AI Marketing Pulse: Key Developments from Tuesday, 13th May 2025

Meanwhile, newer AI models are showing concerning error rates up to 79%, creating risks for marketers. However, a technology called RAG (Retrieval-Augmented Generation) offers a solution to AI reliability issues—and it's something I've spent considerable time researching for our AI implementations.Story #1: AI Overviews Have Doubled (25M AIOs Analyzed)What Happened: Ahrefs Brand Radar tracked the total number of AI overviews between 6th February and 6th May, revealing the most substantial day-to-day change in the dataset (6.26%) occurred between 12th and 13th March, coinciding with the rollout of Google's latest Core Update. The total number of AI Overviews grew by 116% between March 12th (pre-update) and May 6th. AI Overviews Have Doubled (25M AIOs Analyzed)Why It Matters: Search volume shows similar growth: AI overviews now account for 11.8% of all US-based keyword search volume, compared to 6.2% on March 13th. AI Overviews Have Doubled (25M AIOs Analyzed) This expansion directly impacts organic search visibility and CTR for marketers who rely on search traffic.Suggested Actions:* Monitor your website traffic in Google Search Console to identify drops that may correlate with AI overview expansion* Consider shifting from traffic-centric metrics to brand awareness indicators to better understand your position within the AI ecosystemSource: Read the full Ahrefs analysisStory #2: New AI Models Make More MistakesWhat Happened: The newest AI tools, built to be smarter, make more factual errors than older versions. Tests show errors as high as 79% in advanced systems from companies like OpenAI. OpenAI's latest system, o3, got facts wrong 33% of the time when answering questions about people. That's twice the error rate of their previous system. Its o4-mini model performed even worse, with a 48% error rate on the same test. New AI Models Make More Mistakes, Creating Risk for MarketersWhy It Matters: These aren't just abstract problems. Real businesses are facing backlash when AI gives wrong information. Last month, Cursor (a tool for programmers) faced angry customers when its AI support bot falsely claimed users couldn't use the software on multiple computers. New AI Models Make More Mistakes, Creating Risk for Marketers For marketers using AI for content creation or customer service, this trend represents significant brand risk.Suggested Actions:* Implement rigorous fact-checking processes for all AI-generated content* Consider using older, more reliable AI models for customer-facing applications* Build human review into all AI-powered customer service workflowsSource: Read the full SEJ reportStory #3: RAG - The AI Reliability Solution Marketers NeedWhat Happened: MarTech's latest analysis highlights Retrieval-Augmented Generation (RAG) as the most important AI tool marketers haven't heard of. RAG enables AI systems to access and reference specific information from curated sources, dramatically reducing hallucinations and increasing accuracy—a technology I've been researching extensively for marketing applications.Why It Matters: With AI error rates climbing (as we've seen in story #2), RAG offers a solution by grounding AI responses in verified data sources. Instead of generating information from scratch, RAG-powered AI pulls from trusted databases, making it ideal for brand-safe marketing applications.Suggested Actions:* Explore RAG implementations for customer service chatbots* Consider RAG-based content creation tools that reference your brand guidelines* Investigate how RAG could improve AI accuracy in your marketing stackSource: Read the full MarTech articleComprehensive SummaryThe landscape on 13th May 2025 presents both challenges and solutions for AI in marketing. Whilst Google's AI Overviews have expanded dramatically—now affecting 11.8% of US search volume—the technology powering AI is becoming paradoxically less reliable, with newer models showing error rates up to 79%.However, the emergence of RAG (Retrieval-Augmented Generation) offers a promising solution. Having spent considerable time researching this technology, I can attest to its potential for solving AI's reliability crisis. RAG grounds AI responses in verified data sources, eliminating the hallucination problems that plague standard AI models.This convergence of trends—expanding AI presence in search, declining AI reliability, and emerging solutions like RAG—suggests we're at an inflection point. Marketers who understand and implement these technologies strategically will have a significant advantage as AI continues to reshape the digital landscape.FREE BETA ACCESS: We've launched our AI for SEO course to help you implement these cutting-edge strategies. Get free access during the beta period at https://seoaicourse.indexify.co/ Key Takeaways* AI Overview dominance requires new strategies: With AI Overviews doubling and affecting nearly 12% of search volume, traditional SEO metrics need updating.* Verify everything: With error rates up to 79% in newer AI models, human oversight is critical for brand protection.* RAG offers a solution: Retrieval-Augmented Generation technology can solve AI reliability issues by grounding responses in verified data.How-To SpotlightImplementing RAG for Marketing Applications:* Identify your trusted data sources (brand guidelines, product databases, verified content)* Choose a RAG-compatible platform (many enterprise AI tools now offer RAG capabilities)* Create a structured knowledge base that RAG can reference* Test with low-risk applications first (internal tools before customer-facing)* Monitor accuracy improvements compared to standard AI* Gradually expand to customer-facing applications once reliability is proven* Maintain and update your RAG knowledge base regularlyPro tip: Start with FAQ chatbots—they're perfect for RAG implementation as they require accurate, consistent responses based on verified information.Additional Resources* Understanding RAG Technology* Google's AI Overview documentation* Ahrefs' AI Overview tracking guideSubscribe & ListenSubscribe to our daily updates: and listen to our podcast: https://indexify.substack.com/podcast Get full access to Jon’s Substack at indexify.substack.com/subscribe