Training · Hero
The useful detailsSettle the tool and the rules first. Then pick two or three real tasks per role, and run a hands-on session where people practise those tasks on safe examples and learn to check the output. Give managers a 30-day plan afterwards. Training that skips the rules leaves people unsure what's allowed. Training that skips the follow-through fades within a month.
Training people on a tool they aren't allowed to use for real work wastes the session.
Try it freeWant to see a session first? The free AI Hour is online on the first Friday of every month. Next: Fri, Nov 6, 12 PM ET.
Send a short anonymous survey a week ahead. Anonymity matters: people won't admit to using personal accounts if their name is on the answer.
The answers tell you where to aim. If half the team is already using personal accounts, the security block matters more than prompt tips. If most people answer "never tried it", start slower.
Good first tasks are frequent, text-heavy, easy to check and low-risk. Examples that work in most offices:
| Role | Good first tasks | Keep with a person |
|---|---|---|
| Admin and operations | Meeting notes into actions; first drafts of procedures; summarizing long email threads | Commitments, dates and anything sent externally |
| Client service | Reply drafts to common questions from approved wording; rewriting for tone | Anything about money, liability or exceptions |
| Finance | Explaining or writing spreadsheet formulas; month-end checklists; plain-language summaries of reports | Every number; anything with account details |
| HR and people | First drafts of job postings and onboarding checklists; interview question banks | Every decision about a person |
| Managers | Status updates from bullet points; meeting agendas; planning a difficult conversation (no names) | The judgment and the conversation itself |
An HR note for Ontario employers: since January 1, 2026, employers with 25 or more employees must say in publicly advertised job postings if AI is used to screen, assess or select applicants. Using AI to help write the posting doesn't trigger this, but screening résumés with it does. Cover this in any HR session.
A demo shows what's possible. Practice shows what people can actually do. A three-hour session can look like this:
| Time | Block | What people do |
|---|---|---|
| 20 min | What these tools are and aren't | See a confident wrong answer on purpose, so nobody trusts the tool blindly |
| 20 min | The information rule | Sort ten example items into red / yellow / green together |
| 40 min | Asking well | Rewrite a one-line request using role, task, context and format. Compare the two outputs side by side |
| 15 min | Break | |
| 45 min | Hands-on lab by role | Each person does one task from Step 3 on a fictional example, start to finish |
| 25 min | Checking the output | Use the review checklist below on their own result. Ask the tool "What did you assume that I should check?" |
| 15 min | Wrap-up | Each person writes down the one task they'll repeat three times this week |
Two exercises that work well:
Use fictional examples that feel like the job. Don't use lightly edited real files.
"Review the output" means nothing unless people know what to look for. Give everyone this list:
If checking takes longer than doing the task the old way, that task isn't a good fit yet.
A next stepRather have someone run it? See the public workshop or private team training, or book a free 30-minute call.
| Week | Team does | Manager does |
|---|---|---|
| 1 | Each person uses their chosen task three times | Asks one question in the team meeting: "What worked, what didn't?" |
| 2 | Start a shared "prompts that work" document | Adds the best two prompts; removes anything with client details |
| 3 | Try the task on a slightly harder case | Spot-checks two outputs against the checklist |
| 4 | Each person notes roughly how long the task takes now | Decides: keep, change or stop. Picks the next task |
What to measure. Keep it light: how many people use the approved tool for work each week; the time one repeated task takes, measured on a few real examples before and after; and how often outputs need major rework. Don't count attendance as the result.
Ask these questions before you pay:
Red flags: guaranteed time savings before anyone has looked at your work. A session that is only "prompt hacks". Pressure to paste real client data into a tool. No answer to "what do we do next Tuesday?"
About funding: government training grants usually have rules about which providers qualify, which training is covered, and approval before the training starts. Check the program's rules before you plan a budget around one. Hero's training isn't set up as grant-eligible training today.
Hero runs this kind of training: public workshop seats open to anyone, 1:1 sessions (one trainer, one person) and private team training on your own work, with a live course coming soon. Online anywhere in Canada, or in person in the GTA. Prices are on the pricing page. You can also run everything above yourself. The free guide has starter prompts by role and a 30-day plan you can print.
Sources checked (October 7, 2026): Ontario, Your guide to the ESA: publicly advertised job postings · Microsoft Learn, Copilot Chat privacy and protections · OpenAI Help Center, ChatGPT Business FAQ.
A CLEAR NEXT STEP STARTS HERE
Tell us about your team and the tools you use. We’ll tell you whether a public workshop, 1:1 coaching, or private team training fits.
Free 30-minute call with our founder · Video call, camera optional