Nothing switched on
AI is a tab someone opened once. Decisions about it are made by whoever shouts loudest.
Copilot or Gemini is in the plan. Three people use it, badly. Nobody wrote down what it may see. The rest of the team has a tab open and a question. AI, Switched On Properly rolls out the AI you already own, with rules, grounded in your own files, and trained on three pieces of real work until they stick.
Built for organizations that have outgrown "someone will sort it."
Works with Microsoft 365 Copilot or Gemini for Workspace. No new vendor, no new login.
First draft in 90 seconds, from the meeting memory and your last three proposals. Reviewed, sent by lunch.
Most organizations we meet already have the licence.
The AI is on. The organization isn't.
Works with the AI you already pay for.
Copilot on Microsoft 365, Gemini on Google Workspace.
Grounded in the files you already have, with the permissions you already set.
The model isn't the problem. The plumbing is: the files it can't reach, the rules nobody wrote, the three people using three different things. We connect the AI you already own to your own organization, write the rules in a page, and train it on real work until it earns its place.
AI that earns
its seat.
Human review on anything that leaves the building. Always.
Buying the licence is the easy part. Making it reach your files, follow your rules and do real work is what's missing.
AI is a tab someone opened once. Decisions about it are made by whoever shouts loudest.
Copilot or Gemini is in the plan. A few people use it for emails. Most seats are empty. No rules.
Grounded in your files. Rules everyone read. Three workflows live and measured. Training that stuck.
We write the operating rules, connect the AI to your own files with the permissions you already have, and pick three real workflows to make it stick.
One page, written with you, read by everyone: what the AI may see, what it may say on your behalf, what it keeps, and what never goes in. Which tools are approved and which are retired. Human review on anything that leaves the building. Vendor terms and data residency checked before a single seat is switched on.
Copilot or Gemini connected to the files, mail and chat you already have, with exactly the permissions people already hold. If someone couldn't open the folder before, the AI won't answer them from it. The meeting memory, if you have Voice Memory, becomes something it can draw on. Twenty real questions tested until the answers are right and cited.
Three workflows your team actually does, made faster with the AI: usually proposal first drafts, pre-call briefs and the shared inbox. Two 90-minute working sessions on your real documents, not demos. A champion in each team. Usage measured by team at 30 days, and the rules updated for what you learned.
Measured, not promised. Every one of these is checked with you before we call it done.
One page on what the AI may see, say and keep, which tools are approved, and where a human signs off. Read and acknowledged by the whole team.
Answers come from your documents, mail and meetings, with the permissions you already set, and they cite where they came from. Twenty real questions pass.
Proposal drafts, pre-call briefs, inbox triage, or the three that matter most to you. Live, trained on real work, and measured at 30 days.
The AI & Operations Diagnostic shows where AI earns its place. This module switches it on there first. Everything else builds on it.
These four are the floor. Agreed in week one, written in a page, acknowledged by everyone, and updated as the tools change.
The AI sees what the person already could, nothing more. No folder opens because an AI asked.
Payroll, HR files, legal advice, client financials outside the tenant, and whatever you add to the list.
Drafts are drafts. Proposals, client emails and quotes are reviewed by a person before they're sent.
One AI inside your tenant. Personal accounts retired. New tools scored before anyone signs up.
The prompts, the agents, the rules, the templates, the data and the accounts. Everything is set up in your tenant, in your name. Leave any time and take all of it; nothing GoodOps builds is a reason to stay.
Three questions: what AI are you already paying for, who's using what (including the personal accounts), and which three pieces of work would you most like to be faster? Every tool lands in one of four states. That list is the plan.
Copilot or Gemini seats paid for, barely used. We switch them on properly.
e.g. Microsoft 365 Copilot, Gemini for WorkspacePeople use it, nobody governs it. We write the page and connect the files.
e.g. Copilot on, SharePoint not groundedPersonal ChatGPT, Claude and Gemini accounts. We move everyone to one, inside the tenant.
e.g. Personal AI accounts, pasted client dataYour file system and permissions. We plug in, we don't rebuild.
e.g. SharePoint, Google Drive, your identity providerGoodOps drafts proposals and briefs from its own meeting memory with the same rules. The demo is a real proposal.
You approve before we move to the next step. We start with one team or one cycle, prove it, then roll it out to everyone.
Operating rules written with you, approved tools named, personal accounts listed. Vendor terms and residency checked.
AI connected to your files, mail and meetings with existing permissions. Twenty real questions tested until they pass. Personal tools retired.
Three workflows live. Two working sessions on real documents. A champion per team. Usage dashboard by team.
Runs on your GoodOps retainer. Monthly AI recommendations, rules updated as tools change, usage reviewed at 30, 60 and 90 days.
Then AI, Switched On Properly runs on your GoodOps fractional retainer. Rules, grounding and training are the module. Seats are what you already pay for.
Software: Copilot is now built into Microsoft 365 Business plans; Gemini is included with Workspace Business. Most organizations add nothing.
The things owners ask us on the first call.
Write the rules, connect it to your files with your permissions, retire the personal tools, and train your team on three real workflows until they stick. The licence is the easy part; most organizations we meet have 80% of their seats sitting unused.
Not inside Microsoft 365 Copilot or Gemini for Workspace on business plans; both commit to that in their terms. We check the terms and the data residency before anything is switched on, and we write what we found into your rules.
Whichever one you already pay for. Copilot if you're on Microsoft 365, Gemini if you're on Google Workspace. One tool, inside your tenant, with your permissions. We don't make anyone switch for fun.
Inside the tenant, the AI only sees what the person already could, so the blast radius is already their own access. The rules name what never goes in, and the retired personal accounts remove the usual leak. Monthly review catches the rest.
The ones that waste the most time in your organization. Usually proposal first drafts, pre-call briefs and shared inbox triage. The mini-diagnostic picks them with you.
The 30-minute mini-diagnostic is free. From it you get a fixed price for the four-week rollout, typically CAD 6,000–12,000 depending on team size and how many tools need retiring. Seats are what you already pay for. Once live, it runs on your GoodOps retainer.
Book the 30-minute mini-diagnostic. We'll find your unused seats, your personal accounts, and the three pieces of work that should be faster by next month.
A conversation about your organization.
Not a demo of software.