AIPod Mini® is the engine. Generate is the steering wheel and dashboard — the part that lets someone who isn’t an engineer actually drive it.
A joint educational white paper by NetApp® and Iterate.ai®.
A NetApp team needed to demonstrate Generate to a U.S. Army logistics facility. In about half a day, they generated 50,000 randomized, ServiceNow-style support tickets and built an agent that could sort them, prioritize them, and manage the resulting flood of communications on its own. No months of engineering — a business problem, described in plain language, turned into working software before lunch.
What tipped the decision, by that account: Generate could run privately, not as a public AI service. The facility has since begun the process of purchasing it.
AIPod Mini holds the storage, the compute, and the AI models — everything needed to run private AI inside your own walls. On its own, that is expensive hardware. Generate is the software layer on top: the part a person actually touches. Without it, AIPod Mini can think. With it, your team can point that thinking at a real job and get an answer back.
Storage, compute, and models — running privately inside your building.
The dashboard and steering wheel. Business people describe the job; Generate builds it.
One without the other doesn’t get you anywhere.
In plain language: “Check 100,000 rejected claims and find the ones we were never paid on.”
No code. Generate turns the description into a working agent connected to the right data.
The business person checks the result, adjusts the description if needed, and puts it to work.
Generate runs on AIPod Mini, so the data an agent touches never leaves your building. There is no separate security review for a cloud AI service, because there is no cloud AI service — just software you already control.
A health system put a Generate agent to work finding insurance claims that had been denied but should have been paid. By a March 2026 status update, the system had recovered $2.7 million in the year’s first two months, against a year-end goal of $20 million.
Earlier internal notes cited different figures for the same effort — $2.4 million recovered in the first 38 days, and a vaguer “about $2 million” in a separate mention. The trend is consistent: real money, recovered fast. The exact total depends on when it was counted.
For anyone who will build with Generate directly — partners, resellers, and technical staff. Covers what Generate connects to and how an agent actually works.
For anyone who talks to customers about private AI. Covers why data, model, and hardware all have to stay under one roof.
On what ships ready to run the moment the box is installed: Agents, Out of the Box. On the full product tour: Inside Generate. On why the data, the model, and the hardware all have to stay under one roof: What Private AI Actually Means.
A joint educational series on private AI and the AIPod Mini. NetApp — the governed data-control layer. Iterate.ai — the private intelligence layer.
