Trustpilot
top of page
Search

The Professor's Guide to Teaching AI Search: From Keywords to Conversations

  • Writer: Clark Boyd
    Clark Boyd
  • Jul 9
  • 10 min read

The Search Revolution Your Students Need to Understand


Search is fundamentally changing.


While you've been teaching keyword research and traditional SEO, your students are spending 6.5 minutes on ChatGPT versus 2 minutes on Google. They're writing 23-word conversational prompts instead of 4-word keyword searches. And 70% of their searches now end without ever clicking through to a website.


"We're not teaching search anymore—we're teaching conversation with information," says Clark Boyd, CEO of Novela digital marketing simulations and faculty professor at Hult Business School. "The shift from keywords to natural language represents the biggest change in information retrieval since Google replaced the Yellow Pages."


This guide provides the data, frameworks, and practical exercises you need to teach AI search effectively.


The New Search Landscape: Platform Breakdown


Understanding Platform Correlations


Research from Grow and Convert analyzing over 400 commercial keywords reveals crucial insights about how AI search platforms relate to traditional SEO:


  • ChatGPT: 67% correlation with Google rankings (Grow and Convert, 2024)

  • Perplexity: 77% correlation with Google rankings (Grow and Convert, 2024)

  • Google AI Mode: Uses the existing index with new retrieval methods


But these numbers tell only part of the story.


Why ChatGPT Shows Lower Correlation


The 67% correlation reflects fundamental differences in how ChatGPT operates:


  1. Training Data Influence: Despite having web access, ChatGPT still heavily relies on its training data. When it searches the web, it uses Bing's API, but synthesizes results through the lens of its pre-trained knowledge (Grow and Convert, 2024).

  2. Source Preferences: Research shows 47.9% of ChatGPT citations come from Wikipedia, compared to just 5% for traditional Google searches (BrightLocal ChatGPT Search Sources Study, 2024).

  3. Authority Bias: ChatGPT heavily favors established, authoritative sources even when searching the web, often prioritizing well-known sites over newer, potentially more relevant content.


"This is critical for students to understand," Boyd explains. "Even with web access, ChatGPT's responses are filtered through its training data. It's more likely to cite established sources it 'knows' from training than to surface brand-new companies or emerging trends, even when they appear in search results."


Why Perplexity Mirrors Google More Closely


Perplexity's 77% correlation stems from its hybrid approach:

  • Real-time web crawling similar to Google

  • Shows sources for every claim

  • Updates with current information

  • More sensitive to recent ranking changes


Google AI Mode: The Query Revolution

Google's AI Mode represents a fundamental shift through "query fan-out":

  • Takes one search and generates hundreds of related queries

  • Retrieves information at the chunk level, not page level

  • Personalizes based on user context and history

  • Blends traditional results with AI-generated summaries


The Fundamentals That Haven't Changed


Before diving into what's new, it's crucial to understand what remains constant. Not everything about search has been revolutionized—the fundamentals still matter.


Quality Content Still Wins


"AI hasn't changed the fundamental truth that helpful, accurate information wins," Boyd states. "What's changed is how that information needs to be structured and delivered, not the need for quality itself."


The evidence is clear:

  • AI systems still prioritize authoritative, well-researched content

  • E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) is more important than ever

  • User intent remains the north star—understanding what people need hasn't changed


Authority Signals Have Evolved, Not Disappeared



Traditional authority signals haven't vanished—they've transformed:


Old Authority: Backlinks, domain age, PageRank


New Authority: Brand mentions, structured data, expert identity


An analysis of 75,000 brands revealed that brand web mentions have a 0.664 correlation with AI visibility (Master AI-Optimized Search and Content in 2025 Report, DMI). The same research found that top brands earn 10X more mentions than the next quartile. Authority still matters—it's just measured differently.


Technical Excellence Remains Critical


Fast, accessible, well-structured sites were important for Google. They're essential for AI:

  • Page speed still matters (AI crawlers have limited patience)

  • Mobile responsiveness remains crucial

  • Clean HTML structure is now non-negotiable


User Intent Is Still King


"Whether typed or spoken, whether 4 words or 40 words, users still have the same fundamental needs," Boyd notes. "The sophistication of how they express those needs has changed, but not the needs themselves."


Students must understand:

  • Information needs remain constant

  • Commercial intent still drives business

  • Navigation queries still exist (even if they bypass websites)

  • Transactional searches still convert


"The biggest mistake I see is educators throwing out everything they know about search," Boyd warns. "Build on the foundation. Don't demolish it."


The Engagement Revolution: From Seconds to Minutes


Session Duration Tells the Story


Data from multiple industry sources reveals a dramatic shift in how users engage with search:

  • Google Classic: <2 minutes average session (Datametrics, 2024)

  • ChatGPT: 6.5 minutes average session (Datametrics, 2024)

  • Perplexity: 12.28 minutes average session (Datametrics, 2024)


"This isn't just about time—it's about depth," Boyd notes. "Students are having extended conversations with AI, refining their questions, and exploring topics iteratively. Traditional 'bounce rate' metrics become meaningless when users never leave the platform."


The Query Length Revolution


Research from SparkToro and Semrush reveals:


  • Traditional Google searches: 4.2 words average (SparkToro, 2025)

  • ChatGPT prompts: 23 words average (SparkToro, 2025)

  • Maximum prompt lengths: Up to 2,712 words observed (SparkToro, 2025)


This 5x increase in query length represents a shift from keyword thinking to conversational thinking—a skill students must develop for the AI era.


The Technical Barrier That Changes Everything


The AI Crawler Limitation


Here's what every digital marketing professor needs to teach about AI crawlers versus traditional search crawlers:


Traditional crawlers (Googlebot): The marathon runners

  • Follow every link methodically

  • Execute JavaScript

  • Render full pages

  • Return multiple times


AI crawlers (GPTBot, ChatGPT-User, ClaudeBot): The snapshot takers

  • Can't execute JavaScript

  • No page rendering

  • One quick visit only

  • Grab raw HTML and leave


"Think of it this way," Boyd explains. "Google's crawler is like a food critic who dines in, tastes everything, and experiences the ambiance. AI crawlers are like delivery apps—they just need the menu with prices and ingredients listed."


The Critical Test Every Student Should Know


Teach students the "View Source Test":

  1. Visit any webpage

  2. Right-click → "View Page Source"

  3. Search (Ctrl+F or Cmd+F) for the main content

  4. If it's not in the source, AI can't read it


Real-World Implications


This isn't about all JavaScript—it's about content that ONLY loads via JavaScript:


Invisible to AI:

  • Product descriptions that load after page load

  • FAQ sections in JavaScript accordions

  • Reviews loaded dynamically

  • Content behind "Load More" buttons


Visible to AI:

  • Server-side rendered content

  • Static HTML

  • Pre-rendered pages

  • Content in the initial page source


"A React site isn't automatically invisible to AI," Boyd clarifies. "But if your critical content only appears after JavaScript executes, you've got a problem. AI crawlers are impatient—they grab the HTML and leave."


Solutions to Teach


  1. Server-Side Rendering (SSR): Generate HTML on the server

  2. Static Site Generation: Pre-build pages at compile time

  3. Progressive Enhancement: Core content in HTML, enhanced with JS

  4. Hybrid Approaches: Critical content static, interactive elements dynamic


From SEO to AEO and GEO: The New Optimization Paradigm


Answer Engine Optimization (AEO)


AEO represents a fundamental shift in optimization strategy:


Traditional SEO Goals:

  • Drive clicks to websites

  • Improve rankings

  • Increase traffic


AEO Goals:

  • Become the cited answer

  • Optimize for zero-click satisfaction

  • Focus on being THE source, not A source


"With AEO, success isn't measured in clicks—it's measured in citations," Boyd explains. "Would you rather have 1,000 visitors or be the answer that ChatGPT gives to millions?"


Key AEO Strategies to Teach


  1. Question-Based Content Structure

    • Use questions as H2/H3 headers

    • Provide complete answers in 40-60 words

    • Think "featured snippet" for every section


  2. Direct Answer Format

    • Answer first, explain second

    • Use clear, definitive language

    • Avoid "it depends" without specifics


  3. Schema Markup Evolution

    • FAQ schema becomes critical

    • Speakable schema for voice/AI

    • Article schema with clear structure


Generative Engine Optimization (GEO)


GEO goes beyond AEO by optimizing for how AI understands and synthesizes information:


Core GEO Principles:

  • Entity Optimization: Build clear connections between concepts

  • Topic Authority: Demonstrate expertise across related subjects

  • Semantic Richness: Use varied vocabulary AI can understand

  • Original Data: Provide unique statistics and research


"GEO is about becoming a knowledge node, not just a webpage," Boyd notes. "AI doesn't read pages—it builds knowledge graphs. Your content needs to feed that graph."


The New Metrics That Matter


Moving Beyond Traditional KPIs


Teach students these emerging metrics:


  1. AI Visibility Score: How often content appears in AI responses

  2. Citation Rate: Frequency of being referenced by AI

  3. Share of Voice: Percentage of AI responses mentioning your brand

  4. Semantic Coverage: Breadth of topics where you're cited

  5. Authority Score: Credibility rating in AI systems


Platform-Specific Metrics


For ChatGPT:

  • Training data inclusion (pre/post cutoff)

  • Wikipedia citation correlation

  • Domain authority influence


For Perplexity:

  • Real-time citation frequency

  • Source ranking position

  • Follow-up query generation


For Google AI Mode:

  • AI Overview inclusion rate

  • Chunk selection frequency

  • Query expansion coverage



The 2,300% Success Story


The Search Initiative documented an industrial manufacturer that saw zero AI visibility despite strong traditional SEO. After implementing AEO strategies (Master AI-Optimized Search and Content in 2025 Report, DMI):


  • 2,300% increase in AI referral traffic

  • 90 keywords appearing in AI Overviews (from 0)

  • 6,175% increase in AI-driven conversions


Key changes:

  • Converted all content to Q&A format

  • Added comprehensive FAQ sections

  • Implemented proper schema markup

  • Created topic clusters around product categories

The E-commerce Revolution

Auto parts retailer Hedges & Company implementing AI-friendly structure saw (Master AI-Optimized Search and Content in 2025 Report, DMI):

  • 14% higher engagement from AI traffic versus organic

  • 6% better conversion rates

  • Lower bounce rates despite zero-click searches

"These aren't anomalies," Boyd states. "Companies optimizing for AI search are seeing explosive growth while others wonder where their traffic went."

Practical Classroom Exercises to Teach AI Search

Exercise 1: The Platform Comparison Lab (20 minutes)

Objective: Understand platform differences

Students search the same commercial query across all platforms:

  1. Document result differences

  2. Analyze source preferences

  3. Time the interaction

  4. Compare information depth

Debrief: Which platform serves which user intent best?

Exercise 2: The Correlation Test (30 minutes)

Objective: Verify the 67% and 77% correlations documented by Grow and Convert (2024)

  1. Select 10 keywords from a client/project

  2. Check Google rankings

  3. Test same keywords in ChatGPT and Perplexity

  4. Calculate correlation

  5. Identify patterns in what appears/doesn't

Insight: Students discover why newer brands struggle with AI visibility

Exercise 3: The JavaScript Audit (15 minutes)

Objective: Identify technical barriers

  1. Choose 5 competitor websites

  2. Run the View Source test

  3. Use Chrome DevTools to see rendered content

  4. Calculate % of content that's AI-invisible

  5. Propose solutions

Application: Students can immediately apply this to client work

Exercise 4: AEO Content Transformation (45 minutes)

Objective: Convert traditional content to AI-friendly format

  1. Take existing blog post or product page

  2. Restructure with question headers

  3. Add direct answers (40-60 words)

  4. Implement FAQ schema

  5. Test in AI platforms

Results: Immediate visibility improvement

Teaching the Future of Search

The Skills Students Actually Need

"Stop teaching keyword density. Start teaching conversational intelligence," Boyd advises. Essential skills include:

  1. Natural Language Query Construction

    • Moving from keywords to questions

    • Understanding user intent layers

    • Crafting follow-up queries

  2. Multi-Platform Strategy

    • When to use which platform

    • Cross-platform optimization

    • Understanding platform limitations

  3. Technical Implementation

    • Schema markup mastery

    • JavaScript workarounds

    • Structured data optimization

  4. Content Architecture

    • Topic clustering for AI

    • Entity relationship building

    • Semantic content mapping

The Curriculum Evolution

Traditional SEO Module (Still Relevant):

  • Keyword research (evolved to topic research)

  • On-page optimization (now includes AI-friendly structure)

  • Link building (expanded to mention building)

  • Technical SEO (with AI crawler considerations)

AI Search Module (New Additions):

  • Conversational query analysis

  • Multi-platform visibility strategies

  • Answer optimization techniques

  • Entity-based SEO

  • Zero-click optimization

  • AI crawler behavior

  • Schema markup as core curriculum

"Don't replace your SEO curriculum—expand it," Boyd advises. "Students need both skill sets. Traditional SEO for Google's 93.57% market share (SparkToro, 2025), AI optimization for the future that's rapidly arriving."

Industry Implications to Discuss

The Zero-Click Economy

With 60% of searches ending without clicks (SparkToro, 2024), students need to understand:

  • Brand impressions matter more than visits

  • Being the answer beats being a result

  • Measurement shifts from traffic to influence

The Platform Wars

"Google processes 373x more searches than ChatGPT today," Boyd notes, citing SparkToro's 2025 research. "But ChatGPT users are 2x more likely to take action. Students need to understand it's not about volume—it's about intent and engagement quality."

The Technical Divide

Companies will split into two camps:

  • Those visible to AI (static, structured, semantic)

  • Those invisible to AI (dynamic, JavaScript-heavy, unstructured)

Action Items for Professors

Immediate Curriculum Updates

  1. Add AI platform comparison to keyword research modules

  2. Include schema markup as core, not advanced topic

  3. Teach question-based content structure

  4. Cover technical barriers to AI visibility

  5. Update metrics to include AI-specific KPIs

Resources to Create

  • Platform comparison templates

  • AEO optimization checklists

  • Schema markup generators

  • AI visibility audit frameworks

  • Updated case studies featuring AI success

Industry Connections

"Bring in practitioners who are doing this work," Boyd suggests. "Theory is important, but students need to see real campaigns, real results, and real failures. The field is moving too fast for textbooks alone."

The Bottom Line

Search is fundamentally changing. It's no longer about matching keywords—it's about understanding and optimizing for how AI systems comprehend, synthesize, and present information.

"The professors who adapt their curriculum now will prepare students for reality," Boyd concludes. "Those who don't will send graduates into the job market with obsolete skills. The choice is yours."

Quick Reference Guide

Platform Cheat Sheet

  • ChatGPT: 67% Google correlation, 2023 knowledge cutoff, Wikipedia-heavy

  • Perplexity: 77% Google correlation, real-time data, shows sources

  • Google AI: Query fan-out, personalized, chunk-level retrieval

Key Metrics to Teach

  • AI Visibility Score

  • Citation Frequency

  • Zero-Click Optimization Rate

  • Entity Coverage

  • Semantic Richness Score

Technical Must-Knows

  • 98.8% of sites use JavaScript (hard for AI to read)

  • View Source test for AI readability

  • Schema markup is now essential, not optional

The New Optimization Framework

  • SEO: Search Engine Optimization (still relevant)

  • AEO: Answer Engine Optimization (be THE answer)

  • GEO: Generative Engine Optimization (feed the knowledge graph)

About Novela

Novela provides digital marketing simulations and AI Marketing training for universities worldwide. Our platform helps students practice real-world AI optimization strategies in a safe, measurable environment.

For curriculum resources, guest lectures, or platform demos, contact:

This guide reflects the AI search landscape as of July 2025. For quarterly updates and new research, subscribe to our educator newsletter.

References

Note: Where available, direct links are provided. Some studies are referenced through industry reports or are proprietary research not publicly available online.

BrightLocal. (2024). ChatGPT Search Sources Study. Analysis of Wikipedia and source preferences in AI citations. [Study findings referenced in industry reports]

Datametrics. (2024). Average Session Time Comparison: Google vs ChatGPT vs Perplexity. User engagement metrics across search platforms. [Industry data]

Digital Education Council. (2024). Survey: 86% of Students Already Use AI in Their Studies. Student AI adoption research. https://campustechnology.com/articles/2024/08/28/survey-86-of-students-already-use-ai-in-their-studies.aspx

Digital Marketing Institute. (2025). Master AI-Optimized Search and Content in 2025: Complete DMI Masterclass Research Report. Analysis of 75,000 brands and AI visibility correlations. [Internal research document]

Grow and Convert. (2024). Does Google SEO Affect LLM Optimization? We Analyzed 400+ Keywords to Find Out. Platform correlation study. https://www.growandconvert.com/ai/google-seo-and-llmo/

Hedges & Company. (2024). Auto Parts E-commerce AI Implementation Case Study. AI traffic engagement metrics. [Case study findings]

Pew Research Center. (2025). Share of Teens Using ChatGPT for Schoolwork Doubled from 2023 to 2024. U.S. teen AI usage statistics. https://www.pewresearch.org/short-reads/2025/01/15/about-a-quarter-of-us-teens-have-used-chatgpt-for-schoolwork-double-the-share-in-2023/

Semrush. (2024). ChatGPT Prompts Analysis: 80 Million Clickstream Prompts Study. Analysis of prompt types and usage patterns. [Referenced in SparkToro analysis]

SparkToro. (2024). Zero-Click Search Study. Analysis of searches ending without clicks. [Industry report]

SparkToro. (2025). New Research: Google Search Grew 20%+ in 2024; receives ~373X more searches than ChatGPT. Market share and query analysis. https://sparktoro.com/blog/new-research-google-search-grew-20-in-2024-receives-373x-more-searches-than-chatgpt/

The Search Initiative. (2024). Industrial Manufacturer AI Visibility Case Study. Implementation results and traffic metrics. [Case study findings]

bottom of page