SUMM

Scrunch AI was founded after recognizing the disruptive potential of LLMs on how people interact with businesses online.
Early on, businesses mainly wanted chatbot plugins, but the founders realized most consumers don't want to engage with site-specific chatbots.
The core insight was that AI is fundamentally altering online discovery and customer journeys, not just traditional SEO.

Scrunch monitors how major AI platforms (ChatGPT, Gemini, Perplexity, Claude, Meta AI) surface and reference businesses in their search and conversational interfaces.
Focus is on platforms with significant user adoption; the coverage will expand as new platforms emerge.
Many brands are more concerned with AI search features (search box enabled) because user actions with commercial intent increasingly involve search, which pulls real-time or recent web content.
For pre-training crawlers, brands can control what is exposed, particularly avoiding including sensitive or short-lived promotion info.

There is growing debate and experimentation around compensating content creators for AI's use of their data (e.g., Cloudflare's paywall for bots).
Businesses must balance content restrictions with their need for exposure and traffic via AI/search engines.
New browsing paradigms (AI browsers, agent-based retrieval) are emerging, making it harder to entirely block LLMs from consuming web content.
The shift is not just AI replacing search but AI potentially replacing much of traditional web browsing for many user intents.

Most AI search platforms still use traditional ranking algorithms (BM25, PageRank, etc.) before applying AI-based re-ranking and summarization.
Traditional SEO practices continue to matter, but effective re-ranking depends on having descriptive, relevant page metadata and content.
Structurally clear, informative pages perform better in AI-generated responses than those optimized just for human clickthrough or visuals.

Businesses must decide how much product/context detail to publish for AI consumption—more detail improves discovery and inference by AI models.
AI search does not yet pass user query data to destination sites, mainly for privacy reasons, but inbound traffic analysis can guide content development.
Real-world example: Monitoring what topics or products inbound AI-driven visitors engage with helps refine content for future queries.

AI models sometimes hallucinate features or information; this can reveal demand for non-existent features, which can guide product/content development.
Clear, straightforward descriptions of offerings on product and homepage pages are crucial, especially for startups and B2B SaaS providers.
There is growing attention on how different AI systems re-rank and select content, akin to historical SEO tracking of Google algorithm updates.

The field is sometimes termed AEO (AI Engine Optimization) or GEO (Generative Engine Optimization), but Scrunch prefers to frame it as “agent experience.”
Unlike classic SEO, the goal expands from being discovered to managing the full AI-mediated user journey, as more customer actions complete within AI tools.

Personalization and user profiles (memories, preferences) already influence AI search results and content selection, with increasing effect.
Common high-value prompt clusters: coding, shopping, B2B software comparisons, and broader problem-solving rather than classic “search” intent.
Many business users leverage AI to directly solve problems (“do the thing for me”), which is higher intent than information gathering alone.

LLMs and AI search engines do not execute JavaScript when indexing; ensure key content is server-rendered for inclusion.
File formats like LLM.txt (or similar) are not yet major factors in discoverability—traditional, well-structured HTML content remains primary.
Over-focusing on embeddings and “semantic” hacks yields little benefit for exposure; clarity and structure are more effective.
Programmatic SEO/content generation is widely used; it works best if content brings unique, helpful insights rather than just regurgitated web material.
Effective cases include turning support tickets into structured how-to content for AI and users.

Clerk (developer authentication) saw a 6x increase in AI-driven traffic and a 9x increase in conversions by optimizing docs and content for AI platforms.
Scrunch acts as a feedback/experimentation tool rather than a content generation service, supporting customers to iterate based on data.
Typical uplift for actively engaged clients is double-digit percentage traffic growth within 1-2 months.

ChatGPT is the dominant AI platform in terms of traffic, engagement, and brand relationship, with AI overviews (Google, Bing) gaining exposure.
Perplexity follows as a strong, niche favorite, especially in tech/early adopter circles, though still smaller in volume.
Meta AI and Claude have passionate user bases but lower market share compared to ChatGPT.
Market shares are volatile; new entrants can spike in usage but ChatGPT has had the most enduring traction.

The web is rapidly shifting toward AI as a primary interface; being “AI compatible” is increasingly critical for exposure and conversion.
Avoid “snake oil” solutions; focus on clear, honest, well-structured content that helps both humans and AI agents.
Scrunch is actively hiring, especially for those interested in AI/search/web intersections.

AI is Eating Search