GenAI Essentials for Enterprise Productivity

Bloom-aligned objectives • Remember / Understand: key terms and model behaviours • Apply: safe usage patterns for everyday tasks • Analyze: when GenAI is likely to fail (and why) Topics • AI vs ML vs GenAI: what changes for enterprise work • How LLMs generate text (high-level): tokens, context windows, next-token prediction • Common behaviors and limits o Hallucinations and overconfidence o Sensitivity to phrasing (prompt dependence) o Stale/unknown information and when to verify • Enterprise value levers o Time-to-first-draft reduction o Information synthesis and decision support o Standardization of outputs (templates, tone, format) Activity (20–25 min): “AI in my workflow map” Participants map 5 recurring tasks and classify them into: • Drafting / rewriting • Summarizing / extracting • Structuring / formatting • Analyzing / comparing • Decision-support (with verification) Micro-lab 1 (30 min): “Grounded summary” Input: 2–3 page internal-style document (sanitized). Output: structured summary with: • Key points • Decisions needed • Risks/unknowns • Follow-up questions Deliverable: one-page summary + verification checklist.

Bloom-aligned objectives

• Apply: a prompt blueprint for consistent results
• Analyze: prompt failures and iterate systematically
• Create: reusable prompt templates for personal use

Core prompting framework (P-T-C-F-T blueprint)

• Persona: assistant role and expertise
• Task: verb + objective
• Context: inputs, constraints, audience, definition of done
• Format: required structure (bullets, tables, sections)
• Tone: style, length, formality, language constraints

Topics

• Zero-shot vs few-shot prompting (when examples help)
• Iterative prompting loop: produce → critique → refine → finalize
• Output control
  • formatting requirements
  • length limits
  • “ask-me-questions-first” pattern
• Quality checks
  • factuality flags (“what could be wrong?”)
  • completeness checks (missing assumptions)
  • consistency checks (terminology, numbers)

Lab 2A (45 min): “Prompt deconstruction clinic” Participants diagnose 6 prompts (good/bad) and rewrite them using P-T-C-F-T. Lab 2B (60 min): “Prompting patterns workshop” Pairs practice 3 patterns across 2 tasks each:

1. Role + rubric: “Act as X, grade against Y”
2. Critic + reviser: “Generate draft, critique, revise”
3. Clarifying interviewer: “Ask 5 questions before drafting”

Deliverable: personal prompt library (minimum 8 prompts).

Bloom-aligned objectives

• Apply: GenAI to draft and refine communication artifacts
• Evaluate: outputs for accuracy, tone, and compliance
• Create: reusable templates and style guides

Topics

• Email workflows
  • drafting in multiple tones (executive, peer, customer)
  • summarizing long threads into actions + owners
  • converting a thread into a meeting agenda
• Documents and reports
  • first drafts from bullet notes
  • rewriting for clarity and conciseness
  • converting a long doc into an executive brief
• Presentation support
  • slide narrative outline from a document
  • speaker notes generation
  • converting insights into a “1-slide” story

Lab 3A (45 min): “Inbox triage sprint” Input: simulated email thread bundle. Outputs:

• executive summary
• action list with owners and due dates
• 2 draft replies (different tones)

Lab 3B (60 min): “From notes to proposal” Input: rough notes / meeting bullets. Output: one-page proposal with:

• background, objectives, scope, timeline, risks
• acceptance criteria
• 3 questions to confirm with stakeholders

Deliverable: proposal v1 + revision v2 after peer critique.

Bloom-aligned objectives

• Apply: extraction and synthesis from unstructured text
• Analyze: themes, sentiment, and contradictions
• Evaluate: confidence and verification needs

Topics

• Summarization patterns
  • executive brief
  • decision memo
  • “what changed?” delta summary
• Extraction patterns
  • key entities (people, dates, KPIs, commitments)
  • risks, blockers, dependencies
  • “claims vs evidence” separation
• Handling ambiguous inputs
  • missing context prompts
  • converting messy notes into structured records

Lab 4A (60 min): “Customer voice synthesis” Input: 40–60 short feedback snippets. Outputs:

• top themes
• representative quotes
• sentiment summary
• recommended actions + expected impact

Lab 4B (45 min): “Meeting-to-execution” Input: meeting transcript excerpt (sanitized). Outputs:

• minutes with decisions
• action items (RACI format)
• follow-up email draft

Bloom-aligned objectives

• Apply: GenAI to speed up spreadsheet work
• Analyze: trends and anomalies
• Create: reusable analysis prompts and checklists

Topics

• Natural-language-to-analysis workflows
  • KPI calculation guidance
  • formula generation and explanation
  • pivot-style summarization prompts
• Data quality and safety
  • identifying missing values/outliers
  • sanity checks for totals and units
  • avoiding sensitive data leakage
• Visual insight generation (tool-dependent)
  • chart selection rationale
  • narrative interpretation with caveats

Lab 5A (75 min): “Sales performance quick analysis” Input: sample sales dataset (products, regions, months). Tasks:

• compute revenue, margin, and growth
• identify top contributors and underperformers
• generate short insights memo with recommended next actions

Lab 5B (30 min): “Formula clinic” Participants bring 2 real formulas they struggle with; GenAI helps generate, explain, and test them using a checklist. Deliverable: “Analysis prompt pack” (minimum 6 prompts).

Bloom-aligned objectives

• Understand: core risk types and governance vocabulary
• Evaluate: outputs and usage for risk/compliance
• Create: personal playbook + team charter

Topics

• Key risk areas for enterprise use
  • privacy and data leakage (PII, contracts, credentials)
  • IP and confidentiality
  • bias and harmful content
  • hallucinations and decision errors
  • prompt injection (indirect instructions in documents/emails)
• Practical guardrails
  • data classification rules (“never paste” list)
  • verification checklist for factual claims
  • citation/traceability expectations
  • human approval thresholds
• Adoption enablement
  • role-based prompt libraries
  • change management patterns (champions, office hours, community)

Workshop (75 min): “Playbook builder” Outputs:

• Individual playbook (safe-use rules, best prompts, verification checklist)
• Team charter (approved tools, do/don’t, escalation, audit approach)

Final simulation (45 min): “A day in the AI-augmented office” Teams complete a timed sequence:

1. Summarize a thread
2. Draft a response
3. Create a mini brief
4. Produce spreadsheet insights
5. Flag risks and propose mitigations

Deliverables: final artifacts + short “how we prompted” explanation.