Elements of AI
A free online course that explains AI in plain language, no math required.
Visit site (opens in a new tab)AI tools by task
Which AI tools should a small business start with?
Start with the AI already inside software you pay for. Examples are for orientation only; we do not earn commissions from any vendor.
| Task | What it does | Examples | Start safely |
|---|---|---|---|
| Writing and everyday assistance | Draft emails, summarize documents, brainstorm | ChatGPT, Claude, Microsoft Copilot, Gemini | Choose a business or team plan so you control data use. |
| Meeting notes | Transcribe calls, list action items | Built into Microsoft Teams, Zoom and Google Meet; Otter, Fireflies | Tell attendees when a meeting is recorded. |
| Customer service | Answer FAQs, draft replies, route tickets | Zendesk, Intercom, Tidio, Freshdesk | Limit the assistant to public information first. |
| Marketing and design | Social posts, images, campaign ideas | Canva, HubSpot, Adobe Express | Check brand tone and image rights before publishing. |
| Bookkeeping and finance | Categorize transactions, read receipts | QuickBooks, Xero, Dext | Keep your accountant reviewing the output. |
| Automation | Connect apps and move data between them | Zapier, Make, Microsoft Power Automate | Start with one workflow and log every run. |
| Data and reporting | Summarize spreadsheets, build dashboards | Copilot in Excel, Gemini in Sheets, Power BI, Looker Studio | Clean your data before you trust the charts. |
Free download
The AI Readiness Toolkit
Four practical templates we use with clients, free to adapt for your business.
- AI use policy template. A one-page policy for teams and a client-facing version for freelancers.
- Vendor security questionnaire. 20 questions about data location, retention, training and breach notice.
- Use-case scoring sheet. Rank ideas by value, effort and risk to choose your first pilot.
- Funding application checklist. What funders ask for and how to prepare it.
Email me the toolkit
Vetting AI professionals
How do I choose a trustworthy AI consultant or developer?
Anyone can call themselves an AI expert. Ask these questions, watch for red flags and score each candidate the same way.
12 questions to ask
- Can you show two or three similar projects with measurable results?
- Who exactly will do the work, and what are their qualifications?
- Where will our data be stored and processed, and who can access it?
- Will our data be used to train any model? Can we opt out in writing?
- How will you measure success, and what is the baseline?
- Can we start with a small, fixed-price pilot?
- Which tools or models will you use, and why those?
- What happens to our data and accounts when the contract ends?
- How will you train our staff to run this without you?
- What are the ongoing costs after launch (licences, usage, support)?
- How do you test for errors and bias before going live?
- Can we speak to two clients you have worked with in the past year?
Red flags
- Guarantees of results, or ROI promised before they understand your business
- Vague answers about where data goes or who owns what is built
- Pressure to sign a long contract or buy their own product
- No references, no case studies, or only demos
- Plans to replace staff judgment with no human review
- Unclear pricing, or costs that scale without limits
Where to find qualified help
- An NRC IRAP advisor can often suggest vetted experts
- Regional innovation centres and tech associations in your province
- University and college partnerships through Mitacs
- Referrals from businesses that have completed a similar project
Provider scorecard
Tick what the candidate has shown you in writing. Use one scorecard per provider.
Keep learning
Free, trusted places to learn more
Canadian Centre for Cyber Security
Small business guides on phishing, passwords and safe cloud use.
Visit site (opens in a new tab)Office of the Privacy Commissioner
Guidance on privacy obligations for businesses using new technology.
Visit site (opens in a new tab)Canada's national AI institutes
Amii (Edmonton), Mila (Montréal) and the Vector Institute (Toronto) run programs for businesses adopting AI.
Visit site (opens in a new tab)Want an independent opinion on a proposal?
Send us the proposal you are considering and we will review it against this checklist.