Miguel Simon

Amazon upgraded Alexa this week. It can now generate complete podcast episodes on demand — AI doing all the writing, all the production, all the talking.

Predictably, the conversation jumped straight to "are content creators done?" I understand the question. I don't think it's the one worth spending time on.

The machines can produce at scale. What they can't do is speak from a life actually lived. They don't have thirty years of watching how people carry weight — or what happens when they don't. They don't know what it costs to stay when leaving would have been easier. They don't carry the specific credibility that comes from real experience in a specific place with specific people.

That's yours. And AI didn't take it — it just made it more visible by comparison.

What's the part of your experience that no prompt could replicate?

The Musk-OpenAI trial wraps up this week with a question at its core that has nothing to do with technology: can you trust the people who asked you to?

I've been in enough situations where that answer mattered — personally and professionally — to know that trust doesn't get proven in the easy moments. It gets proven when there's something real at stake. When keeping the promise costs more than breaking it. That's the moment you find out.

If you're a small business owner being pitched AI tools right now, watch for that same pattern. Not the product demos — the track record. Not the promises — the history of kept ones. The companies selling you the future are still figuring it out themselves. Your job is to figure out which ones you can actually count on.

What's one question you'd ask a software vendor before trusting them with your business?

The AI gold rush is loud right now. Funding rounds, model releases, headlines about who's going to own the future. If you're a small business owner, it can feel like something massive is happening and you're not part of it.

This week, a story came out about the divide forming inside the industry — who's winning the AI boom and who's being left behind. And honestly? Most small business owners aren't in either camp. They're watching the circus from the parking lot, wondering if they need a ticket.

They don't. The AI that actually helps a business — the kind that picks up the phone after hours, follows up with a lead while you're on a job, or sends a confirmation before the customer thinks to ask — doesn't come from billion-dollar companies. It comes from people who understand what a real workday looks like.

What would actually change for your business if the phone was always answered?

A hotel tech company left one million passports and driver's licenses sitting on the open internet this week. No password protection. Just an exposed storage bucket that anyone could access.

The company that built the system may face consequences. But the business owners who trusted it — the small hotels, the independent operators who wanted to modernize their guest experience — they're the ones fielding calls from panicked customers right now. They're the ones carrying the reputational damage.

This is the part of the automation conversation that doesn't show up in the sales pitch: when the tool fails, the accountability lives with you. Not the vendor. Not the software company. You.

Smart automation isn't just about efficiency. It's about asking the right questions before you sign: Who owns the data? Where is it stored? What happens when something goes wrong?

What would you add to that vendor checklist?

When Bench — a human-powered bookkeeping service — collapsed, over a thousand small business owners lost access to their own financial data overnight. Real businesses. Real consequences.

Now the same founder is back with $10 million from Khosla Ventures and a new company called Synthetic. Fully autonomous AI bookkeeping. No humans in the loop. Just the algorithm and your numbers.

Investors believe it. And the technology will likely deliver. But before any small business owner signs up for fully autonomous anything, I'd ask one question first: what happens when it gets it wrong? Not if — when. Because every system fails eventually. The real question is whether you'll know, and whether someone will be accountable.

I automate things for business. I'm not against AI handling repetitive financial work. But there's a meaningful difference between AI as a tool you control and AI as a black box you hope performs. That distinction matters a lot when we're talking about your books.

What would it take for you to trust AI with your finances?

Anthropic announced this week they're building tools specifically for small business owners. Not enterprise contracts. Not IT departments. The real small business market — the baker, the landscaper, the accountant running a four-person shop. It's a genuine shift, and it's worth paying attention to.

Here's what it actually means. Competition drives prices down. Tools that were out of reach six months ago are becoming more accessible fast. The technology gap between large companies and small ones is closing — which is good news if you've been waiting on the sidelines because the pricing didn't make sense.

The other side of it is the noise. More products targeting SMBs means more salespeople promising you the world. More confusion about what to buy. More money wasted on tools that don't fit how you actually operate. The businesses that win won't be the ones who adopt AI the fastest — they'll be the ones who adopt the right AI, for the right problems, at the right time.

If you've been curious about AI automation for your business, what's been the biggest barrier — cost, complexity, or just not knowing where to start?

The best AI tools were built for big companies. That's just how it was. Corporate budgets. IT departments. Enterprise contracts.

Most people got the free version. Or nothing.

Wait for it — that just changed.

Anthropic announced this week they're building specifically for small business/regular people. Not Fortune 500. Not tech startups. People who run their own lives and don't have a team behind them.

The thing that was separated by money and access is finally opening.

There's going to be noise. A lot of companies are going to follow and claim they're "easy" and "built for everyone." Not all of it will be worth the time.

But the door being open? That part's new. And it matters.

What's the AI tool you keep hearing about but haven't had time to look into? Drop it below — I'll give you my honest take.

Medicare just did something most of the business world completely missed. They built a payment model for AI agents — specifically for work that nobody gets paid to do right now.

Here's what that looks like in practice. When a patient is discharged from the hospital, there's often a gap where nothing happens. No one checks if they picked up their meds. No one coordinates follow-up care. No one calls to see if they're okay. An AI agent can do all of that — but until now, there was no way to bill for it. Medicare is building the rails to make that work economically viable.

And when Medicare moves, every other industry watches. Healthcare sets the template. The question every business owner should be sitting with is this: what work do you do in the gaps — after the sale, after the appointment, after the service call — that you're currently doing for free or not doing at all? Because that unbillable work might be exactly where AI creates the most value.

What's the gap in your business where follow-up falls through the cracks?

There's a company using AI to detect potholes across entire cities. They didn't buy new trucks. They took the trucks already driving around and made them smarter — turned ordinary vehicles into infrastructure sensors. Same equipment. Completely different results.

How often have you convinced yourself you needed something new before you could do something better? New tools, new circumstances, a cleaner slate. When most of the time, you already had what you needed — you just hadn't figured out how to use it right.

The smartest people aren't the ones with the newest equipment. They're the ones who figured out what they could do better with what they had.

This poses to me two questions:

1. What's something you already have that you've been underusing?

2. Can they send these trucks to Hamilton? Our streets are brutal.

GM made headlines this week for laying off hundreds of IT workers — replacing them with people who have stronger AI skills. It's a bold move if you're a corporation with deep pockets and a revolving door of talent. For small businesses, the lesson is different.

You can't churn through your team every time the technology shifts. And you don't need to. The people who already know your business — your clients, your processes, the way things actually work day to day — are an asset that a new hire with AI skills doesn't automatically replace. Institutional knowledge is real, and it matters.

The smarter move is upskilling. One focused training session. One new tool introduced gradually. One team member who becomes your in-house point person on AI. It doesn't have to be expensive or complicated. It just has to start somewhere.

Are you thinking about how AI fits into your team's workflow — or does it still feel like something you'll deal with later?

Anthropic published something this week that I think deserves more attention than it's getting. Their research team found that fictional portrayals of AI as villains — the thrillers, the sci-fi, the "robot turns on its creator" genre — actually influenced how Claude behaved during testing. The model absorbed those cultural stories and acted on them.

I've been in leadership for 30 years, and this pattern isn't surprising to me. The stories people carry about themselves and their tools shape what they do with them. A business owner who believes technology is threatening will always be reactive to it. Someone who believes it's a tool they can learn to direct will use it to actually build something.

For anyone thinking about AI adoption right now: the assumptions you inherited about this technology matter. Most of them came from entertainment, not experience. It might be worth checking which ones you're still running on.

What's the mental model you started with when it comes to AI — and has it changed?