MKTG2023 Chap.6 AI Personalisation and Service Script Prompts
AI Personalisation and Service Script Prompts
Three intelligences, three benefits
Huang and Rust order AI by difficulty. Mechanical AI automates routine tasks and delivers standardisation; thinking AI analyses data and delivers personalisation; feeling AI handles two-way, emotional interaction and delivers relationalisation. The framework runs through research, strategy and action, and actions feed data back into research.
The lecture adds a caution: true feeling AI does not yet exist, so emotional, high-value moments still call for people.
Matching AI to the flower
The seminar assigns an intelligence to each petal.
Routine petals such as billing and payment suit mechanical AI, data-rich petals such as information and consultation suit thinking AI, and emotional petals such as hospitality and exceptions need feeling AI or a human.
Prompts as service blueprints
Whatever an AI agent says at a touchpoint is its service script.
Prompt engineering writes that script with eight elements, from persona and brand voice to escalation rules and guardrails, and Australia’s AI ethics principles of transparency, fairness, privacy, accountability and contestability belong inside the prompt itself.
What this chapter covers
- 01
Mechanical, thinking and feeling AI
- 02
Standardisation, personalisation and relationalisation
- 03
The research, strategy and action cycle
- 04
Matching AI to the petals of the flower
- 05
The eight-element service-script prompt
- 06
The National AI Plan and AI Ethics Principles
Worked example · free
Rebuild a weak prompt for a library chatbot
- 1Persona: the library’s loans assistant.
- 1Brand voice: calm, plain and never accusatory.
- 1Context: the student’s loans, due dates and fine history, passed in as variables.
- 1Task: explain the fine and show how to pay or renew.
- 1Constraints: never waive fines or quote rules not supplied.
- 1Escalation: hand over to staff when the student reports illness, hardship or a disputed return.
- 1Output format: under 100 words with one next step.
- 1Guardrails: disclose that it is an AI assistant and ask for no health details.
Key terms
- Mechanical AI
- AI that automates repetitive, routine tasks such as classification or payments, delivering standardised, consistent outcomes.
- Thinking AI
- AI that processes data to reach conclusions, as recommender systems do, delivering personalised outputs.
- Feeling AI
- AI meant to recognise and respond to human emotion; the lecture notes that true feeling AI does not yet exist.
- Service Script
- Words an AI agent says to customers at a touchpoint, governed by a prompt that sets its voice, limits and escalation.
- Prompt Guardrails
- Explicit never and always instructions in a prompt, covering privacy, disclosure of AI and a ban on invented policy.
- Contestability
- AI ethics principle that customers can challenge an automated decision and reach a human who can review it.
AI Personalisation and Service Script Prompts FAQ
Which AI intelligence suits the hospitality petal?
Feeling AI, or a person. Hospitality is an emotional petal, and the lecture warns that current tools only imitate empathy by running thinking AI on emotional data. Customers pull back once a bot seems to lack empathy, so the course recommends humans or feeling-augmented tools for high-emotion, high-lifetime-value moments.
What is the difference between personalisation and relationalisation?
Personalisation tailors the output, such as content, offers and recommendations, and comes from thinking AI. Relationalisation personalises the relationship itself, responding to emotion and building rapport over time, and is associated with feeling AI. The lecture says relationalisation pays off when a customer’s lifetime value is high, and that the best experiences often need both.
What are the eight elements of a service-script prompt?
Role or persona, brand voice and tone, context, task, constraints, escalation rules, output format and guardrails. Together they make an AI agent consistent and on brand, and the guardrails carry privacy, disclosure that the customer is talking to AI and a ban on inventing policy.
How do Australia’s AI rules affect personalised marketing?
The National AI Plan relies largely on existing privacy, consumer and anti-discrimination law, and the AI Ethics Principles set the benchmark. In practice that means disclosing AI, having a lawful basis for personal data, auditing scripts for bias, naming a human owner for each script and giving customers a way to contest automated decisions.
Assessment move
Take one chatbot you have used and write down its weakest reply. Rebuild the prompt behind it with all eight elements, then check it against three of the AI Ethics Principles. Finally assign mechanical, thinking or feeling AI to each petal of the same service and circle the petal you would keep human.
Working through AI Personalisation and Service Script Prompts in MKTG2023? Sia is AskSia’s AI Marketing tutor — ask any MKTG2023 AI Personalisation and Service Script Prompts question and get a clear, step-by-step explanation grounded in how MKTG2023 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.