Room 2202, 22/F,
Mega Trade Centre,
1 Mei Wan Street,
Tsuen Wan, New Territories
Tel: 91375571

Invoice Number h-i20200635328
Invoice Date 29.06.2021
Total Due $59,000.00

C/ include this Summary of Works(general work from MINT)

1. Prepare detailed project plan, project schedule and implementation schedule. Monitor project progress jointly with user and provide periodic progress report to the Project Team.


2. Prepare User Requirement Specification/Functional Specification/and related documentaiton


3. Conduct storyboards and walkthrough for some cases


4. Perform system development and program construction for individual functions and perform Unit Test.


5. Develop comprehensive test plans and test cases in consultation with user and conduct user acceptance test jointly with user according to the test plan and test cases. Document all findings of the user acceptance test. Assist users in carrying out the UAT and perform UAT refinements to fix all problems / issues identified.


6. Develop comprehensive test plans and test cases in consultation with and conduct user acceptance test jointly with user according to the test plan and test cases. Document all findings of the user acceptance test. Assist user users in carrying out the UAT and perform UAT refinements to fix all problems / issues identified.


7. Perform performance tuning to meet the Performance Requirements as stated in this tender.

Provide troubleshooting and prompt support services to user especially during the initial production run of the system.

Provide training to end users and technical staff, including the system Administrator.


8. Provide all the source codes (including programs and scripts), and brief us on the source codes, system internal structures, processing logic and processing flow, etc.


9. Provide free warranty services during the Warranty Period.

Project timeline:

1 month development setup & data collection from users

1 month data training and adjustment of the chatbot

1 month beta testing for users

1 month bug fix and adjustment

1 month launch & project completed

Development Scope:

chatbot development with diaglow flow

general support


1. bug fix for questions and answer


2. adding within 10 questions


3. fb chatbot update on color or popup setting


4. api bug checking


5. office hour call support for users


all of chatbot-requirement.txt feature are supported? If not, any alternative or else?


Just to say the match rate is about 85% as some times the user may have typo and the questions was not proper which may confuse the machine learning so we wouldn’t commit as 99%


The chatbot should be able


1. to answer questions correctly based on the pre-defined questions and answers; - yes


2. to answer questions rationally or provide alternatives when no existence of exact-match answers; - yes


3. to answer simple questions rationally; e.g. How are you? What’s your name kind of answer; - yes


4. to collect the questions and answer from users and save them in database for model optimization or


statistical purposes; yes


5. to collect the metrics from users on whether the chatbot answers are correct or appropriate corresponding


to the users’ questions. For example, the chatbot will ask follow up question after an answer from the chatbot like, “Does this likely answer your question?” and then the user can answer by selecting the star(s) from one to five; - yes, will be adding a questions after they finish the conversation and ask them to rate


6. to let administrator to define ‘branch questions or answers’ from the backend in order to limit the scope of answers. For example, “CILL - “Which subject do you want to ask? Science or Maths?”. Please select the programme which you want to know more as answer; yes


7. to provide emulators for testing on web and mobile environments before pushing the model to UAT or production. For example, we can test the chatbot messenger behaviour after questions and answers were added in the testing model; can create an UAT chatbot on Facebook for testing before switching to production


8. to be implemented on various platforms like website, mobile devices like iOS and Android; depends on the language if they use native or hybrid, but mainly we can integrate with inappbrower


9. any jdk to integrate with native mobile app(IOS / Android), how? Will be install the fb messages sdk if the client agree with fb chatbot


Custom CMS to support Q&A updating by administrators.


1. To set up many-to-many Q&A (in English) meaning that a set of different forms of a question maps to a set of different forms of the answer or better solution, to cater the variation of enquiries and provide the diversity of the answers. Yes


2. To set up different forms of selection menus for signal or multiple options. Yes


3. To set up branch questions and answers to be triggered in different scenarios. For example, if the user


asked a question, but the chatbot has no such question and answer, the user will be redirected to similar


question that exists. Or showing we do not understand and ask if this prompted question can help them


4. To set up follow-up questions, e.g. “Does this likely answer your question”, “Do you like this chatbot”


* for the 4 items above, the set up can be performed by manual input or data feed, e.g. csv. yes


5. To view the users’ input and responses in a meaningful way (a kind of visualization). For example,


a. the top ten frequently asked questions within a particular period, yes


b. the ranking of the perfect or imperfect match questions and answers, yes


c. the statistical data regarding whether the chatbot can answer users’ questions correctly. It will be shown on the backend of the machine learning logic but for being wrong or correct shall need user to give the comment


UI Frontend (for all users):


1. Able to accept but not limited input text, options in the form of radio or checkbox. Only text and selection of the answer it would become a form if having chatbox


2. Able to output but not limited text, text-to-speech, images, links, different forms of selection menu. Text, images, links is fine for the text to speech is just using the phone or the computer input


3. Spelling correction - do not support


4. Design of the chatbot (image, theme, color scheme) including PolyU/ELC logo in the title bar of the


chatbot agent - yes




Traction logs:


Record all enquiries per session with timestamp - the complete questions and answers should be logged by each session (i.e. multiple conversations between the chatbot and the user within a session) for chatbot model optimization and statistical purposes.

- yes


first payment - 24000

Final payment - 59000 * this payment

Total payment - 83000

Hrs/Qty Service Rate/PriceAdjustSub Total
1 chatbot integration with portal

Integrate with web & app
web/app integration
Install on the client web portal
Web chatbot prompt function
Web notification features
Email notification from chatbot
Backend reply portal if client reply by human

1 Training service

Training - 3 hours
free warranty for 3 months

Any security and privacy is included?

1 Chatbot- Setup

chatbot setup fb
chatbot setup
API setup
auto-reply setup
developer setup on Facebook bot

1 Chatbot- machine learning development

Logic Adapters control
Storage Adapters
Data Training
Words Filters
Frequently Asked Questions
Machine Learning

Sub Total $59,000.00
Tax $0.00
Total Due $59,000.00

*Cheques should be crossed and made payable to “Hazedawn Limited”
*Company bank account - Bank of China - 012-720-0-003526-8

Hazedawn Limited
FPS ID: 160462198