Creating a ChatBot

This assessment was completed as part of ADL6001 Engaging Learners in Digital Environments

About

After reviewing the LinkedIn course about phishing scams and designing a course to stay safe from scams, I've challenged myself to create a product.


But it's not a boring PowerPoint pack or an infographic - it's a chatbot.

https://landbot.online/v3/H-1519225-2TWJUHMW00EAZIE9/index.html

Meet the scamsafeBOT

Click on the icon above to try the scamsafeBOT.

The Problem

Scamwatch reported that in 2022 Australians have reported losses of over $569 million.


Research conducted by Commonwealth Bank (2022) found:

  • 3 in 5 Australians are more concerned about scams over the last 12 month
  • 57% of respondents find it harder to identify a scam as fake
  • Australians would receive 4.98 scam calls/ SMS/ email/ social media messages per week.

The User

''Any individual can fall victim to a scam. Scammers don’t discriminate and will target people of any gender, age-group or socio-economic background - though scammers often do target the most vulnerable people in society such as the elderly."


Australia New Zealand Bank, Broome, Hewett (2019)

The Australian Bureau of Statistics (2023) revealed 2 in 3 Australians over 15 years old were exposed to a scam in 2021-22.

The User Story

As an Australian exposed to scams,

I want to be educated on the topic of scams,

so that I can protect myself from scams.

As a potential victim of scam,

I want to know what I need to do,

so that I can limit my losses.

User Personas

Phil and Claire are in their 80s and are not comfortable with technology. They have some electronic devices like their mobile and an iPad.

Jake, a single dad, has recently connected with Amy online. They've fallen in love, but they haven't met in person.

Anne is currently a university student living away from home. She's in a sharehouse right now. She's also casual barista, so money is tight.

A close friend of theirs has recently been scammed. The couple are frightened and would like to learn about scams.

Amy has asked for some money so she can fly to Australia to meet him.

She's received an Instagram message with a great crypto investment opportunity.

Why scamsafeBOT?

The User’s Objective

The Developer’s Objective

Learn more about the different types of scams 

Educate the user about the different types of scams and provide useful resources and organisations 

Question the legitimacy of unsolicited contact 

Suggest potential red flags indicating it might be a scam 

User has been scammed

Refer the user to support services 

Technology

diagrams.net was used to draft and design the conversation flow. The events, activities and gateways were documented.

Landbot was the platform used to develop the chatbot. Landbot allows the chatbot to be intergrated into existing platforms like WhatsApp. The selling point was that Landbot required no coding experience and skills.

Laws of UX

Yablonksi's (2023) collection of laws of UX was considered and implemented in the design of the scamsafeBOT.


Miller's law

  • When the user wanted to learn about scams, less than 7 options were provided. For scam types outside of the range provided, the user was provided external resources.
  • At the end, the top 5 tips to avoid scams were provided.


Jakob's law

  • Research was undertaken to ensure the bot aligned with the user experience of chatbots created by other Australian banks.
  • The tested chatbots only provided button options, not free flow text


Serial Position Effect

  • The top 5 tips were intentionally placed at the end as users will best remember the last item.

User centred design

  • The bot was created following the user centred design process noted by Goel, Tanwar, Sharma (2022).
  • The concept of the product was first brainstormed (educating about scams).
  • Research was undertaken, with a focus on user research.
  • The user story was created (as a… I want... so that…).
  • The bot was designed and went through user testing.


Other UX design considerations

  • The user only gets button options, not free text, as the user’s objective was the learn, not ask for help on a topic.
  • Users prefer having options and expect buttons for common inputs (Budui, 2018).
  • Button interactions strengthen user experience (Haughland et al., 2022).
  • A high level introduction was provided but for an indepth explanation, the user is referred to the website.

Challenges

Challenge #1: Text heavy

  • Scam education can be content heavy.
  • Explaining every single characteristic of a scam type would disengage the user.
  • User testing provided feedback that the responses were text heavy.


Solution

An extremely high level introduction to each scam type was provided. Instead of giving just the textbook definition, the

answer provided educates the user on the deception technique used by the scammer and how the victim is manipulated into the scam. This allows the user to personalise their learnings to themselves.


Images were placed into the bot's responses to break up the text and provide examples for visual learners. The user is then given the opportunity to seek additional resources that will provide full definitions.


Challenge #2: Mapping the conversation paths

  • Although there are only 3 user objectives, each initial objective breaks up into many paths.
  • For example, the types of scams path then branches out to 6 different types of scams.
  • Similarly, the path of a potential victim of scam branches out depending on whether their details were compromised, and whether they have contacted their bank. This conversation has a different ending but can lead back to scam education.


Solution:

The process map drafted at the beginning was extremely helpful when developing the conversation paths. Where required, information was replaced by a link to an external resource. This reduced the number of some paths. When possible, the conversation would revert back to original block (learn about scams, report a scam, verify unsolicited contact).

User testing

User testing was completed by staff employed in fraud and scams at a bank

and students specialising in digital instructional design.

Feedback

Modification

Pilot name ‘scamBOT’ can be mistaken as a bot that will scam the user

Chatbot name has been modified to scamsafeBOT. The organisation can personalised the chatbot name to suit their organisation’s mascot.

Lack of referral to support services

The chat flow was altered to refer the user to support services like Scamwatch after each path (excluded scammed user yet to report to bank)

Lack of visuals

Incorporated images in text heavy blocks (e.g., education about the different types of scams)

Text heavy in the types of scam education segment

An additional task was created to provide useful resources instead of offering additional links at the end of each scam type. 

Future Potential Developments

  • Expand the chatbot to include more active learning components (e.g., quiz, interactive activities)
  • Create an automation to send the user resources directly to their inbox
  • Provide an option to speak to a human (e.g., when a victim is reported they have been scammed)

100/100 High Distinction

References


Australia and New Zealand Bank, Broome, M & Hewett, C 2019, INFOGRAPHIC: the Aussies falling victim to scams, bluenotes.anz.com, viewed 16 March 2023, <https://bluenotes.anz.com/posts/2019/12/scam-statistics-australia-retail-commercial-customers-losses>.


Australian Competition & Consumer Commission 2016, Scam statistics, Scamwatch, viewed 16 March 2023, <https://www.scamwatch.gov.au/scam-statistics?scamid=all&date=2022>.


Australian Competition and Consumer Commission n.d., Scams, Australian Competition and Consumer Commission, viewed 16 March 2023, <https://www.accc.gov.au/business/business-operations-and-costs/scams>.


Budiu, R, 2018, The User Experience of Chatbots, Nielsen Norman Group, viewed 15 March 2023, https://www.nngroup.com/articles/chatbots/


Commonwealth Bank of Australia 2022, Research shows the average Australian receives over 250 scam attempts a year, CBA Newsroom, viewed 16 March 2023, <https://www.commbank.com.au/articles/newsroom/2022/10/commbank-scams-and-fraud-research.html>.


Goel, G, Tanwar, P & Sharma, S 2022, ‘UI-UX Design Using User Centred Design (UCD) Method’, 2022 International Conference on Computer Communication and Informatics (ICCCI), Computer Communication and Informatics (ICCCI), 2022 International Conference on, pp. 1–8, viewed 16 March 2023, <https://discovery.ebsco.com/linkprocessor/plink?id=25d0e813-45d3-32da-888d-7817ac32c278>.


Haugeland, IKF, Følstad, A, Taylor, C & Bjørkli, CA 2022, ‘Understanding the user experience of customer service chatbots: An experimental study of chatbot interaction design’, International Journal of Human - Computer Studies, vol. 161, viewed 15 March 2023, <https://discovery.ebsco.com/linkprocessor/plink?id=85b35c2c-9b3b-31b5-8f05-b28215a61667>.


‌Yablonski, J 2023, Laws of UX, Laws of UX, viewed 16 March 2023, <https://lawsofux.com>.