AI Foundations for Marketing Teams — Training Guide
Produces a facilitator-ready, four-module AI literacy guide (about 150 minutes) that frames AI as augmented intelligence and ends with supervised practice; the prompt-writing workshop follows as a second session.
Use When
- Staff are using AI tools without guidance and need a beginner session on how they work, their limits and hallucinations.
- We need rules for safe use: client confidentiality, Uganda DPPA 2019 personal data, disclosure and human review.
- A team that knows the basics needs a hands-on prompt-writing workshop using the Alpha-Beta-Gamma-Delta-Epsilon structure and PAS and AIDA copy frameworks.
- Managers want supervised practice and proof that trainees can critique and improve prompts.
Do Not Use When
prompt-engineering-library for a reusable library of text, image, audio and video prompt templates.training-client-team for a social-media handover workshop or DIY content handbook.policy-ai-content-ethics for the organisation's AI content policy.- Stop before exercises put real client personal data into public AI tools; use synthetic examples instead.
Required Inputs
| Artefact |
Source/provider |
Required? |
If absent |
| Client business name, industry, country/city and primary goal |
Client lead |
Yes |
Default to Uganda / East Africa; ask for the goal and hold the guide until it is stated. |
| Team size and prior AI experience (none / basic / intermediate) |
Client lead or pre-session survey |
Yes |
Assume no experience and deliver foundations before any prompt-writing module. |
| Primary platforms used (e.g. Facebook, Instagram, WhatsApp) |
Client lead |
Yes |
Use Facebook, Instagram and WhatsApp and label the platform table provisional. |
| Training format and time available (2-hour express / half-day / 4 weekly sessions) |
Client lead |
Yes |
Plan the 150-minute half-day and mark cuts for a 2-hour express. |
| Tool access: Android, free tier, 3G connection for the five hands-on tools |
Facilitator check before the session |
Yes |
Mark unverified tools not assessed and run those segments as demonstrations. |
| Synthetic practice material (no real client personal data) |
Facilitator |
Yes |
Write synthetic briefs; never paste client personal data into public tools. |
Workflow
- Run the intake in training-guide-plan.md and decide whether the team needs foundations, the prompt-writing follow-on module, or both in sequence.
- Verify tool access (Android, free tier, 3G) and prepare synthetic exercise material; stop before any exercise would put real client personal data into a public AI tool.
- Write the Training Overview block with the four sources, then open with the augmented-intelligence frame and junior assistant analogy from foundations-and-limits.md.
- Generate Modules 1–2 (what AI is; what it can and cannot do, with the EA platform table) from foundations-and-limits.md, adding the Co-Pilot vs Co-Thinker and Three Waves segments.
- Generate Modules 3–4 (five hands-on tools; the human quality standard) and the exercises summary from tools-and-human-review.md.
- If the team is ready, schedule the follow-on session from prompt-writing-module.md after this guide, never before it.
- Check the guide against the Quality Standards, apply
anti-ai-slop and block release on an F from ai-slop-audit; correct any failing section and rerun the affected check before hand-off to the facilitator.
Outputs
| Artefact |
Consumer |
Acceptance condition |
| Facilitator-ready training guide, four modules, about 150 minutes |
Client facilitator or marketing manager |
A non-technical manager can run it without extra preparation; client name, industry, platforms and city appear throughout. |
| Supervised exercise set (Co-Pilot/Co-Thinker audit, Three Waves, tool practice, 3-step edit) |
Trainees and facilitator |
Each exercise uses synthetic or approved material and has a stated debrief question. |
| Follow-on prompt-writing session plan (when chosen) |
Facilitator |
Scheduled after foundations and built from the prompt-writing module. |
Evidence Produced
| Evidence |
Format |
Acceptance condition |
| Tool-access check |
Table: tool, Android, free tier, 3G, date checked |
Every hands-on tool is verified or marked not assessed. |
| Trainee critique record |
Before/after prompt or edited output per trainee |
Shows each trainee critiqued and improved an output using the 3-step edit. |
Capability and Permission Boundaries
Read and search only; analysis is read-only. Publishing, spend, live account changes, outreach and personal-data processing need explicit, action-specific client authority. Exercises use synthetic data; any use of client data in AI tools needs a Uganda DPPA 2019 check and client sign-off.
Degraded Mode
Without confirmed team experience level and tool access, return the narrowest qualified result and mark the affected checks not assessed. The augmented-intelligence frame, Modules 1–2 and the human quality standard can still be delivered as a classroom session without live tools.
Decision Rules
| Condition |
Action |
Failure or risk avoided |
| Learners lack a shared AI mental model |
Teach literacy and risk before prompt technique |
Prompt recipes create confidence without judgement |
| Learners understand basic AI limits and need repeatable prompting practice |
Deliver the prompt-writing module from prompt-writing-module.md with worked exercises, human review and verified facts |
Participants copy prompts without checking outputs |
| An exercise would use real client or customer personal data |
Replace it with synthetic data; escalate any real-data use for DPPA 2019 review |
Personal data leaked into public AI tools |
| A hands-on tool fails on Android, free tier or 3G |
Swap to a demonstration or a lighter alternative and note bandwidth guidance |
A session stalled by access the team does not have |
| The client asks for a tool, feature or tier claim to be stated as current |
Verify it on the date of training or mark it not assessed; the guide teaches judgement, not a tool catalogue |
Out-of-date AI tool claims taught as fact |
| The client wants trainees to publish AI drafts during the session |
Keep outputs as reviewed drafts; publishing needs the approval rule and client authority |
Unreviewed AI copy on live client accounts |
Quality Standards
- Augmented intelligence framing is used consistently throughout (AI assists humans, it does not replace them); the junior assistant analogy is included.
- All three AI types (Mechanical / Thinking / Feeling) are explained with Uganda/East Africa marketing examples.
- The "What AI cannot do" section is specific and EA-calibrated, not a generic global list; Luganda/Swahili limitations and cultural intelligence gaps are named explicitly.
- All 5 hands-on tools (ChatGPT, Gemini, Canva, FeedHive, Otter.ai) are verified as accessible on Android, free tier, and 3G connection; bandwidth guidance is included.
- The human quality standard section includes the banned vocabulary list, the 5 signs of AI text, and the 3-step edit process.
- Platform AI applications are presented as a table with EA-specific notes per channel, not a single generic list.
- Output is structured so a non-technical marketing manager can facilitate the session without additional preparation.
- British English spelling throughout; imperative language used in all instructions; Uganda/East Africa, EAT, UGX and WhatsApp-first assumptions stated where they apply.
Anti-Patterns
- Teaching prompt recipes before the team shares a mental model of AI limits. Fix: run foundations first and schedule the prompt-writing module after it.
- Using real customer lists or client files in live exercises. Fix: use synthetic briefs and keep real-data use behind a DPPA 2019 check.
- Presenting a generic global "what AI cannot do" list. Fix: name Luganda/Swahili limits and local cultural intelligence gaps.
- Assuming every trainee has a laptop, paid tier and fast connection. Fix: verify Android, free tier and 3G access and plan demonstrations as a fallback.
- Treating the Co-Pilot mode as the whole of AI use. Fix: run the Co-Pilot vs Co-Thinker audit so the team practises thought-partner work.
- Letting AI output go out unread because it "looks fine". Fix: teach the 3-step edit and the approval rule, and record each trainee's edited output.
References
