Most owners and leaders we talk to have a few worries they never raise on a first call. Some feel awkward asking. Some assume the person selling AI help will say whatever closes the deal.
Here are those questions about using AI in a Canadian business, with our answer to each and an independent source you can check yourself. Where we have written about a topic in more depth, the link is underneath. Read the ones that apply to you, and bring any of them to the call.
Is it safe to put our business information into AI?
Yes, a Canadian business can use AI safely if basic security comes first: strong passwords, staff who can spot phishing, and a sign-in code sent to a phone for email, files and finance systems. Then use a paid business AI plan, not free personal accounts, and decide which systems hold sensitive information before connecting AI to them.
In our client work so far, the security gaps we have found were weak passwords, phishing and missing sign-in codes. None came from the AI tool itself.
The order we suggest:
- Your IT provider confirms that your email, files and finance systems have strong passwords and need a code sent to a phone to sign in.
- You decide which systems hold sensitive information, and what is in them.
- You pick the AI tool that works with what you already run, such as Microsoft 365 or Google Workspace, and that your client agreements allow.
On paid business plans, which are different from free personal accounts, OpenAI, Anthropic and Microsoft all state that they do not use your data to train their AI by default. Where your data is stored depends on the tool and the plan. We confirm that with you and your IT provider before anything is built.
The honest trade-off: the more you connect AI to, the more it can do for you, and the more can go wrong. Connect it to your inbox and something you did not want in a draft will eventually show up in one. That is why every setup we build starts with written rules on what to leave out, and a person reviews anything before it goes out.
Sources: How to prevent phishing for small businesses (Get Cyber Safe, Government of Canada). OpenAI business data. Anthropic: is my data used for training?. Microsoft 365 Copilot enterprise data protection.
More from us: Is my data safe with AI? What will IT say?
What if our IT provider says no?
If your IT provider says no to AI, the setup changes until they can approve it. Bring them in from the start: they usually approve the AI accounts, which systems AI can connect to, and the sharing settings. In our experience, IT providers worry more about staff quietly using personal AI accounts than about a planned setup.
We work with your IT provider directly or send them a checklist, whichever you prefer. Your own staff sign in and approve each connection. We never ask for your passwords or keep logins to your systems.
Source: Microsoft 365 Copilot enterprise data protection explains how Copilot follows the permissions your IT provider already manages.
More from us: Sidekick for IT providers, a page you can forward to yours.
What will AI cost us every month?
Standard business AI plans from Anthropic, OpenAI and Microsoft list at roughly CAD 28 to 41 per person per month, as of our October 2026 check. Microsoft's Copilot is added on top of a Microsoft 365 licence. Outside help and your team's time usually cost more than the tools. Features that charge by usage can be capped before anything is switched on.
Our own fees are agreed in writing, and paid before each stage of work starts. If your firm has at least $1 million in annual revenue, BDC's LIFT program offers advisory help and financing for digital and AI plans. Check current terms with BDC.
Source: BDC LIFT program.
More from us: What AI costs a Canadian business (tools, help and time, with current list prices) and AI grants and financing in Canada.
Our information is a mess. Can we still start?
Yes, a business with messy files and records can still start using AI. The best first jobs sit in a leader's day, such as email, meetings, proposals and follow-ups, and those do not need tidy records. Where messy records do get in the way, cleaning them up becomes one of the jobs, and AI can do much of the sorting.
A common example is a client list kept in three places. AI can match and merge the records, and your team checks the result before anything is replaced.
More from us: The hours test, a 10-minute worksheet to find which task to start with.
Won't AI make things up?
Yes, AI tools sometimes make things up, so a person should review anything AI produces that leaves the business or feeds a real decision. Canada's privacy regulators give the same advice: check AI output before using it in decisions or making it public. A good setup builds much of that checking in, but it does not replace your judgment.
In every setup we build, your rules and review steps are written into the AI's instructions on day one, so it checks its draft against them before you see it. That catches some of the obvious misses.
Source: Principles for responsible, trustworthy and privacy-protective generative AI technologies (Office of the Privacy Commissioner of Canada and provincial regulators).
More from us: Why your AI keeps getting things wrong, and what to check first
Will AI replace my people?
AI does not have to replace your people. Whether to cut staff is your decision, and it is not work we help with. Statistics Canada found that about a third of Canadian businesses using AI chose to train the staff they already had because of it. We help leaders decide where the time AI frees up should go, such as backlogs, delayed projects and work the team could not take on before.
Your staff will wonder about this whether you mention it or not. Tell them early where you stand.
Source: Statistics Canada, AI use by businesses in Canada, second quarter of 2026.
More from us: What happens to your staff when you bring in AI
We're too busy. Should we wait?
If your next month truly has no room, wait to start AI rather than start and stall. Being too busy is one of the most common reasons we see AI work stall. The first few weeks need some of a leader's time: a kickoff, a short coaching session most weeks, and using the setup on tasks from your actual week.
We pick the first job by what takes up most of your week, so that is where your time comes back first.
More from us: The hours test shows in 10 minutes how many hours one task could give back.
We tried ChatGPT and nobody kept using it. Why would this be different?
When a team tries ChatGPT and stops using it, the tool is rarely the problem. AI tools can now connect to your email, calendar and files and follow written instructions about how you work. What is usually missing is everything around the tool: a setup that fits your business, rules for what the AI should and should not do, and coaching through the first few weeks.
One more piece matters: a way to update those instructions each time someone corrects the AI, so the same mistake does not come back.
More from us: Why AI doesn't stick
Should we wait until AI settles down?
You do not need to wait for AI to settle down, because the habits your team builds carry over as the tools change. You would not be going first, either. Statistics Canada reports that the share of Canadian businesses using AI tripled in two years, from 6.1% in 2024 to 19.2% in 2026, and reached 27.8% among businesses with 100 or more employees.
Source: Statistics Canada, AI use by businesses in Canada, second quarter of 2026.
More from us: The four levels of AI adoption
What happens if something happens to you?
If Chase Bernier, who leads every Sidekick engagement, became unavailable, your setup would keep running. You keep everything we build and can go on using it, it is written down as it is built, and your team runs it after handover. Today there is no second consultant, so Chase would refer you to partner firms he trusts.
The goal of every engagement is that your team does not need us in the long run.
More from us: Before you say yes: how working with us works, including what you commit to and when you can stop.
Do we have to tell our clients we use AI?
Where AI is part of a decision about a person, Canada's privacy regulators say that person should be told, and your business stays accountable for the decision even when AI helped make it. Where AI only helps your team draft work a person reviews and sends, the main question is what your client agreements allow.
If your firm is in a regulated profession, such as accounting, law or engineering, check what your professional body says too. A one-page AI use policy gives your staff clear answers to most day-to-day questions.
Source: Principles for responsible, trustworthy and privacy-protective generative AI technologies (Office of the Privacy Commissioner of Canada and provincial regulators).
More from us: AI policy for Canadian businesses, with a one-page template.
About Sidekick
Sidekick Orchestration helps leaders put AI to work, then rolls it out across the company. In practice, that means setting up AI tools such as Claude, ChatGPT or Copilot for a company's leaders first, coaching them, then helping the rest of the team use them. Chase leads every engagement and is your single point of contact.
If AI does not fit your business, we will tell you on the first call. And if one of these questions is the real reason you are holding off, say so. It is usually the most useful part of the conversation.
