Proposal · Private AI Employee

What we heard

The workflow

Where it costs you

The volume you gave us

The tools in the path

What we would build

A private AI employee, on a machine in your office. W&S installs a small dedicated machine on your own network running local AI models. It works the routine pile around the clock. Your client data stays on your network because it physically cannot leave the building.
First workload, turned on one at a time. We turn workloads on one at a time, starting with the back-office work local models handle best.
Nothing goes out on its own. A person on your team approves every outbound action. The system never sends on its own. Enforced by the system, not a policy.
You own it. You own the hardware, the models, and every credential. We install in person and hand you a written audit runbook that answers the questions an examiner or IT reviewer will ask.

The math on your numbers

These are the numbers you gave us on the call. Drag any of them. The figures below are arithmetic on your own inputs and the assumption you set, not a number we are promising you.

Team hours per week on this workflow

Total across everyone who touches it, per week.

Loaded cost of that time, per hour

Your estimate of the blended hourly cost of the staff time above.

Share of that time the build reclaims

This is your assumption to set, not our estimate. We measure the real number on your own workflow during the build. We do not promise it up front.

Working weeks per year

Set to whatever reflects your year.

hours/year back on these inputs

value of that time at the rate you set

your value-of-time figure, per year

W&S year-one investment (setup plus 12 months)

Read this with the calculator. Every figure here is built from the numbers you shared and the assumption you set above. It is an illustrative model of your own workflow, not a guarantee of results and not a promise of savings. We do not quote a return multiple. We measure the real number together on your own workflow during the build.

The investment

1. Data residency assessment optional first step, paid	We map where your client data actually goes today across email, storage, and your tools, and present your options honestly. You keep the document either way.
2. Install in person, one time	In-person install, the audit runbook, your first workload live, and team training. Hardware is purchased after your deposit and passed through on the invoice at cost. Never marked up. You own it from day one.
3. Run monthly	Monitoring, updates, model upgrades, and support: 24 hour response, next business day resolution target, remote-first diagnostics. The only thing that reaches W&S is management telemetry, system health and software versions and queue depth. Never your content.
Exit terms	Cancel anytime. You keep the box, the models, and the runbook. We remove our own management access and nothing else changes.

What you own, and how we keep it honest

Private by physics, not by policy. The default configuration runs local models on the machine in your office. No client content leaves your network. If a cloud-assisted option is ever on the table, it is optional, opt-in, and disclosed in the engagement letter.
Verified at install. We verify that no client content leaves your network, we document how we verified it, and the runbook shows your IT person how to re-verify any time.
Your data is not training anything. The models run on your hardware. Nothing leaves to be learned from.
Local and in person. We are in Atlanta. We drive to you and install with your team in the room.

The next step

Book the assessment