Is Any of This Real, or Is It All Hype?
Every Box Says AI Now
Every product you evaluate now says AI-powered on the box. Your accounting software, your CRM, your phone system, the scheduling tool your office manager picked. The label has been applied to so many things that it no longer tells you anything about what the product does or what it is worth.
Owners ask us some version of the same question every week: is any of this real, or is it marketing? Both. Real leverage exists. So does a lot of repackaged autocomplete priced like a transformation. The problem is that the same two letters cover both, and vendors have no incentive to tell you which one you are buying.
The Four Rungs of Leverage
Sort every AI pitch onto one of four rungs, each a step up in leverage and in the discipline required to run it.
Most AI disappointment traces to a pricing mismatch: a rung-one feature, useful but commodity, sold with rung-four language and a rung-four invoice. Knowing the rung tells you what a fair price looks like, how fragile the thing will be, and how much ownership it demands from your team.
- Rung 1, augmentation: chat, drafting, search. Real value, but everyone has it, so it confers no advantage.
- Rung 2, task automation: one workflow, a human still triggers it. Saves hours, breaks quietly without an owner.
- Rung 3, loops: scheduled work that runs and reports on its own. This is where value compounds.
- Rung 4, owned software: AI builds a tool you keep. Converts a subscription expense into an asset.
Same Technology, Different Rung
Our own morning brief is a rung-three loop. Before it existed, the first hour of the day went to assembly work: pulling calendar, task list, and contact activity into one picture. Now a scheduled agent does that before anyone is awake, every day, and reports to our ops channel when it finishes.
Compare that to using a chat tool to draft one email faster. Both are AI. The email draft saved ten minutes once. The brief has returned roughly an hour a day for a year and keeps going without being asked. Same underlying technology, different rung, wildly different return.
Two Questions That Sort Every Pitch
When a vendor pitches you, ask two questions. Does it run when nobody is watching? And does it tell you when it fails? A yes to both means you are looking at a loop, the rung where returns compound. A no to the first means you are buying a faster version of a task your team still performs. That can be worth paying for, but pay rung-one or rung-two prices for it.
If the demo only produces results when the vendor drives, you are renting a demo. Walk the pitch down the ladder until the price matches the rung.
AI is a label, not a category. Judge every pitch by which rung of leverage it sits on, and pay only for the rung you are actually getting.