Franc Talking – Weekly Round-up, 25 August 2025

Download our guide on AI-ready training data.

AI teams need more than big data—they need the right data. This guide breaks down what makes training datasets high-performing: real-world behavior signals, semantic scoring, clustering methods, and licensed assets. Learn to avoid scraped content, balance quality and diversity, and evaluate outputs using human-centric signals for scalable deployment.

Download the guide now

Welcome

Hello to existing and welcome to all new Franc Talking subscribers.

Great to have you here!

If you work in marketing, product, or tech - and want to keep pace with the shifts AI is driving across search, software and strategy - here’s what you need to know this week. Lets get to it! 🏋🏽‍♀️

  • 🔎 AI is reshaping search habits - Nielsen Norman Group research finds users are changing how they search, with more trust in conversational answers and less reliance on scanning multiple links. Rather than replacing traditional methods, people are engaging with AI search and traditional search in tandem, suggesting incremental change rather than outright replacement.
    Key takeout: although AI search is having a major impact on information discovery, adoption isn’t as fast as those of us working in tech might assume.
  • 🏠 Google brings Gemini to Nest - The Verge reports Google is integrating Gemini directly into Nest speakers, shifting them from simple smart devices to full AI companions. Voice search was all the rage in 2018, with some research firms even predicting that 50% of all searches would be voice by 2020 - a future that never arrived. Home assistants feel extremely dated compared with their chatbot counterparts, making this shift a welcome move for Google.
    Key takeout: Google is repositioning its smart-home devices for the AI era, moving away from the failed promise of voice search towards chat-driven interaction.
  • 🌐 OpenAI taps SerpApi for search data - According to The Information, OpenAI has been sourcing Google results via SerpApi to fuel ChatGPT’s real-time answers. The eight-year-old scraping firm even listed OpenAI as a client until last year. Google executives have quietly pushed back against SerpApi’s methods, underscoring the tension between AI providers needing fresh data and search incumbents guarding their ecosystems.
    Key takeout: ironic, isn’t it? Google built an empire by scraping the web and now ChatGPT is scraping Google’s results to power its own answers. It’s a reminder that in AI search, it may be cheaper to build on someone else’s index than to create your own from scratch.

    • Thanks to each and every one of you for subscribing and reading. Have a great week!

    Andy Francos

    Andy

    📖 What I’ve been reading this week in business, marketing & tech

    👇 Tap-worthy reads worth your time.

    ✨ AI

    🔍 Google & Search

    📈 Growth & Insights

    👋 Got 30 seconds?

    If you’ve got half a minute spare over your morning espresso, I’d really appreciate you taking 30 seconds to fill in this short survey.

    No weird personal questions — just a few quick ones to help me better understand who’s reading this, so I can keep improving and work more efficiently with the right partners.

    Bonus: I may run a small prize draw with one respondent chosen at random.

    Take the survey
    Recorded episode

    🎬 That is a wrap

    Andy and Duane Forrester in split screen chatting about search

    A snapshot from the recording: me and Duane Forrester in split screen, diving into the past, present and future of search.

    When I sat down with Duane Forrester - who helped shape Webmaster Tools at Microsoft, launched Schema.org and lived through Bing’s challenger years - the conversation quickly moved from nostalgia to urgency. We talked about the 2010s, when Google trained the world to type a word in a box and get an answer with remarkable consistency and how that shaped habits that still define the web today. We also covered how Google’s direction has shifted over the years - from social pivots that didn’t land, to revenue pressures that slowed change and why AI now feels like a completely different vehicle.

    The lesson? Search has always been about being the answer. That hasn’t changed. What’s changing is the mechanism: we’re shifting from scan-and-click SERPs to direct answers from AI systems that don’t forget, weigh trust differently and cite selectively. The challenge for brands is clear - adapt your content so it’s structured, entity-rich and discoverable by LLMs, while also rethinking how you measure visibility when clicks decline.

    Key actions Duane flagged were simple but effective: start building your content around entities and create a tracking matrix across AI systems (Gemini, ChatGPT, Perplexity, Claude) to see where you’re being cited and linked. Those who do this now will be ahead when AI discovery becomes the default.

    Don’t miss it - and if you know someone working in search who’d value this discussion, share it with them and encourage them to subscribe.

    🎯 Hey content creator!

    Thinking about launching your own newsletter? I recommend Beehiiv — it’s built for creators, with all the tools you need to grow:

    ✏️ Beautiful templates

    🚀 Powerful growth features

    📈 Built-in analytics

    💸 Monetisation tools

    All in one place, no fuss.

    🐝 Try Beehiiv with my link

    (You'll help support me if you sign up via this link 💙)

    📰 Recommended newsletter of the week -
    Evolving AI Insights

    Evolving AI Insights

    Evolving AI Insights

    Stay ahead with AI insights delivered every Monday, Wednesday and Friday.

    Reply

    or to participate

    © 2025 Franc Talking.

    Privacy policy

    Terms of use

    Powered by beehiiv