July 2025
🚀 The Content Revolution
Content is multiplying across every platform—powered by AI, people can generate endless articles, videos, even ads in minutes. That flood of content empowers big aggregators like Google, YouTube, and Facebook, since their AI summaries let people find answers without clicking through to the original sources.
So simply producing more won't be enough. Instead, the real winners might be those who build trust and genuine connections. Community will matter. Clients will want voices they can rely on, not just more noise.
Meanwhile, smart marketers might see another shift: AI agents acting on our behalf. These digital helpers will browse, shop, and sort information for us. That means our messages must reach these agents and humans alike—so we'll need ideas that speak to both. It's like a new era of Search Engine Optimization.
Finally, content will get smaller and smarter. Think 30-second clips tailored to each person, powered by AI. To stay ahead, focus less on volume, and more on sincerity, shared purpose, and bite-sized insights that will spark real conversations.
🎥 Watch the video summary of this newsletter
-
Insightful article by Ben Thompson: AI has made content production effortless and infinite, accelerating the commoditization of media. This flood of content strengthens the dominance of aggregators like Google, YouTube, and Facebook — especially with AI features like AI Overviews that summarize information directly in search, reducing clicks to original sources. Since everyone can scale, scale alone isn't enough and creators must add something new to their focus. What might truly matters now is trust, connection, and community. In an age of infinite content, the real differentiator is sincerity — the ability to build genuine relationships around shared meaning and intent.
| #AI #Content #Cost #Marketing | Ben Thompson | 2025/07/21
-
Interesting vision that presents a future where the internet is no longer primarily shaped by humans, but by AI agents acting on our behalf. As tools like ChatGPT begin browsing, buying, and emailing for us, our role shifts. This changes the incentive to post on social media, since audiences may be machines, not people. It marks a new chapter where human creativity must adapt to an online world increasingly navigated by artificial agents.
| #Agent #Automation | sineadbovell | 2025/07/21 -
An insightful article outlines key trends shaping future internet content user engagement: dominance of hyper-personalized microformats, rise of short-form videos (Reels, Shorts), elimination of universal Trending charts, prominence of timeless content on LinkedIn, and heightened risk of echo chambers. It recommends that creators leverage AI-driven tools and bite-sized formats to build niche communities; marketers embrace data-fueled, context-specific short-form ads; and users diversify their feeds deliberately to balance personalization with broader perspectives. My take: In the bigger picture, it has always been about shorter and more mobile content — video formats evolved from 60+ minute TV shows to 10–20 minute YouTube videos, then 1–2 minute social media clips, and now 30-second Reels/Shorts — reflecting the shift from TV screens to laptops and finally mobile phones, with content becoming even more precisely tailored to individual interests. AI tools will definitely accelerate this hyper-personalized, bite-sized trend.
| #Content #Marketing | Abir "Bandy" Bandyopadhyay | 2025/07/20
-
The Economist explores a growing dilemma, as AI-powered search engines reduce site traffic by showing answers directly. Some advocate a pay-as-you-crawl model charging AI bots for access. Others develop tools to trace chatbot responses back to sources, enabling proper compensation. Another article explains why, we have this phenomenon, AI Overviews are changing how we see clicks and impressions. While impressions are going up, clicks are going down. This means more people see your content in the search engine, but fewer are actually visiting your website. Impressions are increasing because AI Overviews now give companies two chances to log an impression for a given keyword: once as a "traditional" blue link in the search results, and again as a citation in an AI Overview. My take : It's important to understand this shift, as it affects how we approach online marketing and how we measure success in this new landscape.
| #AI #Marketing #SEO | The Economist | 2025/07/19
-
AI Overviews are changing how we see clicks and impressions. While impressions are going up, clicks are going down. This means more people see your content, but fewer are actually visiting your website. My take : It's important to understand this shift, as it affects how we approach online marketing and how we measure success in this new landscape.
| #AI #Marketing #SEO | Ahrefs | 2025/07/17
-
OpenAI's COO describes agents as autonomous systems solving new, complex tasks autonomously, enhancing productivity. For example, they can write and test code or manage many leads in a sales process. This ability allows teams to focus on important work while agents take care of routine tasks.
| #Agent #Leader | OpenAI COO Brad Lightcap | 2025/07/17 -
OpenAI has begun rolling out ChatGPT Agents, a powerful new feature built for complex task automation. It integrates autonomous text and visual internet browsing, image generation tools, deep research capabilities, operator functions, and conversational intelligence into a single system. The new OpenAI feature is the closest realization of the AI agent vision available today. In OpenAI demos, the Agent created a slide deck from Google Drive documents in just 7 minutes. In another, it planned an entire wedding itinerary—including outfits, hotels, and gifts—using online tools. It also generated 500 anime mascot stickers in 7 minutes and developed an MLB stadium travel plan complete with maps and spreadsheets. Some early testers start to confirm its potential—for example, generating a 2,000-word report from 1300 customer emails in just 15 minutes
| #Agent #Leader | OpenAI | 2025/07/16 -
There are also some less positive or cautious opinions about the new OpenAI Agent feature. Some point out that its efficiency in general tasks—like ordering a cupcake online—is questionable. In one example, the process took 60 minutes and included many unnecessary steps. It did eventually finish, but very slowly and with limited effectiveness. My take: I agree that, at this stage, allowing a large model to perform tasks autonomously can sometimes lead to wasted time or even failure to complete the task. While there are simpler use cases—like serving as an Excel assistant—we still lack enough real-world testing to draw solid conclusions.
| #Agent #Leader | nate.b.jones | 2025/07/15 -
B2B teams are struggling with tool overload and a lack of clear strategy. Many are experimenting with AI solutions without a unified plan, resulting in confusion and limited growth. Experts recommend prioritizing system design and integration that aligns signals, rather than adding more tools. Real success comes from leveraging meaningful signals and fostering cross-department collaboration. My take: This may sound a bit harsh, but it's a fair point. The proliferation of marketing tools can lead to spaghetti architecture and a loss of efficiency. It's something we should definitely keep in mind—especially today, when something new seems to emerge every single day.
| #Marketing #Tools | martech.org | 2025/07/14 -
Claude now connects directly to popular tools like Notion, Canva, Stripe, and Figma. A new directory makes it easy to link Claude to your favorite apps. My take: Leading generative AI companies are betting on ecosystem-wide integration. It's a strong sign that the AI with the best connections—not just the best model—may come out on top.
| #AI #Leader #Tools | Anthropic | 2025/07/14
-
Grok 4 is the latest version of xAI's AI system, presented as a major step forward. The xAI team claims it outperforms PhD-level benchmarks, excels in large-scale reinforcement learning, and performs exceptionally well on standard tests.
My take: When it comes to novelty, we've seen similar announcements for years—yet another model beating benchmarks. That's why, for me, the most interesting part of this update was the in-car integration. This could be genuinely useful: just like we listen to podcasts while driving, we might soon be able to have real conversations with an AI assistant on the road. This kind of natural, frictionless access could make AI feel more helpful—and potentially start shifting consumer behavior. Lately, I feel that raw model power isn't the most important factor anymore. What really matters is how well the model connects to real-world systems, and how easy it is to build agent frameworks around it that keep humans in the loop. At the same time, Grok 4 has stirred both excitement and criticism—especially around concerns of political bias and the controversial "MechaHitler" incident. So it's hard to form a clear opinion at this stage. The team also demonstrated how Grok 4 could autonomously operate vending machines in a simulated business environment, but that part felt a bit too optimistic. As with all new technologies, the real test will be how it performs in everyday use.
| #AI #Automation | xAI | 2025/07/09 -
Nvidia has reached a historic milestone by achieving a market value of $4 trillion, making it the first public company to do so. This success is largely due to its leadership in artificial intelligence and semiconductor technology. The company also plans to expand its operations in Israel, reflecting its ongoing growth and innovation in the tech industry. My take: This milestone reminds us of past giants like Intel, IBM, Nokia, and Cisco—once leaders during their industry's defining boom.
| #Leader | Reuters | 2025/07/09 -
Interesting reflection: AI can automate tasks that are clearly defined and can be answered with a simple true or false. However, tasks that involve ambiguity require human judgment. For example, creative work like writing or designing still needs a human touch. It's important to recognize when AI can operate independently and when human involvement is necessary to achieve better outcomes.
| #Automation #Marketing | Andrew Ships | 2025/07/09 -
According to ahrefs, Google does not punish AI-generated content in search results. A study of 600k web pages found that 86.5% of top-ranking pages include some AI-written material. The research shows there is no strong link between AI content and search rankings. Google focuses on whether content is helpful for users, regardless of how it was created. My take: Interesting question arises here: with content creation easier than ever, will search engines start promoting more authentic, human-made material? So far, Google stands by its promise — the best content wins. Whether created by humans or AI, the key to success is knowing how to produce truly valuable content.
| #AI #Marketing #SEO | Ahrefs | 2025/07/07
-
H&M has launched a new marketing campaign featuring AI-generated digital models known as "digital twins." These virtual models replace real ones in photoshoots, allowing the company to accelerate production and reduce costs. To maintain transparency, H&M clearly labels AI-generated images and shares revenue with the original human models. More AI-driven campaigns are planned to gauge customer reactions to digital fashion. This move reflects a broader shift among other industry players, such as Zalando. Matthias Haase, Vice President of Content Solutions at Zalando, told Reuters: "We are using AI to be able to be reactive." He explained that generative AI has reduced image production time from six to eight weeks down to just three or four days, while also cutting costs by 90%. According to Haase, AI-generated content also leads to higher customer engagement. My take: Creative work is deeply intuitive, but early wins from H&M and Zalando show automation's potential. The question remains: will this shift improve customer satisfaction and relationships, and boost employee fulfillment beyond just cutting time and costs?
| #AI #Influence #Marketing | FASHIONUNITED | 2025/07/04 -
Marketers are losing nearly 12 hours each week due to ineffective communication. Many are stuck in constant meetings and messages. To improve productivity, it's suggested to block time for focused work, reduce unnecessary meetings, and use AI tools effectively. These changes can help marketers focus on what truly matters. My take: Even if the article sounds harsh, it reflects a real trend — tool proliferation often requires rationalization to clarify core workflows the whole team aligns with, which then boosts overall efficiency.
| #Automation #Marketing #Team #Tools | Dave Gerhardt | 2025/07/02 -
Google launches the A2A protocol, enabling AI agents to collaborate across platforms. As Google partners explain it, GrowthLoop (a marketing-tech partner) states: "A2A will optimize campaigns across the marketing ecosystem," while Box highlights its potential for "empowering organizations to better automate workflows, lower costs, and generate trustworthy AI outputs." My take: A2A lays the groundwork for next-gen advertising—automated, scalable, and deeply personalized through agent collaboration. It also echoes Aravind Srinivas's (co‑founder and CEO of Perplexity AI) vision of a business model where we pay agents to consume advertising. We're already seeing signs of this with Google Search's AI Overviews, where users click less on websites—suggesting a shift toward "pay-as-you-crawl" or "pay-as-you-show" models. Finally, it resonates with ChatGPT's offer to let brands be visible inside the assistant, opening new formats for in-context advertising. Definitely a topic to watch.
| #Agent #Marketing | Google | 2025/07/01
-
According to Zalando, it now uses generative AI to create marketing visuals in just 3–4 days, down from 6–8 weeks, cutting costs by 90%. Zalando also creates "digital twins" of models to scale campaigns. My take: This mirrors what we've seen from H&M, which is also accelerating marketing through generative AI. It also aligns with the broader trend of exploding content volumes online as production costs continue to drop.
| #AI #Cost #Marketing | Reuters | 2025/07/01
-
Andrej Karpathy, the former Director of AI at Tesla, recently shared his thoughts on context engineering, which he believes is more effective than prompt engineering for improving large language model (LLM) performance. Context engineering focuses on filling the model's context window with the right information for a given task. This approach helps optimize performance, manage costs, and make better use of available data. As LLM applications continue to grow, context engineering is becoming a crucial part of their development and functionality.
| #AI #Analytics | Andrej Karpathy | 2025/07/01 -
Optimizing how we input information can improve the effectiveness of large language models (LLMs). By managing context better, we can get clearer and more useful outputs. Using standard message formats is good, but creating custom formats can make the process even better. Understanding user needs and providing the right context is key to maximizing LLM performance.
| #AI #Analytics | HumanLayer | 2025/07/01 -
AI is changing the search engine market. Apple's Safari experienced its first drop in searches, likely due to more people using AI tools. It looks like that's started to happen, an Apple executive testified in court. This shift worries Google, which is now trying to adapt by using its own AI, Gemini. Despite these efforts, Google's stock fell over 7%, indicating challenges ahead in maintaining its search dominance.
| #AI #Marketing | Business Insider | 2025/07/01 -
A new study shows that more Americans are getting their news from social media than from TV. Now, 54% of people prefer social media for news, compared to 50% who still watch TV. This shift is especially pronounced among younger people, particularly those aged 18 to 24, many of whom watch news videos on their phones each week. My take: We already know that platforms like Facebook, X, and YouTube are increasingly integral to the customer funnel and influence decision-making. But the real question now is: just how dominant have they become?
| #Marketing | Oxford's Reuters | 2025/07/01
-
Aravind Srinivas, CEO of Perplexity AI, proposed a bold future for advertising: AI agents—not only users—will receive ads. Brands will compete for agents' attention, enabling personalized, invisible monetization where agents filter offers and take commissions without disrupting the user experience. My take: Crazy as it may sound, many parts of the ecosystem are aligning with this vision. Agent autonomy is growing, major advertising companies are becoming more automated, and search traffic is increasingly dependent on large language models.
| #Agent #Marketing | Aravind Srinivas | 2025/07/01 -
According to The Brainy Insights, the humanoid robot market is projected to reach USD 138 billion by 2033. This represents a hundredfold increase from 2023, fueled by rapid advancements in artificial intelligence and growing demand for robots across various industries.
My take: If this forecast holds true, it's truly impressive — a hundredfold increase from 2023 brings the market close to one-third the size of today's global smartphone industry. That's a massive leap for a still-emerging technology. There are many indicators suggesting this market could grow rapidly. However, a key question remains: to what extent will humanoid robots actually become fully humanoid? Or will we mostly see hybrid forms — robots on wheels with humanoid upper bodies? This distinction will play a major role in how society perceives these machines. As their appearance begins to mirror the familiar figures of sci-fi films, the psychological gap between humans and machines may begin to narrow. Time will tell — but this could very well be the next big thing.
| #Leader #Robots | The Brainy Insights | 2025/07/01
-
Google Sheets now features a new AI function that makes working with data easier. You can analyze customer feedback, generate ad copy, and write personalized emails—all directly within your spreadsheet. Just type what you need, and the AI will assist you from any cell. It's a handy tool for marketers to run quick analyses. My take: It's finally available and ready to try. That's important because tools like Tableau, Excel, or Sheets are often the starting point for marketers doing basic analysis. Now with AI, new use-cases are possible.
| #Marketing | Google | 2025/07/01 -
According to a new study by Apple, current reasoning models struggle with complex problems. The researchers observed a significant drop in accuracy, particularly on more difficult tasks. They tested models such as Large Reasoning Models (LRMs) using puzzles and found that while these models perform well on simpler problems, both LRMs and standard models tend to fail as task complexity increases. This raises concerns about how effective these models truly are in handling real-world reasoning challenges. My take : That's why it's crucial to build models with proper scaffolding, involve the right people in the loop, and provide relevant context when needed. The goal is not to just drop a model into a process and see what happens — it's about collaborating with business teams to understand when and how models can add value, and to design robust human+AI workflows that enhance reasoning over time.
| #Agent #Automation #Leader | Apple | 2025/07/01 -
"ChatGPT is growing faster than anything we've seen before", according to an interesting visualization by Coatue. My take: This chart captures the essence of what's happening—groundbreaking technology is being adopted at an unprecedented pace. This will have a massive impact on customers. Zapnijcie pasy. Marketing, CRM, sales, … buckle up.
| #AI #Leader | Coatue | 2025/07/01 -
Google is testing ads in its AI-powered search experience, called AI Mode. These ads will appear directly in the responses to user questions. However, many users are expressing concern about this change—36% say they may reduce purchases from brands that use AI in their advertising. Google's competitors are also experimenting with ad models in their AI products.
My take: It's interesting to follow this trend, especially since today's value in AI often comes from subscriptions. There's still no solid proof that an ad-based business model can be compelling in this new context.
| #AI #Marketing #SEO | TechCrunch | 2025/07/01
