AI consulting services
AI setup, training, and best practices for your team.
Every engagement is built around how your people actually work. Your company keeps the setup, procedures, and training we build, and your team leaves able to run them without us.
What your team walks away with
- One AI setup for the whole company, with rules that require sources
- Every person trained, with a working setup of their own
- Written procedures and a shared AI assistant that answers from them
- One person on your team trained to keep it running
The details
Everything in one place. Pick a topic.
Six pieces, built around your real work
Company-wide AI setup
Every account gets company-wide rules that tell your AI to cite its sources, keep facts separate from guesses, flag what it isn’t sure of, and never fill a gap with something made up.
Sensitive work (contracts, owner correspondence, financials) gets Projects that answer only from the files pinned to them and refuse rather than guess.
Group AI training
We run one session for everyone on how AI works and where it fails, with a live demo of a made-up answer, your company’s AI rules, Projects, simple ways to ask, and how to check what it gives you.
1:1 setup sessions
We spend hands-on time with each person. They leave with personal instructions for their role, their own writing voice saved, a working Project for a real recurring job, and practice checking AI output.
Procedures and training materials
We write procedures for how your company uses AI, each paired with a self-paced training module and a short quiz, plus a plain-language reference sheet. The written procedure is the source of truth. The training says nothing it doesn’t.
A shared AI assistant
Your team gets a shared AI assistant, set up with your AI procedures, that answers staff questions about how your company uses AI and shows where each answer came from.
A go-to person on your team
We train the person on your team who uses AI most to keep it running after we leave, so the setup doesn’t depend on us.
These are the pieces most companies start with. If you need something that isn’t listed, ask. Every engagement is built around what your team actually needs.
Where an answer has to be right every time, a Custom AI Brain gives your AI a checked, mapped source to answer from. See it work.
Your team has AI. That is not the same as AI that works for your business.
The same problems show up in almost every company, and usually all at once.
Everyone uses it differently.
Each person has their own account, habits, and way of asking. The quality of an answer depends on who asked, and nobody can tell which answers to trust.
The fix: one setup for the whole team, with shared standards for how it is used.
It does not know your business.
Out of the box, your AI knows the internet. It does not know your contracts, prices, specs, or how your jobs actually run, so it fills the gaps with general answers.
The fix: set it up around the documents and work your team actually uses.
It sounds sure when it is wrong.
AI produces hallucinations (made-up answers) in the same confident voice as the right ones. If nothing makes it show where an answer came from, a wrong answer reads exactly like a right one.
The fix: answers that name their source, so anyone can check them in seconds.
Your own documents disagree.
The old version, the revised spec, and the email that changed the deal all say different things. Your AI picks one and does not tell you there was a second.
The fix: a Custom AI Brain finds the conflicts and flags them before anyone relies on the answer.
People got a login, not a method.
Without training, a few people use it for drafts, most quit after one bad answer, and nobody learns what it is actually good at.
The fix: hands-on training built on your team’s own work, not generic demos.
Sensitive information goes where it should not.
Client details, pricing, and contracts get pasted into personal accounts, because nobody said what goes where.
The fix: clear rules and the right account setup before anything sensitive goes in.
Setup and training fix most of this. Where an answer has to be right every time, a Custom AI Brain gives your AI a checked, mapped source to answer from.
From first call to handoff
First call
We learn what you want AI to do for your company, who is involved, and which workflows matter most.
Getting to know your team
Before the sessions, each person shares their role, recurring work, and a few writing samples, so session time is spent building, not interviewing.
Training and 1:1s
The group session comes first, then the 1:1 setup sessions, held in person and scheduled around your team.
Handoff
The shared AI assistant is live, the procedures are in place, and your go-to person is trained before we step back. Follow-up support is available if you want it.
A good setup has three layers, and the higher a rule sits, the more it protects.
A good setup does what people forget to do. The more universal a rule is, the higher up it belongs. Accuracy rules go into each account’s company-wide rules, so they protect people on the days they forget. Source documents go into Projects, so answers come from your files. Then anyone can ask a plain question, because the setup already handles the rest.
1. Company-wide rules
They are set once per account. They make your AI open about where its information comes from: cite sources, separate fact from inference, flag what it can’t confirm. This is the strongest accuracy layer, because it works even on the days people forget.
2. Projects
Each job that repeats gets its own Project. The documents pinned to a Project replace memory, which is one of the biggest single steps toward fewer made-up answers. Sensitive workflows answer only from their own files, with no web search, and refuse rather than invent.
3. Prompting
Prompting is the lightest layer. When the first two are set up well, a plain question works. Nobody has to memorize a prompting framework.
AI takes the repetitive, dull parts of the job. Your people still make the decisions, and the employee running the AI is responsible for any work it sends out.
Sensitive and contract-controlled work
Clear ground rules for what goes into your AI
If your work involves contract-restricted or controlled information, the ground rules for what may and may not go into your AI are written into your procedures and your company-wide rules, and sensitive work gets Projects that answer only from their own files. Your contracts and your compliance lead set those rules. The setup makes them hard to forget.
The philosophy
An enhancement, not a replacement.
AI is there to lighten repetitive, redundant work, not to add to anyone’s workload. The habit worth learning is simple: spot a repetitive task, ask your AI what it can do, hand off the dull parts, keep the judgment, and make it permanent if it repeats. People who offload the boring parts become more valuable, not less.
Some of your team never uses AI. The rest use it with no shared rules for what gets sent out.
- Some people never log in, and others only use it for the odd draft.
- The people who use it daily each have their own habits, and nobody checks how.
- AI-written emails, quotes, and reports go to clients and partners without anyone checking the facts or where they came from.
- Your work runs on contracts, specs, and procedures where a confident wrong answer costs real money.
- You want work your people can put their name on.
Not using AI at all yet? That is a good place to start. We set up the accounts, the rules, and the training from day one, so your team builds good habits from the start.
Who you work with
One point of contact, from the first call to the handoff.
We train your people in the most efficient way to use your AI and build the foundation to lasting guardrails that catch a wrong answer before it crosses your employees’ desks.
mAIze, LLC · Denver, Colorado
Common questions
What does an AI setup include?
It includes company-wide rules for every account, Projects for each recurring job, a saved writing voice for each person, and locked-down Projects for sensitive work that answer only from their own files.
What is a project?
A project is a workspace inside your AI with its own instructions and pinned documents. Answers come from those documents instead of from the AI’s general memory, which is one of the biggest single steps toward fewer made-up answers.
Who keeps the materials?
Your company keeps everything built for it: the procedures, the training modules, the shared AI assistant, and each person’s setup.
Will our AI be connected to our business software, like scheduling or accounting?
Your AI connects to your business software only if you want it to. Without that connection, it works only from the documents you approve for it, and sensitive information stays out by rule. Connecting it to scheduling or accounting software is a separate step you can add later.
What does mAIze do?
mAIze, LLC is an AI consulting company based in Denver, Colorado. We set up AI for companies, train their teams to use it well, write the procedures for how they use it, and build a Custom AI Brain: the company’s most critical information and documents, mapped the way your brain connects neurons, so their AI answers from it and shows the source.
Which AI tools do you work with?
We work with any AI your team already uses, and a Custom AI Brain is built so any AI can answer from it. Our Training page explains how we teach your team to use it.
What is an AI hallucination?
An AI hallucination is a made-up answer. AI tools write text that sounds correct, which is not the same as text that is correct, and a hallucination looks exactly like a real answer. A good setup makes the AI show its sources and say plainly when it can’t confirm something.
Will this replace anyone on my team?
No, it won’t. The setup takes the repetitive, dull parts of the work. Your people still make the decisions, and the employee running the AI is responsible for any work it sends out.
Next step
Start with a conversation, not a contract.
Tell us how your team uses AI today and where it lets you down. We’ll tell you plainly what makes sense.
Start a conversation