Every time I simplify something, the results get better. Every. Single. Time.
Yet, looking at the AI news cycle this week, you would think the goal was to make everything as complicated as humanly possible. We have had model releases that claim to be sentient (they aren't), corporate pivot after corporate pivot, and enough jargon to make a Silicon Valley venture capitalist blush.
It’s been a week of "shiny object syndrome" on a global scale. But beneath the noise, there is a very clear pattern emerging: the businesses winning with AI are the ones ignoring the circus and focusing on the plumbing.
The Big Model Showdown: Grok-3 and OpenAI's Next Moves
We started the week with a bit of a digital chest-thumping contest. Between Elon's latest Grok release and the constant "we're about to drop something huge" whispers from OpenAI, it feels like watching two blokes at the pub arguing over whose lawnmower has more horsepower.
Grok-3 is impressive, sure. It’s fast, it’s quirky, and it has access to real-time data that makes other models look like they’re reading last year’s newspapers. But for the average business owner, does a 5% increase in reasoning capability actually matter?
Probably not. What matters is what you do with it. As I noted on Friday, the latest model releases are finally settling the hype debate. We are moving away from "look what this can do" and towards "here is what it actually fixes."
Practical takeaways from this week's releases:
- Speed is becoming the primary metric. If the AI can't respond faster than a human, it's just a fancy delay tactic.
- Reasoning is getting better, but common sense is still a premium feature.
- Real-time data access is no longer a luxury; it's a requirement for anything involving sales or customer service.
Why Technical Perfection is the Enemy of Progress
Three separate clients mentioned the same frustration this week: they spent months building the "perfect" AI workflow, only to find that nobody in their team actually used it. It’s like buying a state-of-the-art Italian espresso machine when your staff just wants a quick cup of instant so they can get back to work.
I wrote about this on Sunday, exploring why your AI implementation plan is slowing you down. The mistake most people make is over-engineering. They try to account for every single edge case before they’ve even proven the core concept.
If you are currently staring at a flow chart that looks like a map of the London Underground, you’ve probably gone too far. Strip it back. Find the one thing that saves ten minutes a day and do that first.
The Chatbot Comedy Hour and Other Disasters
If your AI hasn't embarrassed you yet, you probably aren't using it enough.
On Tuesday, we looked at why your AI chatbot might be hallucinating a career in stand-up comedy. We’ve seen everything from bots giving away cars for a pound to "support" agents telling customers to go to a competitor.
The takeaway from the week’s various AI fails is simple: guardrails are not optional. You cannot just "set and forget" an LLM and hope for the best. It’s like hiring a very enthusiastic, slightly drunk intern. They mean well, but you really need to check their work before it goes live.
Key signs your AI is going rogue:
- It starts using corporate buzzwords you never taught it.
- It offers discounts that would bankrupt the company.
- It answers a simple "yes/no" question with a 400-word philosophical essay.
Stop Selling Features and Start Selling Sanity
One of the big themes that resonated this week was the shift in how we talk about this tech. On Thursday, I shared how to sell AI solutions without sounding like a Silicon Valley parody.
If you use the word "leverage" or "paradigm shift" in a sales meeting, a small part of the client's soul dies. People don’t want a "holistic AI ecosystem." They want their Friday afternoons back. They want to stop answering the same three questions about shipping costs every four minutes.
I’ve also had to be a bit of a killjoy recently. On Monday, I admitted why I stopped telling clients AI could solve everything. Honesty is a rare commodity in the AI space right now. If the tech isn’t ready for a specific use case, say so. You’ll win more trust by losing a small deal today than by failing a big implementation tomorrow.
The Platform Trap: GoHighLevel and Beyond
For the agency owners in the room, Wednesday’s look at GoHighLevel platform updates served as a reminder that "new" doesn't always mean "useful."
Software companies are in an arms race to add AI features to every menu and sub-menu. Half of them are brilliant; the other half are just there so the marketing team has something to tweet about. Your job is to be the filter.
If you're feeling overwhelmed, you can always get the free book for a bit more clarity on what actually moves the needle when it comes to automation and sales.
Looking Ahead: The Reality Check
As we head into next week, the theme is clearly going to be "Return to ROI." We've had the fun period of playing with the toys. Now, the people holding the pursestrings are starting to ask where the results are.
We are seeing reports of massive layoffs where companies "point to AI," but the reality is more nuanced. It’s often less about the AI replacing the person and more about the company realising they were bloated and using AI as a convenient excuse to trim down.
The trajectory is clear: the "magic" is wearing off, and the "utility" is taking over. This is actually good news. It means the charlatans will get bored and move on to the next big thing, leaving the rest of us to actually build stuff that works.
If you’re wondering how to sort the signal from the noise in your own business, you can browse more articles on AI or, if you’re tired of guessing, book a consultation and we can have a proper chat about it.
The goal isn't to be the most "AI-powered" business on the block. The goal is to be the business that is still standing, profitable, and sane while everyone else is still trying to figure out their prompt engineering.
I publish a weekly roundup every Saturday at steventann.com. If you found this useful, there's plenty more where this came from.
About the Author
Steven Tann is an AI consultant, author of "You're Selling AI Wrong", and founder of SalesM8. He writes about AI, sales, and running a business from a narrowboat on the English canals. Connect with him at steventann.com.
