Discussion prompt
20 minutes. Choose one AI proposal that is, or could be, put to you. In 150 to 200 words, answer the five questions (problem, data, evidence, risk, owner) and state what you would need to see before approving a pilot.
Private: nothing you type here is sent or stored.
This course has no discussion board yet, so the prompt is a private reflection.
Recap
- Three families: rules-based, predictive, generative; each has its own risks.
- A good use case is frequent, data-rich, error-tolerant and measurable.
- New cost = tool fee + work still done by hand + cost of correcting errors; saving = old cost − new cost.
- Ask five questions: problem, data, evidence, risk, owner; and beware headline accuracy when events are rare.
Module notes
Key terms. Rules-based automation: behaviour from written rules. Predictive model: learns patterns from data. Generative AI: produces new content. Baseline: a simple method to beat. Accuracy: the share of cases right. Value case: a comparison of costs with and without a tool.
Rules and formulas. Cost = items × minutes × cost per minute. New cost = fee + manual work + error cost. Saving = old cost − new cost.
Limits of this module. All figures are invented for teaching. Real value cases include set-up, training, support and benefits that are hard to quantify.
Next
Modules 2 to 8 are in preparation. Return to the course overview or the Executive Track.
References and further reading
- Specialty Digest: What Is Artificial Intelligence, Really?
- Specialty Digest encyclopedia: Artificial intelligence, Machine learning, Risk Scoring, Security Risk Assessment and Defining Project Scope and Managing Risk.
- All figures in Module 1 are invented for teaching and labelled as such; no product or market claims are made.