the issue is traditional SEO, not AI.

The correlation between AI Overview presence and click-through rate decline is strongest for informational queries. Transactional queries ("buy X", "pricing for Y") and navigational queries ("brand name login") are less affected because AI Overviews are less common on those query types.