🗣️ Top Stories

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 from the week just gone.

🧩 OpenAI turns ChatGPT into an app platform – TechCrunch reports on the announcement that OpenAI is now allowing third-party apps to run inside ChatGPT via a new Apps SDK, with services like Spotify, Zillow, Canva and Expedia already integrated.

Although integration of apps isn't new to ChatGPT, this is a huge release and there will be many brands looking to following in the early footsteps of the brands listed above to get in front of 800M WAUs.

Key takeout: This is big news and it echoes how Facebook opened up to developers in the late 2000s. But unlike then, this move also directly challenges the traditional app store model - one that Apple and Google will be watching closely.

Many of the brands that I have spoken to are actively looking to build apps to capitalise on 'first move advantage' (as I'm sure everyone is). I also caught up with Simon Thillay of AppTweak to get his thoughts on the threat to Apple and Google.

“ I think OpenAI's announcement might be the most significant towards mobile so far, though it's worth noting Google was first in announcing an SDK (namely, the EngageSDK) to power AI features for apps. That being said, OpenAI's demo video certainly looks more impressive - ‘chat with any app’ - than what Google has announced so far around the EngageSDK. We might soon have an interesting race between both companies, with OpenAI benefitting from an OS-agnostic SDK while Google could play on the potential benefits not only for Gemini users but also for both consumers and developers in the Play Store.

Simon Thillay

Head of ASO, AppTweak

📰 UK journalists flooded by automated AI PR content – Press Gazette reports that a PR automation system is spamming UK media outlets with AI-generated press releases at scale -overwhelming reporters’ inboxes and flooding the news cycle with what many now call “AI slop.” For those spending real time crafting thoughtful, human-led outreach, visibility is vanishing fast.

Key takeout: It’s not entirely new, but the rise of tools like “Olivia Brown” is creating a major headache in UK newsrooms. Automated PR workflows - from idea generation to email distribution - are accelerating the volume problem already facing editors. One veteran journalist I’ve known for many years described the tool as “horrendous” and “truly dreadful,” warning of “huge ramifications - more slop, more distrust and less accountability.”

📊 CMOs scramble: “What’s our Reddit strategy?” – Marketing Advice captures the sudden panic in marketing teams as Reddit becomes the new battleground for attention. Sparked by a few misleading “engagement charts” doing the rounds, CMOs are now rethinking how to show up authentically on a platform built around scepticism, in-jokes and community gatekeeping.

Key takeout: Reddit’s rise over the past five years has been extraordinary. It’s become a goldmine of real conversations - and a data source for everyone from Google to OpenAI - but brand participation there can’t be forced. As the author points out, you can’t bulldoze your way in and expect to capture eyeballs overnight.

Reddit success demands patience and credibility. It’s a long-term play rooted in genuine community building, not campaign mechanics. I always advise clients to invest in an internal specialist - someone who lives and breathes the topic, speaks the language naturally and contributes to discussions with useful insight first, brand second.

Thanks for reading and have a great week ahead.

Andy

Exclusive offer for Franc Talking readers 🎉

Get 20% off for 3 months on all AppTweak business plans - including ASO Intell, Campaign Manager, Market Intelligence, and App Review Manager.

Duane Forrester on the evolution of search

When Duane Forrester talks about search, it’s not theory - it’s lived experience. As one of the early figures behind Bing’s Webmaster Tools and Schema.org, he’s seen the internet evolve from meta keywords and link farms to neural networks and large language models.

During his decade at Microsoft, Duane helped shape how search engines interpret intent and context - a foundation that still underpins how AI understands information today. Few people have witnessed the shift from data retrieval to answer generation quite so closely.

Subscribe to Duane's newsletter "Duane Forrester Decodes".

The next search revolution is upon us

At Microsoft, Duane Forrester helped launch Webmaster Tools and played a pivotal role in bringing Schema.org to life - bridging the worlds of technical SEO and structured data long before AI entered the frame. Talking to him feels like opening a living archive of the internet: part historian, part futurist, always a straight talker.

Our conversation moves between nostalgia and foresight - from Bing’s early battles with Google to the rise of ChatGPT and the next wave of generative search. Duane’s take is realistic: practical where others are panicked, curious where others are cynical. Beneath the humour and hard-won perspective lies one constant - a belief that search has always been about helping people find better answers.

As he puts it, search has always been a dialogue between humans and machines. We’ve spent decades learning how to make ourselves understood - and now, for the first time, the machines are learning to understand us.

The uncomfortable truth

Marketers have always been good at renaming things. But whether you call it SEO, GEO or AI Search, the core question remains: how do you make your brand visible when the interface itself is rewriting the rules?

Forrester’s advice is unglamorous: start with chunking. In AI terms, that means breaking content into small, semantically tight entities - facts, relationships, definitions - that large language models can easily ingest and reuse. The goal isn’t to flood the web with content; it’s to make what you already have machine-readable and answer-ready.

Then, build your own visibility matrix. Track how your brand appears - or doesn’t - across ChatGPT, Gemini, Claude, Copilot and Perplexity. Are you cited? Linked? Ignored? The exercise sounds manual, but it’s the first practical way to measure brand presence inside AI-generated responses.

✏️ Beautiful templates

🚀 Powerful growth features

📈 Built-in analytics

💸 Monetisation tools

What marketers need to do now

• ⚡ Optimise for comprehension, not clicks. Think less about meta tags, more about meaning. Short sentences, clear entities, structured relationships.
• 📊 Audit visibility across AI platforms. Create a prompt matrix and track where you appear. This becomes your new “ranking report.” No time? You could always use a platform like Obsero.
• 🧠 Retrain your teams. Every marketing job description should now include AI fluency - understanding how LLMs retrieve, chunk and reference information.
• 🤝 Keep humans in the loop. Forrester calls this his “hill to die on.” Strategy, judgement and creative context remain human strengths - the one advantage machines can’t replicate.