Skip to content

AI FAQ for small businesses

AI in your business: straight answers.

Many businesses still don't know how to use AI, or whether it even applies to them. That's normal. Here are the questions I hear most often, with direct answers, sourced figures and no magic promises.

Last updated: October 2026 · 36 questions, 9 topics

Key figures, no spin

Adoption rates vary widely depending on how AI use is defined. Every figure links to its source.

19.2%

of Canadian businesses used AI to produce goods or services in spring 2026, up from 6.1% two years earlier.

Source: Statistics Canada, Q2 2026

9.9% vs 21%

Rural businesses have adopted AI at half the rate of urban businesses.

Source: Statistics Canada, Q2 2026

78.1%

of businesses with no plans to use AI say it's “not relevant” to them. The barrier isn't cost or technology.

Source: Statistics Canada, Q3 2025

39% → 66%

of entrepreneurs said they used AI; 66% actually did once they were shown a list of tools.

Source: BDC, 2024

53%

of Québec workers lack clear guidance: 21% say their employer has no position on AI, and 32% don't know what it is.

Source: NETendances 2025, Université Laval (in French)

17%

median productivity gain in automation projects at Québec small businesses; 62% pay for themselves in under three years.

Source: CFIB and Investissement Québec, 2025 (automation in general, not only AI)

Getting started

Is AI useful for a small business?

Yes, in most cases. For a small business, AI isn't a robot or a big project. It's a tool that drafts a first version, summarizes a document, analyzes sales or answers frequent questions. Many businesses already use it without realizing it, inside their email or their accounting software.

According to BDC, 39% of Canadian entrepreneurs said they used AI in 2024, but 66% actually did once they were shown a list of tools. So the real question isn't “do we need AI?” but “which repetitive task costs me the most time every week?”. That's where we start.

Source: BDC, April 2024 survey (1,247 business owners)

Related service: The AI assessment

Where should a small business start with AI?

Start with one repetitive, frequent task that involves no sensitive data, such as quotes, emails or a monthly report. Time how long it takes, test a tool for two to four weeks, then time it again. If the gain is real, move on to the next task. Small, measurable steps beat a big plan.

The main barrier isn't cost or technology: 78.1% of Canadian businesses with no plans to use AI say it's “not relevant” to them (Statistics Canada, Q3 2025). So start from your trade, not from the technology. If you'd rather have help, that's exactly what my assessment is for: we review your processes together and you leave with three clear priorities, whether or not you work with me afterwards.

Source: Statistics Canada, Q3 2025

Related service: The AI assessment

Are Québec small businesses behind on AI?

A little, especially outside the cities, but the gap is closing. In spring 2026, 19.2% of Canadian businesses used AI to produce their goods or services, three times more than in 2024. Most small businesses are still at the testing stage, so starting now, with a small project, is perfectly reasonable.

The urban–rural gap is clear: 21.0% of urban businesses use AI, compared with 9.9% of rural ones (Statistics Canada, Q2 2026). One well-chosen project brings you up to date without taking unnecessary risks.

Source: Statistics Canada, Q2 2026

Related service: The AI assessment

Do I need an in-house AI expert to get started?

No. Most small businesses adopt AI with the staff they already have. What's usually missing is a bit of method: writing a clear request, checking the answers, and knowing which data should never go into a tool. A hands-on training session of a few hours is often enough for everyday uses.

According to a June 2026 report from the Diversity Institute, 72% of micro-businesses rely on their current staff to adopt AI, and only 12% hire specialists. When you need to connect software or build a custom tool, an outside developer like me steps in as needed, without you having to create a position.

Source: Diversity Institute and Future Skills Centre, June 2026 (in French)

Related service: Team training

Costs, ROI and funding

How much does AI cost for a small business?

It depends on what you want to do. Paid versions of the common tools cost a few dozen dollars per user per month. Connecting two pieces of software takes a few days of development; a custom tool takes more. The safest approach is a short, measured trial, which costs little and tells you whether to go further.

Cost is the second barrier reported by Canadian businesses (10.6%, and 15.1% among those with 20 to 99 employees), after privacy (Statistics Canada, Q2 2026). For reference, the median investment for a micro-business in Québec automation projects is about $20,000 (CFIB and Investissement Québec, 2025), but a well-targeted first AI project can cost far less. Always check tool prices: they change often.

Sources: Statistics Canada, Q2 2026 · CFIB and Investissement Québec, October 2025 (in French)

Related service: The AI assessment

How do I measure the ROI of AI in a small business?

Measure before and after. Note how long the target task takes and how often it comes up, then measure again after a one-month trial. Multiply the hours saved by that person's hourly cost and you get a concrete return you can compare with the cost of the tool. No guesswork, no hype.

Serious studies show real but modest gains: about 2.2 hours a week for workers who use generative AI (Federal Reserve Bank of St. Louis, February 2025) and 26 minutes a day in a large UK government trial (June 2025). That adds up over a year, but it isn't magic. Be wary of promises of “10x productivity”.

Sources: Federal Reserve Bank of St. Louis, February 2025 · UK Government, June 2025

Related service: Process automation

Why do so many businesses see no return on AI?

Because they start too big, or without a measurable goal. A vague “transformation” produces vague results. The projects that pay off target one specific problem, with a clear indicator, one person in charge and a check-in after a few weeks. Small, measurable and quick: that's how I work with every client.

A KPMG survey of Canadian executives (November 2025) found that only 2% of organizations see a return on their generative AI investments, and 31% can't measure it. By contrast, targeted automation projects in Québec small businesses show a median productivity gain of 17%, and 62% pay for themselves in under three years (CFIB and Investissement Québec, October 2025).

Sources: KPMG Canada, November 2025 · CFIB and Investissement Québec, October 2025 (in French)

Related service: The AI assessment

Are there grants for AI in Québec small businesses?

Yes, but they mostly depend on your size. ESSOR can fund up to 50% of a digital assessment that includes AI for businesses with at least $2.5M in revenue, BDC's LIFT program targets businesses with at least $1M in revenue, and Services Québec can subsidize part of your staff training.

For ESSOR (component 1B), the maximum contribution is $20,000. The rules change often: I can point you in the right direction, but you always apply directly to the organization. See the list of funding programs below, with links to official sources.

Sources: Investissement Québec, ESSOR (in French) · BDC, LIFT (April 2026)

Related service: The AI assessment

Data, security and Law 25

I'm not a lawyer: what follows is a plain-language overview. For a specific situation, talk to a legal professional or the person responsible for protecting personal information in your business.

Can I use ChatGPT and still comply with Québec's Law 25?

Yes, if you pick the right plan and are careful about what you enter. In free or personal versions, never include names, emails or identifiable records. To process personal information, use a business plan or an API, which don't train models by default, after a short privacy risk assessment.

On personal ChatGPT accounts, conversations may be used to train models unless you turn that off; OpenAI's business plans and API don't use them by default. At Anthropic (Claude), personal accounts have let users choose since August 2025, while business plans are excluded from training by default. Privacy is also the top barrier reported by Canadian businesses (13.4%, Statistics Canada, Q2 2026).

Sources: OpenAI, personal accounts · OpenAI, business plans · Anthropic, August 2025 terms update · Statistics Canada, Q2 2026

Related service: AI use policy and Law 25

What does Québec's Law 25 say about artificial intelligence?

Law 25 doesn't mention AI as such, but it fully applies whenever personal information is involved. Among other things, it requires a privacy impact assessment before sending that information outside Québec, which most cloud AI tools do, and transparency about decisions made by automated systems.

All of its provisions have been in force since September 22, 2024. If a decision about a person is made solely by an automated system, you must inform them and let them request a review. A confidentiality incident, such as a client list pasted into a consumer tool, must be logged and sometimes reported. As far as I know, there is no AI-specific law that applies to Québec small businesses as of October 2026.

Sources: Government of Québec, Law 25 (in French) · Raymond Chabot Grant Thornton, October 2024 (in French)

Related service: AI use policy and Law 25

How do I stop employees from putting confidential data into ChatGPT?

With a written AI use policy and a short training session. One or two pages are enough: approved tools and which accounts to use, data that's off-limits, mandatory review before anything goes out, and who to ask when in doubt. Banning AI without offering an approved tool tends to push employees toward personal accounts.

In Québec, 21% of workers say their employer has no position on AI and 32% don't know what it is (NETendances 2025). Employees use it anyway. I can help you write that policy and present it to your team, with examples drawn from your own work.

Source: NETendances 2025, Université Laval (in French)

Related service: AI use policy and Law 25

Can we use AI without our data leaving the business?

It's possible, but rarely necessary for a small business. Models can be installed on your own servers, but they require hardware and maintenance, and they're often less capable. For most uses, a business plan, anonymized data and controlled access offer a good balance between security and efficiency.

Choose based on how sensitive your data really is, not on fear. Replacing names with codes before asking for an analysis already removes much of the risk.

Related service: AI use policy and Law 25

Tools and everyday uses

ChatGPT, Copilot, Gemini or Claude: which one should a small business choose?

The best tool is usually the one that fits what you already use. With Microsoft 365, Copilot works inside Outlook, Word and Excel; with Google Workspace, Gemini is built into Gmail and Docs. ChatGPT and Claude are excellent for writing and analysis. Try two of them on your real tasks before deciding.

In Québec, ChatGPT is by far the most used, ahead of Copilot and then Gemini (NETendances 2025). Popularity isn't enough, though: compare them for two weeks on your own documents, then pick a business plan for the team.

Source: NETendances 2025, Université Laval (in French)

Related service: Team training

Is the free version of ChatGPT good enough for a business?

For trying AI out, yes; for business use, rarely. Free versions have usage limits, don't always offer the latest models and may use your conversations to train models. Business plans add account management, stronger privacy commitments and team sharing, which matter as soon as client data is involved.

At a few dozen dollars per person per month, it's often the most cost-effective purchase you can make, as long as the team is trained to use it well.

Source: OpenAI, personal accounts

Related service: Team training

How can AI help write emails and summarize documents?

It's the simplest place to start. Give the AI the context, the tone you want and the points to cover: it drafts an email, summarizes a long document or email thread, and turns your notes into a clear procedure. You review it, adjust it, then send it. You stay in charge of every word.

Writing communications is the number one use among workers who use generative AI (57%, Federal Reserve Bank of St. Louis, 2025), and text analytics is used by 34.5% of Canadian businesses that have adopted AI (Statistics Canada, Q2 2026). In the UK government trial, this kind of work saved about 26 minutes a day. I train teams on these uses, and I can build an internal assistant that answers from your own documents.

Sources: Federal Reserve Bank of St. Louis, February 2025 · Statistics Canada, Q2 2026 · UK Government, June 2025

Related service: Team training

How can AI analyze my sales data and Excel files?

AI can clean up data, write formulas, summarize a monthly report and spot trends in your sales, then answer your questions in plain language. It doesn't replace your judgment: always check the source figures, because it can get a calculation or an interpretation wrong. Think of it as a fast analyst who needs supervision.

Data analytics is the leading use of AI among Canadian businesses that have adopted it (36.6%, Statistics Canada, Q2 2026). To go beyond a spreadsheet, I build up-to-date dashboards connected to your software, with AI that summarizes the trends. Personal data gets anonymized before any analysis.

Source: Statistics Canada, Q2 2026

Related service: Data interpretation and dashboards

What is a prompt, and how do you write a good one?

A prompt is the instruction you give the AI. A good prompt spells out the context, the role you expect, the format you want, an example and the limits to respect. “Write an email” gets a generic result; a precise instruction gets text that's almost ready to send. It's a skill you can learn in a few hours.

It's what makes the difference between a disappointing tool and one you rely on every day. In my workshops, every participant leaves with prompts tailored to their role and your business.

Related service: Team training

Can AI make mistakes, and who is responsible?

Yes. AI can invent facts, figures or references with great confidence; these are called “hallucinations”. And your business is the one held responsible. The rule is simple: AI prepares, a person checks, especially for prices, commitments, numbers and anything legal. Used that way, it saves time instead of creating problems.

In February 2024, Air Canada was held liable for false information given by the chatbot on its website: the tribunal ruled that the company is responsible for everything on its site. A Harvard and BCG study (2023) also found that, on tasks AI handles poorly, consultants using it were 19% less likely to get the right answer.

Sources: Moffatt v. Air Canada, 2024 BCCRT 149 (McCarthy Tétrault analysis) · Dell'Acqua et al., Harvard and BCG, 2023

Related service: AI assistants grounded in your documents

Employees and training

Will AI replace my employees?

Current data doesn't point that way for small businesses. The vast majority of businesses adopting AI expect no change in headcount. In practice, AI mostly removes tedious work such as data entry, formatting and first drafts, and helps your existing team keep up with the workload.

Among Canadian businesses planning to adopt AI, 69.9% expected no change in staffing (Statistics Canada, Q3 2025). A University of Chicago study of 25,000 Danish workers found no measurable effect on wages or hours worked (2025). With today's labour shortage, AI is mainly extra help.

Sources: Statistics Canada, Q3 2025 · Humlum and Vestergaard, University of Chicago, 2025

Related service: Team training

How do I get a sceptical team on board with AI?

Start from their frustrations, not from the technology. Ask each person which task they'd gladly stop doing, then show on that exact case what AI can do, and what it can't. Be clear about your intentions: saving time, not cutting jobs. And give people room to make mistakes while they learn.

Simple rules reassure as much as they protect. In my training sessions, I work with the team's real documents: that's what turns suspicion into curiosity.

Related service: Team training

What AI training should I give my employees?

Short, practical training tailored to each role: writing a clear request, checking an answer, following privacy rules, then practising on their own tasks. A few hours is enough for everyday uses, as long as people work on the company's real documents rather than generic examples. Follow-up questions a few weeks later help it stick.

According to the Diversity Institute (2026), fewer than one in ten small businesses have access to formal AI training, and nearly half of employees who use these tools have never been trained. I run these workshops in person in Saguenay–Lac-Saint-Jean or remotely anywhere in Québec. Part of the cost can sometimes be subsidized by Services Québec (MFOR).

Source: Diversity Institute and Future Skills Centre, June 2026 (in French)

Related service: Team training

Marketing and branding

Can I create my logo with AI?

You can use it to explore ideas, but I don't recommend using an AI-generated logo as is. Its copyright protection is uncertain, it may resemble an existing trademark, and a logo has to work in vector format, in black and white and at small sizes. AI explores; a designer finalizes.

In the United States, the Copyright Office concluded on January 29, 2025 that a prompt alone isn't enough to obtain copyright. My method: AI speeds up exploration, then I draw and finalize the logo in Illustrator. Before filing a trademark, also search the Canadian Intellectual Property Office database.

Source: US Copyright Office, January 29, 2025 (Sidley analysis)

Related service: Logo and branding

Who owns AI-generated images in Canada?

The question isn't settled in Canada yet: the Copyright Act hasn't been amended and a case is before the Federal Court. What is clear is that the more human work goes into an image, such as choices, retouching and composition, the stronger your position. The tool you choose matters too.

Adobe says it trained Firefly on licensed content and offers contractual indemnification to some enterprise customers, which other tools don't. For your important commercial visuals, I document the tools and steps used.

Sources: Noovo Info, Federal Court (in French) · Computerworld, Adobe Firefly

Related service: AI-assisted graphic design

Can AI write my marketing copy?

It can write a solid first draft quickly, in English or French. But generic copy won't set you apart, and customers quickly recognize the “robot” tone. The right approach: AI structures the message and suggests variations, then a person reworks the text with your words, your examples and your voice.

Always double-check facts, prices and promises: you're responsible for them. Marketing automation is among the AI uses Canadian businesses plan most (25.6%, Statistics Canada, Q3 2025).

Source: Statistics Canada, Q3 2025

Related service: AI-assisted graphic design

Is it a problem if customers can tell something was made with AI?

Not if the result is good. Customers mostly notice AI when it's done badly: distorted hands, empty copy, visuals that have nothing to do with your brand. Using AI isn't a problem in itself; publishing poor-quality content is. What matters is a result that's accurate, consistent with your brand and useful.

I recommend being transparent when it's relevant, for example with a chatbot or a realistic image of something that doesn't exist.

Related service: AI-assisted graphic design

Automation and API integrations

What is an API, and why should a small business care?

An API is a doorway that lets two pieces of software talk to each other directly. A request filled out on your website can create a client in your CRM, a draft quote and a follow-up task, without retyping anything. Less re-entry means fewer errors and more time for your customers.

According to a survey by software vendor Salesforce, growing small businesses are twice as likely as declining ones to have well-integrated tools (66% vs. 32%, 2025). I regularly connect QuickBooks Online, Zoho CRM, Stripe and Amazon Web Services.

Source: Salesforce, SMB Trends Report 2025 (vendor survey)

Related service: API integrations

Which tasks should a small business automate first?

The best candidates are repetitive, frequent and rule-based: sending website requests to your CRM, preparing quotes from a price list, extracting data from supplier invoices, sending appointment reminders or producing the monthly sales report. Start with the one that eats the most time each week, and measure it before you touch it.

Québec small businesses automate administrative tasks first, with a median productivity gain of 17% (CFIB and Investissement Québec, 2025). Not every automation needs AI: sometimes a good integration is enough, and more reliable. That's what I did with PoPadema, a purchase-order and quote app connected to QuickBooks Online.

Source: CFIB and Investissement Québec, October 2025 (in French)

Related service: Process automation

Can old software that doesn't talk to anything still be connected?

Often, yes. Many business applications offer an API, or at least automated exports. If not, we can go through files, structured emails or a small intermediate database. I start with an inventory of your tools, then propose the simplest solution that will hold up over time. No unnecessary rebuild.

I develop in Node.js and MongoDB, and I document everything so you're not dependent on me. For advanced configuration of a specific accounting package or ERP, beyond integration, I'll tell you honestly if a specialist in that software is the better choice.

Related service: API integrations

Should we automate everything at once?

Definitely not. Automate one process at a time, fine-tune it, then move to the next. A well-chosen first project pays off quickly and teaches you how to work with these tools. “Complete transformation” projects are the ones that go off the rails most often, and their return is hard to measure.

Always keep a person in the loop at critical steps, such as sending a quote or a payment: automation should save you time, not take control away from you.

Related service: Process automation

Video

Can a small business get quality video on a small budget with AI?

Yes, for many everyday needs: social media clips, product presentations, testimonials, tutorials. AI mainly speeds up editing, with transcription, subtitles, silence removal, vertical reframing and audio cleanup. Less production time means a lower cost, which puts regular video within reach of a small business.

It doesn't replace a good idea or good footage: I help you plan short, useful clips that are easy to shoot on a phone, then I take care of the editing.

Related service: Video editing

Can you generate an entire video with AI?

Technically yes, and the tools are improving fast. But for a small business, I mostly use them for B-roll, animations or backgrounds. Fully generated videos still often look artificial, raise rights questions, and earn less trust than real people in real places. Authentic footage usually wins.

Your team, your shop and your customers are your best content; AI helps showcase them at a lower cost.

Related service: Video editing

AI assessment

Not sure where to start? That's exactly my job.

We look at your processes, your tools and your pain points together. I'll tell you honestly where AI is worth it, and where it isn't. You leave with three clear priorities.

Request an assessment

Funding programs

The programs below were in effect or announced in 2025–2026 according to their official sources. They change often: check the conditions and amounts with the organization before you apply. I can point you in the right direction, but you always apply directly to them.

QuébecAssessment and implementation

ESSOR, components 1A, 1B and 1C (Investissement Québec)

For businesses with 250 employees or fewer and at least $2.5M in revenue. Component 1B funds up to 50% of a digital assessment or plan, including AI integration (maximum $20,000); component 1C covers implementation (maximum $50,000). External consultant fees are eligible.

In effect until March 31, 2027 · Official site (in French)

CanadaAdvice and loans

LIFT (BDC)

For businesses with at least $1M in annual revenue (digital and AI stream). Guidance from AI advisors and loans from $25,000 to $5M at favourable rates, with possible principal deferral.

Launched April 24, 2026 · Official announcement

QuébecTraining

Workforce training measure (MFOR), Services Québec

Subsidizes part of your staff training costs, with a higher rate when the training is required by a new technology. You apply through a Services Québec business advisor.

Rates to be reconfirmed with Services Québec · Québec.ca (in French)

QuébecSector training

Productivité-Compétences (Ministère de l'Emploi)

Funding paid to organizations (sector committees, associations) that offer AI training to small businesses in their sector: construction, transportation, forestry, manufacturing, tourism and more. Businesses sign up for those sessions.

Projects announced in 2026 · Québec.ca (in French)

QuébecSupport

Offensive de transformation numérique (OTN)

Funding paid to support organizations that offer digital assessments and follow-up to small businesses, often free or at reduced cost. You access it through those organizations.

Extended in the 2025–2028 SME plan · Project list (in French)

CanadaInnovation

NRC IRAP, AI Assist

For small businesses building AI into their own products or services (innovation), with support from an industrial technology advisor. Less relevant if you only want to use existing tools.

Active in 2025–2026 · NRC

Running a very small business?

The ESSOR ($2.5M) and LIFT ($1M) thresholds exclude many small businesses. Realistic options are then MFOR for training, Productivité-Compétences sessions offered in your sector, OTN through a regional organization, and your MRC's economic development office, which can point you in the right direction.

Often mentioned, but not open to you

  • PACME: a 2020 pandemic-era program; no active version found for 2025–2026.
  • Canada Digital Adoption Program (CDAP): applications have been closed since 2024.
  • CDAEIA tax credit: aimed at IT companies that build AI solutions, not at businesses that use them.