AAI101 ECA Guide: Workflow, Charts And Telegram Bot

Quick answer: The AAI101 ECA for July 2026 is a 100-mark, 70% assessment built from five tasks: an integrated AI workflow, an airline data visualisation, a Telegram insurance bot, research documentation and a reflection. Marks go to evidence, not description: screenshots of your own tool use, justified choices and honest limits. Start with the data handling and the screenshots, because they cannot be reconstructed the night before.

You have a Canvas submission due on Friday 6 November 2026 at noon, a brief that asks for screenshots of tools you may not have opened yet, and a marking rubric that rewards reasoning more than polish. This AAI101 ECA guide explains what the SUSS Generative AI Tools and Workflows assessment is really testing, how to sequence the work, and where marks leak away. If you want to see how a finished answer reads first, the AAI101 ECA sample is a reference extract (not for submission).

Need a second pair of eyes on your plan? Our SUSS assignment help team can walk you through the structure while you do the hands-on work yourself.

What the AAI101 ECA asks you to do

AAI101 is a code-free, Level 1 course at the Singapore University of Social Sciences (SUSS), so the ECA tests whether you can design and justify AI workflows, not whether you can program. The full brief is on the AAI101 ECA July 2026 question page. In short:

Referencing is APA. The tasks look separate, but the marker is reading for one thread: do you choose tools for reasons, check what they produce, and say where a human must step in?

Step-by-step AAI101 ECA guide

1. Open the airline dataset first and run your charts. Screenshots need real software, and this is the slowest part.
2. Record your Q4 evidence as you go. Take a screenshot of every AI interaction, and keep the first prompt and its improved version side by side.
3. Draft Q1 and Q3 on paper as diagrams before you write any prose. Boxes, arrows and decision points first; words second.
4. Build the budget table for Q1 and add it up. Show the total against $300 with a margin.
5. Write the two prompt templates for Q3 using the COOL structure exactly as your lessons defined it, with visible placeholders.
6. Write Q5 last, using your own screenshots and mistakes as material.
7. Check the word limits, then reference everything in APA.

Most students do the opposite and write Q1 for a week, then rush the data task. Don’t. The chart question carries the heaviest evidence requirement.

How the 2026 paper differs from the 2025 paper

The July 2025 AAI101 ECA is a useful comparison because it shows what stays constant. The same skills reappear in new scenarios.

The 2026 version asks for more proof of doing. Past-paper habits (describing what a tool could do) will not carry you.

Handling the airline data: the question markers are watching

The brief points you to the Kaggle Airline Passenger Satisfaction dataset and flags two traps. Service ratings recorded as 0 often mean the service did not apply to that passenger, and about 310 Arrival Delay values are missing.

The answer is a decision with a reason, not a silent fix. Treat 0 as “not applicable” and exclude it from averages for that service, or your satisfaction scores will fall for the wrong reason. For missing delays, say whether you dropped those rows (about 310, a small share of a large file) or filled them, and why that choice will not bend the conclusions. Then repeat the decision in the charts’ notes.

Chart choice follows audience. A loyalty team wants segments, an in-flight team wants service-level comparisons and ground operations wants delay patterns. Wilke (2019) and Knaflic (2015) both argue that the chart should answer one question for one reader, which is a useful test for each of your three visuals. If you need help interpreting the numbers, our big data analytics assignment help page explains the support available.

Mark-maximising checklist

• Every screenshot is yours, dated if possible, and referred to in the text.
• Each chart has a stated purpose, an audience and one sentence on why another chart type was rejected.
• The Q1 budget table adds up and stays under $300.
• The brand voice guide is within its word limit. Count it.
• The Q3 flows include an escalation route to a human for disputed claims.
• The data-protection section names the Singapore Personal Data Protection Act 2012 and applies it to the bot’s logging, not just a mention.
• Q4 names at least five sources, evaluated for authority and currency.
• Q5 uses your own course experience, with limits and ethics stated.

Stuck on a particular task? Our ECA assignment help service can give you a model answer and structured feedback to learn from.

Common mistakes that lose marks

The first is describing tools instead of using them. A paragraph about ChatGPT is worth little beside a screenshot with a prompt, an output and your correction of it. The second is a workflow with no human in it. Markers want checkpoints where a person approves, especially for money, claims or legal wording.

Third, a flat budget. Writing “ChatGPT, Make.com, Canva” with no costs fails the brief, which asks for a $300 check. Fourth, treating hallucination as a footnote. Generative models can produce fluent errors (Ji et al., 2023), so your quality-control step must say what gets checked and by whom.

Finally, generic reflection. “I learned a lot about AI” scores poorly. “My first prompt returned an invented policy clause, so I added a source-check step” scores well.

What students worry about most

Three worries are typical for a first AI-tools assessment: whether screenshots from a free tool are acceptable, whether AI use will be treated as misconduct, and what to do when a tool gives a poor answer. The brief itself expects tool interaction, so use the tools, record them, and state clearly where AI helped and where you changed its output. A poor answer is useful evidence if you document the iteration. For wider rules on AI use in Singapore coursework, read our post on using ChatGPT for assignments in Singapore, and check your course’s own guidance on disclosure. Mollick and Mollick (2023) describe structured prompting approaches that you can adapt and cite.

About the expert

AI and data topics are supported by Wei Chen, whose profile lists a Ph.D. in Computer Science from the University of Hong Kong and fields in artificial intelligence, machine learning, data mining and natural language processing.

Where to get help without losing your own voice

We provide model answers, guidance and feedback; you run the tools, take the screenshots and submit your own work. You can see how other students rate the service on our reviews page, and our AI and data specialists can review your plan. For report structure, our report writing help page covers formatting and clarity. If the deadline is close, start with the chart task today.

FAQs

Is the AAI101 ECA an individual assessment?

Yes. The July 2026 ECA is individual, worth 70% of the course and marked out of 100. Submission is through Canvas by 12 noon on 6 November 2026.

How many charts do I need for the airline question?

The brief asks for at least three visualisations, each with a screenshot and a justification. Choose charts that serve the three stakeholder groups rather than three versions of the same bar chart.

What should I do with the zero ratings in the dataset?

Treat them as a decision. Zeros often mean the service was not applicable, so excluding them from averages is usually defensible. State what you did and why, and keep it consistent across all charts.

How do I handle the missing Arrival Delay values?

Report how many rows are affected (about 310), then either remove them or fill them with a stated method. The marker wants the reasoning and a note on whether it changes your findings.

Which data protection law applies to the Telegram bot?

For a Singapore insurer, the Personal Data Protection Act 2012 is the natural choice. Link it to specific bot behaviour: consent before logging, limited fields in Google Sheets and a retention rule.

Can I use ChatGPT for the whole ECA?

The course expects you to use AI tools and document that use, but the analysis, judgement and reflection must be yours. Follow SUSS rules on disclosure and never present unchecked output as fact.

References

Airline passenger satisfaction [Data set]. (n.d.). Kaggle. https://www.kaggle.com/datasets/teejmahal20/airline-passenger-satisfaction

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

Knaflic, C. N. (2015). Storytelling with data: A data visualization guide for business professionals. Wiley.

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

Singapore University of Social Sciences. (n.d.). AAI101 Generative AI tools and workflows. https://www.suss.edu.sg/courses/detail/AAI101

Wilke, C. O. (2019). Fundamentals of data visualization: A primer on making informative and compelling figures. O’Reilly Media.