For information and reference only – not to be submitted as your own work.
This AAI101 ECA sample extract shows how Question 1 of the SUSS AAI101 Generative AI Tools and Workflows End-of-Course Assessment (July Semester 2026) can be answered: an integrated AI system for a fictional Singapore property agency, Urban Nest Realty. It follows the structure of the AAI101 ECA July 2026 question page, and the planning logic is explained in our AAI101 ECA guide. Costs are illustrative and must be re-checked against current vendor pricing before you rely on them.
AAI101 ECA sample Q1(a): integrated workflow map, AI touchpoints and data flow
Urban Nest’s three problems (scattered marketing, slow enquiry handling and uneven agent training) share one root cause: each team holds client information in a different place. The workflow below therefore treats a single shared lead record as the spine, and every AI step reads from or writes to it.
| Stage | Trigger | AI touchpoint | Human checkpoint | Data passed on |
|---|---|---|---|---|
| 1. Lead acquisition | Monthly listing calendar | Generative AI drafts listing posts, neighbourhood guides and email copy from agent notes | Marketing lead approves every item | Campaign tag, source channel |
| 2. Enquiry handling | Web form or WhatsApp message | AI classifies the enquiry (buy, sell, rent) and drafts a first reply from approved facts | Agent reviews drafts about price, loans or legal terms | Lead record: contact, property type, budget band, consent flag |
| 3. Viewing scheduling | Client accepts a reply | Automation checks agent calendars and proposes three slots; reminders are generated | Agent confirms the slot | Viewing date, unit, agent |
| 4. Buyer education | Viewing completed | AI builds a short guide (grants, stamp duty terms, timelines) matched to the client’s profile | Compliance-aware agent checks facts against official sources | Content sent, client response |
| 5. Agent training | New listing, new regulation or feedback trend | AI turns closed-deal notes into role-play scenarios and quizzes | Training manager approves | Quiz scores, recurring client questions |
The data flow runs one way for facts and a loop back for learning. Facts move from the lead record into each AI prompt; client responses, agent edits and quiz scores move back into the record, where the marketing and training stages use them to improve the next round.
Q1(b) Tool recommendations, integration and the $300 monthly budget
Each tool is chosen for one job, and the integration layer is deliberately a no-code platform, because the agency has 35 staff and no developer.
| Tool | Role in the workflow | Illustrative monthly cost (S$) |
|---|---|---|
| ChatGPT paid plan, 2 seats | Listing copy, reply drafts, buyer guides, training scenarios | 54 |
| Make.com paid tier (about 10,000 operations) | Connects the web form, Google Sheets, Google Calendar and messaging | 16 |
| Canva Pro, 1 seat | Listing graphics and education one-pagers | 18 |
| Email marketing tool (about 500 contacts) | Campaign and nurture emails | 30 |
| WhatsApp messaging via a third-party provider | Enquiry replies and viewing reminders | 60 |
| Google Sheets and Calendar (assumed already licensed) | Lead record and agent availability | 0 |
| Contingency (15%) | Price rises, extra operations | 27 |
| Total | 205 |
The total leaves S$95 of headroom under the S$300 cap. These figures are planning allowances, not quotations: convert US-dollar prices at a stated rate and cite the pricing page you used.
Integration is the weak point of most answers, so state it plainly. A form submission triggers a Make.com scenario, which writes a row to Google Sheets, calls the AI model with the row’s fields and a fixed prompt, and saves the draft back to the same row for agent review. Nothing is sent to a client automatically when the message touches price, loan eligibility or legal advice.
Hire a Professional Essay & Assignment Writer for completing your Academic Assessments
Native Singapore Writers Team
- 100% Plagiarism-Free Essay
- Highest Satisfaction Rate
- Free Revision
- On-Time Delivery
Q1(c) Brand voice guide, quality-control checklist and feedback loops
Brand voice guide (under 150 words)
Urban Nest speaks like a well-informed neighbour who knows the property market: warm, plain and specific. We write in short sentences, address the client as “you”, and explain any term, such as ABSD or TOP, the first time we use it. We never promise prices, rental yields or approval outcomes, and we never pressure. Preferred words: “consider”, “typically”, “based on recent transactions”. Avoided words: “guaranteed”, “hot deal”, “once-in-a-lifetime”. Every figure names its source and date. We use Singapore English spelling throughout. Translations into Mandarin, Malay or Tamil are checked by a person before use. Humour is light and never at a client’s expense.
Quality-control checklist
- Every fact (price, size, tenure, grant rule) matches the listing sheet or an official source.
- No guarantee, no comparison that could mislead, no claim the agency cannot evidence.
- Tone matches the voice guide; avoided words are absent.
- Personal data in the draft is limited to what the client supplied for this purpose.
- A named person has approved the item and logged the date.
Generative models can state wrong details with confidence (Ji et al., 2023), which is why item 1 sits first and why price, loan and legal content always passes through an agent.
Feedback loops
- Weekly edit review: the marketing lead logs what agents changed in AI drafts, and the three most frequent edits become new prompt rules.
- Client signal loop: reply rates and viewing bookings per campaign tag show which messages work.
- Training loop: questions that agents could not answer in role-play become next month’s quiz topics.
Prompt iteration should be documented rather than assumed; Mollick and Mollick (2023) show how structured prompts with explicit roles and constraints improve usable output.
Governance note
Because the system handles personal data, it must respect Singapore’s Personal Data Protection Act 2012 (2020 Rev. Ed.), including consent and the Do Not Call provisions for marketing messages. The Personal Data Protection Commission (2024) has also issued guidelines on personal data in AI recommendation and decision systems, and the Infocomm Media Development Authority and AI Verify Foundation (2024) have published a governance framework for generative AI. Both support the human checkpoints in the workflow above.
How this extract earns marks
| Marking area (Q1) | What the extract does | Marks available |
|---|---|---|
| Workflow map, touchpoints, data flow | Five stages tied to the three business areas, a human checkpoint at each, and a shared lead record | 10 |
| Tools, integration, budget | Role per tool, named integration path, itemised total checked against S$300 | 8 |
| Brand voice, quality control, feedback | Voice guide within the word cap, five-point checklist, three loops | 7 |
The remaining parts of the assessment (data visualisation, the Telegram bot, research documentation and the reflection) need their own evidence, including screenshots of your own tool use. For a related chart-based model answer, see the ANL201 data visualisation sample.
If you want subject-specific support with a similar brief, our artificial intelligence assignment help team can guide you through structure and referencing, and the wider SUSS assignment help page lists other modules we cover. You remain responsible for your own submission and your university’s rules.
About the expert
This topic is handled by Wei Chen, who holds a Ph.D. in Computer Science from the University of Hong Kong. The profile lists artificial intelligence, machine learning, data mining and natural language processing as fields of expertise, with 1,902 orders completed.
References
Infocomm Media Development Authority & AI Verify Foundation. (2024). Model AI governance framework for generative AI: Fostering a trusted ecosystem.
Ji, Z., Lee, N., Frieske, R., Yu, T., Su, D., Xu, Y., Ishii, E., Bang, Y. J., Madotto, A., & Fung, P. (2023). Survey of hallucination in natural language generation. ACM Computing Surveys, 55(12), Article 248. https://doi.org/10.1145/3571730
Mollick, E., & Mollick, L. (2023). Assigning AI: Seven approaches for students, with prompts. SSRN. https://doi.org/10.2139/ssrn.4475995
Personal Data Protection Act 2012 (2020 Rev. Ed.) (Singapore). https://sso.agc.gov.sg/Act/PDPA2012
Personal Data Protection Commission. (2024). Advisory guidelines on the use of personal data in AI recommendation and decision systems. Government of Singapore.
Singapore University of Social Sciences. (n.d.). AAI101 Generative AI tools and workflows. https://www.suss.edu.sg/courses/detail/AAI101
