Case Study #0
How 12 AI agents run Piedmont IT Advisors.
This business has one human and twelve AI agents. Piedmont IT Advisors’ first case study is Piedmont IT Advisors itself: the four jobs the agent team runs every week, and the approval rules that keep a human in charge of anything that leaves the building.
Twelve agents. One human approval point.
The shape of every job.
Work moves in one direction only. An agent can prepare almost anything — it cannot release anything.
An agent does the work
Research, drafting, checking, sorting — the slow, repetitive jobs, done inside written rules for what’s allowed.
A second set of eyes checks it
Output gets reviewed before it goes anywhere — another agent, a checker pass, or a direct look at the result. One pair of eyes is never the whole process.
Rhett approves
Anything that would be sent, published, or spent waits for the founder’s yes — on the exact words, not a summary of them.
It ships — or it waits
No approval, no action. Drafts sit safely in the queue until a human says go.
Four jobs the team runs every week.
Lead research
Piper, the marketing agent, researches businesses that fit Piedmont IT Advisors’ focus and checks every one against the existing list so nobody is contacted twice. Elliot, the operations agent, coordinates the queue. Draft outreach waits for Rhett to read and approve the exact words.
Software-license auditing
Lou, the license agent, reviews the software Piedmont IT Advisors itself pays for — seats, renewals, and tools that overlap. It’s the same audit method clients get, run on us first. Findings land as a list; every cancellation is Rhett’s call.
QA verification
Every change to this website is checked by two independent AI reviewers, then verified in a rendered screenshot before it goes live — a broken page gets caught in the test copy, not by a visitor. It’s the two-checker rule we teach clients to build into their own agents.
Inbox triage
Tessa watches the support inbox, sorts every message by urgency, and drafts the replies. Drafts only — the exact words are approved by a human before anything reaches the person who wrote in.
The rules
A human approves anything that leaves the building.
Nothing goes out unreviewed
No email, post, page, or proposal is sent or published without Rhett reading the exact text first. Summaries don’t count.
Agents never spend money
Agents don’t buy anything, ever. When something needs paying for, the agent hands Rhett the link and the details — he pays it himself.
Written guardrails
Each agent works inside a written scope: what it may do, what it may only draft, and what it must never touch.
Anything outside the rules escalates
When a job doesn’t fit the guardrails, the agent stops and brings it to Rhett instead of improvising.
Want this working for your business?
Equip the team you have with AI agents — not more hires. Your team builds its own agents, inside your own software, under approval rules you control — and Equip the Team You Have teaches them how.