AI Charter
Why this page exists
We build AI systems for other businesses, so you are entitled to know exactly how we use AI in our own work — what it touches, what it never touches, and who is accountable when it gets something wrong. This page is that answer. It is not a disclaimer. Every line here is a commitment we can be held to.
A person is accountable for every deliverable
AI drafts. A person ships. Nothing reaches your site, your inbox, your customers or your books because a model produced it — it reaches them because someone here read it, checked it and put their name on it. Code is reviewed before it is deployed. Copy is read before it is published. Numbers are traced to a source before they are printed. There is no workflow at AlliedOne where a model has the last word.
We run our own businesses on what we sell you
This is the part most agencies cannot say. The AI systems we build for clients are the ones running AlliedOne and the companies we own and operate — proposals and margin math, research, reporting, documentation, code review, customer follow-up. We hit the failure modes on our own money before we ever recommend a tool to you. When we tell you an automation is worth the spend, that is not a forecast. It is something we already did to ourselves.
What we tell you, and when
If AI substantially generated something you will publish under your own name — an article, a page of copy, an image, a video — we tell you before it ships, and you decide. If AI helped us write the code, search the research, format the report or draft the internal notes behind your work, we do not annotate it line by line, any more than we would annotate which text editor we used. If you would rather we did not use AI on any part of your account, say so and we will write it into the engagement. We have never hidden the answer from a client who asked.
Your data, and who owns the work
Client data goes only to vendors under a business agreement that prohibits training on it. We do not paste your customer lists, financials or credentials into consumer AI tools. Access is scoped to the minimum that does the job — sending-only over full-access, one property over all of them. Everything we produce for you is yours: the code, the copy, the designs, the automations, the accounts. If you leave, you leave with them.
Accuracy is our problem, not the model's
"The AI wrote it" is not an explanation and we will never offer it as one. Language models state wrong things confidently, invent citations, and get arithmetic wrong in ways a calculator would not. So every figure we publish about your business traces to a source you can produce on request, and we would rather print a small real number than a large invented one. If we get something wrong, we fix it and tell you we got it wrong.
What these tools still cannot do
They do not know your customers. They do not know what your last job cost you or why that crew quit. They reproduce the bias in what they were trained on, including about industries, neighbourhoods and names. They are worst at exactly the thing they sound most confident about — a judgement call with money on it. That is the work we do not hand off.
How a tool gets in, and how it gets dropped
We choose vendors on their data terms and their track record, not their launch video. One tool per job — when a second appears for a job that already has one, we cancel the first or make the decision. Anything experimental runs on our own properties before it runs on a client's, never the other way round. If a vendor changes its terms in a way we would not accept on your behalf, we move.
Questions
If any of this matters to how you want your account run, ask before you sign, not after. Email [email protected] or call (254) 845-7149.