AI workshop vs self-paced course for New Zealand teams.

How do AI workshops and self-paced courses compare?

A workshop is live, led by a trainer, and ideally run on your own tools and documents. A self-paced course is recorded or written material people work through alone. Both can be good; they suit different jobs.

The need is real. In KPMG and the University of Melbourne’s 2025 study, only 24% of New Zealanders had done AI-related training, against 39% globally. Datacom’s 2026 survey of larger New Zealand organisations found 66% had provided AI skills training in the past year, which is a vendor survey but shows how fast employers are moving.

Why does course content date so quickly?

Because the products keep changing. In 2026 alone, Microsoft renamed Microsoft 365 Copilot to Microsoft Copilot, Google renamed NotebookLM to Gemini Notebook on 16 July, Anthropic merged Cowork into the main Claude app on 16 September, and every ChatGPT user moved to GPT-6 on 7 October. A course recorded early in the year can teach menus that no longer exist. Check the date of any course before buying it, and prefer material that teaches habits (checking sources, protecting data, structuring prompts) over button-clicking.

How do you test a workshop and a course on the same team?

Split a team of about 12 into two groups of six with a similar mix of roles. Give both the same ten-task test before training and again two weeks after, then check real use at 30 days.

1. Customer reply from a house brief. Each person drafts three replies to real enquiries in your business AI workspace, using your house brief. Score correct facts and minutes to make each sendable.
2. Check a summary against its source. Give everyone an AI summary of a 20-page document with three planted errors. Count how many errors each person finds.
3. Ten data decisions. Present ten short scenarios, such as pasting a complaint email with a customer’s address into a personal ChatGPT account. Score each answer against your AI use policy.

At 30 days, compare which group is still using the approved tool and reusing saved prompts. Our lesson on measuring adoption after training and the AI training needs assessment give you the instruments.

What does each cost per person who uses it?

Use variables, because fees vary widely and your adoption rate matters more than either fee.

• Workshop total: W + n × Hw × r, where W is the fee, n the number of staff, Hw the hours each spends and r the loaded hourly cost of staff time.
• Course total: n × c + n × Hc × r, where c is the price per seat (sometimes zero) and Hc the hours each spends.
• Cost per person who applies it: total ÷ (n × q), where q is the share still using it well at 30 days.

Illustrative assumptions, to replace with your own: 12 staff, r = NZ$60 an hour, a four-hour workshop and a six-hour course. Staff time alone is 12 × 4 × 60 = NZ$2,880 for the workshop and 12 × 6 × 60 = NZ$4,320 for the course, before any fee. That is the first lesson: time usually costs more than the fee.

The second lesson is q. If a free course reaches q = 0.4, its cost per person applying it is 4,320 ÷ (12 × 0.4) = NZ$900. If a workshop reaches q = 0.75, it matches that only when (W + 2,880) ÷ 9 ≤ 900, so the fee W must be no more than NZ$5,220. If your course group reaches the same q as the workshop group, the course wins on cost. These q values are placeholders; measure your own with the test above.

Which should you choose?

Choose a workshop if

• The team needs shared rules about which data may go into which tool.
• You want practice on your own documents, systems and house style.
• Leaders need to agree priorities together; see executive AI workshops.
• People are unlikely to protect time for a course on their own.

Choose a self-paced course if

• People need foundations before they can use a workshop well.
• Staff work across sites, shifts or time zones, or you onboard new starters often.
• Budgets are tight and your people do finish courses.
• The course is current, dated in 2026 and teaches habits, not just menus.

For most teams, a blend works: free self-paced basics, such as our Learn lessons, NCSC NZ’s AI guide for small businesses and MBIE’s responsible AI guidance, then a workshop on real work, then a 30-day check. Our AI learning plan for teams lays out that sequence.

Where do both fall short?

• Training before tools. Teaching people to use a tool they do not yet have a business licence for pushes them to personal accounts. Buy the seats first.
• Generic data rules. The Privacy Commissioner advises not to put personal or confidential information into generative AI tools unless the provider has confirmed it is not retained or disclosed. Training must translate that into your specific tools and plans.
• Change without support. WorkSafe’s psychosocial risk guidance lists lack of training to use new technology among workplace hazards. It does not mention AI, but the point carries over.
• No follow-up. Both formats fade without someone to answer questions afterwards; an internal AI champion fills that gap.

If a workshop is the right fit, our AI training for business teams runs on your tools and documents.

Written by the Levy Street team in Wellington. Last updated 19 October 2026.