Daden Emerging Technology Seminar - Introduction to Chatbots & their uses within a business - Guest Speaker John Dowd from University of Wolverhampton talks about Chatbots.
2. WHO…
The Numbers….
150 Years
Applied Sciences
Art & Design
Business
8 Schools Education
4 Campuses
Health & Wellbeing
Law, Social Sciences & Communication
Sport, Performing Arts & Leisure
Technology
23,000 Students
3. WHO… WHO…
Strategic Plan, 2009 – 12
From fixed to flexible
From print to electronic
From physical to virtual
6. WHAT…
• Supportive role with regards to the Web Refresh project
• Development of an evidence based FAQ
• A ‘push’ to web site content not being fully utilised
• An adjunct to ASSIST (project)
• Alternate front-end/access method to content on LIS website
• Interactive vs Passive
LIS’s chatbot is an interface which will
provide real-time interaction and
navigation to resources and services
provided by LIS. It supplements the
ASSIST online chat service, by dealing
with routine information enquiries.
enquiries
Created: 24th October 08
8. WHAT… WHAT… WHAT…
Criteria – ‘Lightweight’, ‘Easily Controllable’
• Chatbot database to extract responses from FAQ on web site
• Chatbot database to be maintained by export of FAQ database
to file, then import into Chabot
• Chatbot configured to refer to relevant pages of Website
• Chatbot database to be ‘standalone’ with full configuration of
all questions/responses
13. Quantitative …Chatbot Versus Assist…
% of Correct & Incorrect Answers
48% Correct
52% Incorrect
14. Qualitative …Chatbot Versus People…
Did you find the chatbot helpful? 85% (score 4 out of 5)
Did you enjoy using the chatbot? 80%
Would you use it again? 60% yes/20% maybe
Have you used ASSIST before? 40% yes
Would you use the chatbot first in future? 100% yes
How many questions did the bot get right? = 16 out of 19
15. Additional from survey:
What did you like about it?
Appearance helpful
Ability to ask the question using natural lively
language, not just a key words like it is when enjoyable service
using search engines
Answer was told and written
instant answer
quick answers, links you straight to the page you need, points you in the right direction
reads out the answer as well as it being typed at the it would be useful to new students
side.
more personal without having to talk to a person, accessible from home, gives good information
16. What didn’t you like about it?
voice was annoying, is a white librarian going to be good for all students? does it come in other languages?
The lady at the start talking made me jump! *the bot said hello I am harry but she was a girl
Whether students would bother using the chatbox *sounded a bit rude when it did not understand the question
How students would find chatbox on the web
im dyslexic and i found it slightly hard to word questions to explain what i meant so it would understand.
When I first used the chatbot, I didn't expect it to open the webpage, but to give me an answer. I asked the question
and it said "See if this page contains information you require", I was confused, because I expected the information
will be given by a chatbot, not that I will be redirected to the webpage.
It advised me to contact Learning Centre Direct, but didn't open the webpage with contact details.
17. Other comments
It has some problems with understanding long sentences, but generally works
very well. It's much more user-friendly way of finding information on the
website than search engines.
.. .
My personal favourite
think it's a really good idea because it takes you to pages on the internet you
might never have come across or been aware of. im in third year and still
dont understand how ejournals work but it took me to a page that explained
everything.
19. Quantitative
350
300
250
28%
200
Correct
Incorrect
150
Answered correctly
100 Answered incorrectly
50
72%
0
January February March April May June
20. Patterns
40
ASSIST Queries (Comparative Review)
35 SEPT
250
OCT
30 NOV
200 DEC
25
JAN
FEB
20
150
MAR
APR
15
100 MAY
10 JUNE
JULY
50
5 AUG
00
January February March April May June
10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00
700 No. of chatbot chat sessions by time
600 45
40
500
35
400 30
No. of chats
25
300
20
200 15
10
100
5
0
0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00 12.00 13.00 14.00 15.00 16.00 17.00 18.00 19.00 20.00 21.00 22.00 23.00
January February March April May June
Time (24 hr clock)