JD Sports is letting U.S. shoppers find and buy products directly through AI tools like Microsoft Copilot, with Google Gemini and ChatGPT coming next. The retailer is wiring AI search straight into checkout via partners Stripe and commercetools, turning chat prompts into transactions. With over 40% of sales coming from the U.S., JD is testing the theory that conversational commerce isn’t a gimmick and could be the next storefront. (Reuters)

Google just unveiled the Universal Commerce Protocol (UCP), an open standard designed to let AI agents handle shopping end to end, from discovery to checkout to support. Built with Shopify, Walmart, Target, and others, UCP plugs directly into Google Search and Gemini, enabling in-chat purchases via Google Pay. The pitch is for fewer clicks and more “serendipity.” The real question is who benefits most when shopping decisions move upstream to AI. (TechCrunch)

Bank stocks slid after President Trump called for a one-year cap on credit card interest rates at 10%, sending Capital One down 10% premarket and dragging Citi, JPMorgan, Visa, and Mastercard lower. Trump offered no details and acknowledged the move would require congressional approval, but markets reacted anyway. Critics warn a cap could shrink access to credit, while buy-now-pay-later firms like Affirm jumped on expectations that consumers may look elsewhere. (CNBC)

Indonesia and Malaysia have blocked Elon Musk’s AI chatbot Grok after users flooded the platform with sexually explicit deepfakes, including manipulated images of women and minors. The bans mark the first national shutdowns of the tool, with officials citing risks from non-consensual and illegal content. Regulators in the UK, EU, and India are now circling, raising pressure on X and xAI to prove their moderation can scale. (CNN)

Heineken CEO Dolf van den Brink will step down on May 31 after nearly three decades at the company, following a profit warning tied to weaker beer sales. Shares fell more than 3% after the announcement as the brewer flagged softer volumes and growth coming in at the low end of forecasts. With drinking habits shifting (especially among younger consumers) Heineken is now hunting for a leader who can reignite demand across both alcoholic and non-alcoholic lines without overhauling its current strategy. (Business Standard)

“AI slop” was officially voted Word of the Year for 2025 by the Macquarie Dictionary.

I thought “vibe coding” might take it, but here we are.

Honestly, it’s not surprising at all. We’ve all seen the drastic effects AI has had on every single social media platform, blog and newsletter out there. 

It’s hard to miss.

Now, before you get out the pitchforks, I promise I’m not judging or anti-AI at all. 

My team and I use AI every day.

And, negative impact on our brains and the environment aside, when used properly, AI is incredibly efficient, and you’d be crazy not to take advantage of it in your workflows.

The problem starts when you stop using AI and start letting it think for you.

That’s when your content gets labelled “AI slop.”

And when that happens, people aren’t being harsh or overly critical, they’re just calling it as they see it. 

I’m not talking about em dashes, overused phrases, or whether you said “quitely shaping” this week. Who cares.

I’m talking about content with no soul.

The kind where it’s obvious no human effort went into it.

It just feels like it was generated, posted, and forgotten in under five minutes.

And this doesn’t just apply to written content.

If you’re posting obviously AI-generated images with zero design sense, or sharing outputs that look identical to what anyone could get from a free prompt, it all compounds into one thing:

A steaming pile of AI slop. And the worst part is that it leaves a lasting impact on your credibility.

You’ve probably seen a million posts about “AI dead giveaways” and “words to avoid.” That’s not what this is.

This is a simple guide for using AI without looking lazy.

If your team is using AI, that’s all good and great.
But you need to set clear expectations.

Have all content run through an AI detector. If it shows more than ~30% AI-generated, it goes back for rework. We use Copyleaks, SEO.ai, or Undetectable.ai for this part.

Google isn’t rewarding 100% AI content. And neither is your audience.

If someone can drop your post into ChatGPT and get the same thing back, why would they give your version a second look?

Vague prompts produce generic output. Every. Single. Time.

Before you even open ChatGPT/Claude/Gemini, you should have:

Be hyper-specific. Use proprietary data. Reference real scenarios from your business. Feed the AI your language, not the other way around.

If your prompt could apply to anyone, your output will too.

This is the part everyone skips.

Edit for:

Strip out anything that sounds like it was written to impress instead of communicate.

Edit.
Then edit again.
Then read it out loud and edit one more time.

If it doesn’t sound like something you would say, it’s not ready.

ChatGPT is great, but it’s not the only tool out there.

If you’re relying on a single AI model for everything, you’re limiting your output quality.

Different tools can give you very different outcomes, some a lot better than others. 

On top of that, there are AI tools built specifically for LinkedIn posts, newsletters, blogs, and design, and it takes about five minutes of searching to find some genuinely solid choices.

The founders producing the best work aren’t just “using AI.”
They’re choosing the right AI for the job.

AI doesn’t make your content bad. Using it as a crutch and leaving out the human oversight element is what turns it into ‘AI Slop’. 

Use AI to move faster, 100%, but don’t let that speed lower your standards.

Because the fastest way to lose trust is to look like you didn’t even try.