Changelog

Original Materials (4 Workshops)

Workshop 001

• Slide 11 — Updated Midjourney example to v7.
• Slide 15 notes — Removed claim about processing power doubling every 3–4 months (no longer accurate).
• Slide 16 notes — Updated subscription info.

Workshop 002

• Slides 5–8 — Removed entirely and section notes rewritten. The chatbot features covered in these slides have all changed significantly and the content was no longer accurate.

Workshop 003

• Resource doc — Minimum plan name updated from “Pro” to “Plus” to reflect current naming.
• Slide 11 — Added Canvas to the capabilities list. Canvas is now toggled on by default when creating a custom GPT.
• Slide 12 — Removed “2025” from the AIFNT source tag on the Three C’s Framework so it remains evergreen.
• Slide 13 resource doc — Replaced the obsolete “Use DALL-E” instruction. New wording: “Click the image and upload an icon image (for example, one you’ve generated with ChatGPT’s image tool in a separate chat).”
• Slide 14 resource doc — Replaced the custom GPT checkbox screenshot to match the current interface.
• Slide 15 resource doc — Updated the instruction-refusal tip to clarify that this method alone won’t fully protect data.
• Slide 17 resource doc — Removed “no internet access” from Claude’s cons (no longer true). Updated sharing note: Claude projects can now be shared on team plans. Added note that Gems are now shareable in the same way as Google Docs, and that Claude projects now include project-specific memory.
• Slide 20 resource doc — “ChatGPT Teams” renamed to “ChatGPT Business/Enterprise.”
• Slide 21 quiz resource doc — Q10 updated to reflect that Claude projects can now be shared in a team workspace.

Workshop 004

• Slide 4 notes — Removed the “£14bn invested by government” claim (much of that figure was private investment).
• Slide 5 — Added new use cases with source links in notes (Microsoft medical AI, Forbes healthcare tech 2026, UNEP-WCMC/Google wildlife trade AI, Google adaptive interfaces for accessibility).
• Slide 6 — Added updated news stories with source links (Bloomberg AI data centre water usage, Guardian deepfake study Feb 2026).
• Slide 7 notes — Updated copyright section to reflect the legal developments that followed the original “ChatGPT stole copyrighted work” story.
• Slide 11 notes — Updated the Data Analysis feature name.
• Slide 18 notes — Added note on Suno free-tier limitations.
• Slide 23 notes — Added note on how much agentic AI has progressed since Summer 2025.
• Slide 25 notes — Updated Copilot positioning — it is no longer the weaker choice.
• Slide 35 notes — Updated Claude section to reference Code and Cowork.
• Slide 47 — Removed outdated hallucination statistics and replaced with current information.
• Slide 48 — Removed the Sarah Silverman reference (case was substantially dismissed). Added updates on the now-settled Getty and Suno cases. Added the Thaler v Perlmutter ruling (March 2026): the US Supreme Court ruled that AI-generated output cannot be copyrighted or owned.
• Canva link doc — Updated Canva plan from “Pro” to “Plus.” Updated Ideogram and Runway entries to mention affordable basic plans (free tiers are now very limited).

Brand New Materials (15 Workshops)

Intro to ChatGPT for Business

• Resource doc p.3 — Updated model info to include GPT Go.
• Slides & resource docs — “Connectors” renamed to “Apps” throughout (reflects OpenAI’s current naming).

Intro to Copilot for 365

• No material content changes. Links verified, slide notes added.

AI for Small Business Owners

• Slide 10 — Fixed broken link (now points to the correct document).

Let’s Build Custom GPTs

• No material content changes. Links verified, slide notes added.

Risks, Ethics and Limitations in AI

• Slide 29 notes — Added further detail on recent copyright cases.
• All linked videos checked for accuracy — no changes needed.

Turbo Boost Your Personal Productivity

• No material content changes. All software links verified and working.

Build an AI First Business

• All slides — Direct links to prompts added throughout (quality of life improvement — worth replicating if you use linked prompts).
• The non-techie image prompting guide from Session 5 was previously missing from the materials — it has now been added to slide notes and slides.
• Removed an orphaned “Prompt Workbook (9 business categories)” reference from Session 3 notes (it did not appear in the workbook).
• Removed “Better Social Post Example Prompt” from Session 5 prompt list (not referenced on slides or in delivery).
• Slide 123 — Fact-checked “15–20% of new AI users encounter hallucinations” — still accurate, no change.
• Slide 133 — Fact-checked legal/copyright advice — still current, no change.

GenAI for Business Beginners

• Slide 29 — Updated AI chatbot descriptions to reflect the current landscape:
  • ChatGPT — Still described as a strong general-purpose “Swiss Army knife”
  • Gemini — Better integrated with Google products and more reliable, though occasional inconsistencies remain (particularly Google AI Search)
  • Claude — Very human-sounding and thoughtful, particularly in longer-form writing and analysis
  • Copilot — Deeply integrated into Microsoft 365 and Windows; recently much improved; now the default assistant in many organisations

Strategic AI for Senior Leaders

• Slide 7 — Replaced an outdated executive AI adoption statistic with more future-proof language (“The vast majority of executives now report AI initiatives underway”), aligned with the McKinsey 2025 report.
• Resource notes — Updated Manus and Genspark descriptions for clarity. Added a link to McKinsey reports as a recommended source for AI-in-business statistics (noted: US-skewed).
• Resource notes — The AI Maturity Self-Assessment (referenced in Session 1) has been added as a PDF and linked in the research notes.

Supercharge Your Sales with AI

• Slide 33 notes — Added a 2026 update note (timeline approaching 2027).
• Slide 78 — Removed extraneous text.

ChatGPT for Small Business: Essentials

• Resource doc — Direct links added to each prompt. Added note to “Tools & URLs” section: OpenAI regularly updates pricing and tiers — check before delivery.
• Slide 6 — Replaced legacy pricing tier descriptions (Free/Plus/Teams/Enterprise/Pro) with current 2025–26 positioning. Go and Plus are now described as “full-featured individual plans with GPT-5 access and faster responses” with indicative pricing (£8–£20/month).
• Slide 98/99 resource doc — Added trainer note: many larger organisations now use Connectors in production on Business/Enterprise plans. Treat as advanced features, always subject to your client’s security, data-protection, and IT policies.

Ultimate Copilot for Work: Essentials

• Resource doc — Direct links added to each prompt. Copilot plan naming updated to match current Microsoft terminology. Added link to Microsoft’s plan pricing page. Added note that Copilot pricing and M365 bundling is due to change in July 2026.
• Slide 21 — Added “video” to the list of creative output types (feature introduced November 2025).
• Slide 47 — Corrected the COCO framework: the final “O” now correctly reads “Output” (was mislabelled as “Objective”).
• Resource notes, Session 5 — File upload limit updated from 10 to 20 files (or 100 SharePoint files). Added note on Researcher & Analyst’s 25 combined uses/month limit.
• Resource notes, Session 6 — Added a summary of significant Notebook updates in 2026, plus a link to a useful walkthrough video (slides unchanged).
• Resource doc header — Added a link to Microsoft’s feature roadmap page so you can stay current between update rounds.

Practical AI for Senior Leaders

• Resource doc — Direct links added to each prompt. Claude description updated from “a strong underdog” to “growing in popularity.” ChatGPT description adjusted: innovation is no longer its key differentiator, but it retains dominant market share.
• Slide 7 — Removed Manus (acquired by Meta in December 2025; expected to be absorbed into Meta’s product suite rather than continuing as a standalone tool).
• Slides 12 and 13 — Were previously duplicated. Slide 12 now has the correct content on super agents/tools.
• Slide 15 — Deep Research and Multi-Modal tools previously had identical descriptions. Multi-modal tools description has been corrected.
• Slide 16 — Software engineering has been removed from the “AI weaknesses” slide. The original 2024 stat (~5% success on real-world GitHub issues) is now significantly outdated (current benchmarks show 50%+).
• Resource doc (privacy & security) — Added note: since January 2026, OpenAI displays paid advertising in the free tier of ChatGPT — another reason to recommend paid plans to clients.
• Resource doc — Added a new section on regulation, covering the EU AI Act and its implications for UK businesses.
• Resource doc — Added a more recent productivity study. It supports the 2023 figures shown on Slide 15, so no slide change was needed.

Prompt Like a Pro (Beginner)

• Resource doc — Direct links added to each prompt.
• Slide 33 — Changed “2025” to “in the next 12 months” (evergreen fix — worth checking if you’ve customised this slide).
• Slide 41 — Removed a duplicate slide.
• Slide 46 notes — Added a note for trainer awareness: research (Wharton) suggests that offering “tips” or “penalties” in prompts doesn’t materially improve output quality.

Prompt Like a Pro: Next Level

• Resource doc — Direct links added to each prompt.
• Slide 6 — “Build Your Own Prompt Machine” has been renamed to “Promptathon” to match actual course content.
• Slide 26 — Context Window Token Limits table has been fully updated:
  • Removed the inaccurate “ChatGPT5 and Copilot — 256,000 tokens” row
  • ChatGPT (Plus) updated to GPT-5.3 Instant — 32,000 tokens
  • Claude updated to Claude Sonnet 4.6 — 1,000,000 tokens
  • Gemini updated to Gemini 3 Flash — 1,000,000 tokens
  • Microsoft Copilot removed from the comparison table (no single published token figure; varies by surface)
  • A trainer note has been added to the resource doc explaining the Copilot exclusion and advising you to check current figures before delivery

What to Do With This Information

Please review the updated master files against your own customised versions and make the corresponding changes. Pay particular attention to:

• Factual corrections (token limits, copyright cases, pricing tiers, tool naming) — these affect delivery accuracy.
• Evergreen fixes (date removals, future-proof phrasing) — worth replicating if you’ve customised those slides, so your materials don’t date quickly.
• New prompt links in resource docs — a quality-of-life improvement you may want to carry across to your own versions.

If you have any questions about specific changes, please get in touch with the team.