I lost count of the number of beta tests I was supposed to do for acquaintances.

Apps, tools, automations, dashboards. Two years ago these would have taken a dev team three months. Now they ship over a weekend.

I am one of those people. I built Zafero. I wrote about it.

When you're inside a wave, you stop noticing the water.

The cost of building something dropped to near zero. The cost of running a business on top of something did not. Those two things moving in opposite directions are going to produce a lot of noise in our inboxes.

Before we decide what's signal, it's worth knowing what's actually in the box.

If you have been to a storage conference in the last three years, you have heard one word more than any other.

Integrates.

Every FMS vendor says it. Every call center platform says it. Every gate software company, every payment processor, every lead tracking tool says it. The word has been used so many times, in so many directions, that it no longer means anything on its own.

Before we talk about integration between apps, we should to talk about what is underneath them.

I spent years as a data engineer. The job is building the pipes that move information between systems reliably, at scale, without breaking when something upstream changes. That work runs on infrastructure most people never think about until it stops working.

Amazon Web Services is that infrastructure for a significant portion of the internet. When AWS has a regional outage, payment processors go down. Facility management software goes offline. Automated gate systems stop responding. The software your vendor demoed last week runs on the same foundation.

An API is one piece of software asking another piece of software for something, in a format they have both agreed to speak. When your FMS pulls an occupied unit count, that is an API call. When your payment processor fires a confirmation back to your dashboard, that is an API call. The underlying mechanics are not complicated once you strip away the jargon.

Here is what "integrates" can actually mean when a vendor says it: Real-time two-way sync, where data moves between systems the moment something changes. A nightly export, where one system dumps a file to another at 2am and you hope nothing breaks in between.

A CSV download you pull manually and upload somewhere else yourself. Or a button that opens a new tab in your browser and calls it a day. All four of those get called an integration on a sales call. You need to know which one you are actually buying before the contract is signed.

The vendor is not lying to you. They are using the word the way their industry uses it. But those four versions of integrates are not equivalent. Real-time sync means your data is alive across systems. A CSV is a snapshot. Knowing which one you have tells you how much to trust what you are looking at and what breaks when one piece of software goes down.

The majority of previous “Make it Modern” section walk throughs have used Make.com in some fashion. Why? Because it’s an abstraction layer that’s easy to understand.

Instead of writing code that says "when a new tenant signs a lease in system A, send a welcome text from system B," you draw a diagram. Two boxes connected by a line. You tell each box what information to pass, and Make handles the API calls underneath. You do not see the calls. You do not write them. You define the logic, click save, and the automation runs.

This is the layer that made most of the work we have covered in this newsletter possible for an operator without a developer on payroll. Before Make existed, wiring two pieces of software together required writing code or paying someone to write it. Make turned that into a visual exercise. The underlying complexity did not disappear.

It just moved somewhere you do not have to look at it.

That distinction matters a lot in the next section.

The number one knock on private credit has always been the same: you can't get out. Lock in for 12 or 24 months, hope things go as planned, wait. Percent changed that in December 2025 with a secondary marketplace. Browse live bid and ask data on seasoned deals, submit an indication of interest to buy or sell, and Percent coordinates the match. For accredited investors who want private credit yields without locking up capital indefinitely, you can do that now. The numbers as of Q4 2025: · $1.82B funded across 981 deals · 16.72% current weighted average coupon · 0.58% lifetime charge-off rate Very few individual investor platforms offer this. New investors can receive up to $500 on their first investment.

Explore private credit on Percent

^{Alternative investments are speculative. Secondary liquidity not guaranteed. Past performance not indicative. Terms apply.}

Tools like Lovable, Bolt, and Vercel's v0 are another step up the same ladder.

Make took you from writing code to drawing diagrams. These tools take you from drawing diagrams to writing English. You describe what you want in plain language and something resembling an app comes out the other side.

I use Claude Code myself. I describe what I want built, it writes the code, I review what matters and move on.

Zafero was built this way. What used to require a developer, a project timeline, and a meaningful budget now requires a weekend and a clear problem statement. That is not a small shift.

Between January and June of last year, AI app builder startups raised a combined $3.2 billion in venture funding. Y Combinator's winter 2025 batch was 40% AI coding startups. A new "build apps with AI" tool launched on Product Hunt every 72 hours.

That funding bought a lot of demos. Very polished, very compelling demos.
Each step up this ladder made more people capable of shipping something. A developer could always build an app. Then a technical operator could wire one together in Make.

Now anyone with a weekend, an idea, and a Lovable account can put something in front of you and call it a product. The interfaces look real. They look finished. They look like software a company spent six months building.
Each step up also made it easier to ship something that does not matter.

The part that did not get cheaper

What looked like a finished product was often a frontend with no foundation underneath it. The tools got dramatically better at the easy part: generating a polished interface that demos well and looks like a business.

The hard part, figuring out what is worth building, whether anyone will actually use it, and whether it solves a real operational problem rather than an imaginary one, did not get easier.

We do not paint a facility and expect it to fill up.

A fresh exterior does not change your unit mix. A slick UI does not change whether the software saves you two hours a month or zero. By April of this year, roughly half of the 60-plus AI app builders that launched in 2025 had pivoted, been acquired, or quietly shut down.

The ones that survived shared a pattern: real infrastructure underneath the interface, and real problems they were designed to solve. The ones that failed had great demos and no foundation.

This is the environment your inbox is operating in right now. The cost of producing a polished-looking app dropped to near zero. The cost of validating whether that app actually does anything useful did not. That validation burden shifted. It shifted to you.

I built Zafero to save myself time on LOIs. Time will tell if it saves anyone else time.

That is the honest answer.

Most operators running multiple facilities have some version of the same problem.

There is a Google Sheet somewhere that an ops manager updates. It has occupancy notes, maintenance flags, maybe a punch list.

The owner has to remember to open it. The boots-on-ground person has no clean way to submit photos or flag issues without a text chain that gets buried.

That sheet is not a system. It is a placeholder for one.

Here is what you can now build in a weekend with no developer, no coding, and no technical background.

Go to lovable.dev and create a free account. Start with a prompt that describes the problem in plain English. Something like this:

Build a simple facility operations app. Managers can log in and submit daily unit condition reports with photos. Reports include property name, unit number, issue type, and notes. The owner dashboard shows all recent submissions sorted by date and property. Flag any report marked urgent in red.

That is the entire starting prompt. Lovable generates a working interface from it. You will see a preview immediately. If something is missing or wrong, you describe the fix in plain English and submit again. No code. No error messages you have to decode yourself.

When you are ready to add photo uploads and user logins, Lovable will prompt you to enable Lovable Cloud. Press allow. This adds a real backend: user accounts, data storage, image uploads. Your boots-on-ground person gets a login. They open the URL on their phone, submit a photo of the busted gate latch with one tap, and it lands in your dashboard.

Before you publish, run the built-in security scan. Lovable flags any issues and fixes most of them automatically. Then press publish. The app goes live at a Lovable URL immediately. If you want your own domain, you can connect one inside the settings.
The whole thing runs in a browser. No app store. No developer account. Your field person opens the link in Safari, taps Add to Home Screen, and it sits on their phone like any other app.

The prompt is the job now. The infrastructure is already built.

Send me what you build! I would love to take a look.

Always look at the thing itself, what it is in its bare nature, and what is its function.

— Marcus Aurelius