S

17 atur
2.1 da
6.2 y, M
54
.6 ay 7
by ,
Th 2022
yV
IP h u h from
oa
ng IP
Search Experience to 7
2 2
, 20 u ho
m
fr ng an
o
a an
h
ay hy V
Ad Usefulness:
tu
Sa 6.25
r d
,
4
M
ay 6 by
.
T

Understanding Intent
172
. 1

S
17 atur
March
2 .16 day2022
.25 , M
The Ads 4.6 Quality
ay
by 7, 2 Team
Th 022
yV
u h from IP
oa
ng IP r o m anh
f
an
h 2 0 22 oang
, h
a y 7 y Vu
a y , M y Th
d b
a tur 254.6
S 6.
.1
172

TELUS International confidential

 The Task

The Search Experience to Ad Usefulness task type17will

2.1 urdapresent you with
Sa
t
6.2 y, M
an ad and an accompanying worksheet to guide you in54.6assigning
ay
by 7, 2
Th 022
a final
yV
u h from
rating. 2 f
IP h
rom g an
n
oa
ng IP
a
2 a nh
7 , 20 u ho
M ay hy V
, T
d ay 6 by
As you assess the ad, the worksheet will ask corresponding questions and
tu
Sa 6.25
. 1
r 4 .

172
recommend
Sa
a rating based on your responses.
17 tur
2.1 da
6.2 y, M
54
.6 ay 7
by ,
Th 2022
The worksheet provides an overview of how well the intent andIPad
yV
u h from
oa h
ng IP rom g an
intersect, and each response can alter the final result.
an
h
7 , 2 u
f
022 hoan
y V
y , Ma Thy
da b y
a tur 254.6
S 6.
.1
172

TELUS International confidential 2

 Understanding User Intent

S
17 atur
Before continuing to assess the ad 2.1 da
6.2 y, M
54
itself, it is important to evaluate and .6 ay 7
by ,
Th 2022
yV
research the query IP h in order to u h from
oa
r o m an ng IP
understand202what
2 f ang the user wants.
ho
an
h
7, u
M ay hy V
T
r d ay, 6 by
.
Here we can see the first question
tu
Sa 6.25
1
4
.
172
posed by the worksheet: “After
S
researching
17 atur
2.1 da the query, do you
6.2 y, M
understand
54
.6 ay 7
b ,
what the user wants?”
yT 2
hy 022
Vu fro
ho m I IP
an P m anh
Researching the query is vital, even if
ga r o
f ng
22the
nh In the above example, we can see , 2 0 ho user’s query
a and
you think you understand it in its approximate geographic location,
y 7 V u
Ma Thy as well as the
y , y
entirety. Many words and phrases u rda 4.6 b
worksheet question andS flags
a t
1 6.2
5 for the query.
.
can have alternative meanings we 172
were unaware of!

TELUS International confidential 3

 Understanding User Intent

S
17 atur
Depending on what the user’s query 2.1 da
6.2 y, M
54
is, you may need to perform different .6 ay 7
by ,
Th 2022
yV
amounts of research
IP h to understand it. u h from
oa
m
fro ng an ng IP
2 2 a an
7 , 20 u ho h
ay hy V
Some queries will have a clear and
M
, T
r d ay 6 by
.
explicit intent, while others may have
tu
Sa 6.25
1
4
.
172
multiple likely intents.
S
17 atur
2.1 da
6.2 y, M
Research may also reveal if there are
54
.6 ay 7
by ,
Th 2022
possible but unlikely interpretations,
yV
u h from After research, we can see that this IP query:
oa
ng IP m anh
or even entirely unrealistic an [Woodford Reserve Double Oaked 0
f r o
22 oangBourbon
h , 2 h
interpretations of the query. It is Whiskey] is reasonably straightforward: a y 7 y Vu
the likeliest
a y , M y Th
b
important to be aware of the potential r
u 54.d
user intent is to find aSatspecific 2
6
product. With this
. 16.
meanings of a query as you assess understanding, we will 17 2
be able to assess the ad so
the ad. our response on the worksheet should be “Yes, I
understand at least generally and can rate ads
TELUS International confidential for this query.” 4
 “After Researching the Query, Do You Understand What The User Wants?”

Rateable Queries - click the link for Search Results S

17 atur
2.1 da
6.2 y, M
54
“Yes, I understand at least generally and can rate ads for this query.” .6 ay 7
by ,
Th 2022
yV
IP h u h from
oa
m
fr ng an
o ng IP
● plu ize wimwear: Query is misspelled
2 2
, 20 u ho
a ● 205/65 r16: After research, we can see an
h
7
M ay hy V
but intent is still apparent after viewing this user has a clear intent to find a
, T
r d ay 6 by
.
the search results.
tu
Sa 6.25
1
4 specific model of tire.
.
172

●172.S1atstump
urd removal: Multiple likely intents ● ten sleep: This user is most likely looking
6.2 ay, M
(either
54
.6 ay 7 information- or service-seeking)
by ,
for information about a town in Wyoming,
Th 2022
yV
u h from with a less likely (but still possible) IP h
intent
oa
ng IP
● brazil economy: Query is general but to find one of many local m an
fro businesses with
an
h 2 0 2 oang
2
, h
the intent to find information is clear this name. a y 7 y Vu
, M Th
u r day .6 by
t 4
Sa 6.25
.1
● moma: Query appears nonsensical but ● clpk232a 181: This query looks unclear,
172
research shows it is a well-known but research shows that this user wants a
museum. specific model of drill.
TELUS International confidential 5
 “After Researching the Query, Do You Understand What The User Wants?”

Unrateable Queries - click the link for Search Results S

17 atur
2.1 da
6.2 y, M
54
“No, I don’t understand well enough to rate ads for this query.” .6 ay 7
by ,
Th 2022
yV
IP h u h from
oa
m
fr ng an
o ng IP
2 2 a an
, 20 u ho
● how to st: Query appears incomplete
7 ● near me: We can see that this user h
M ay hy V
, T
(but is still present, so the “Query is
tu r d ay 6 by
4 . wants something near them, but even if
Sa 6.25
Missing” flag is not appropriate)
172
. 1 we have their location we still have no
S way of knowing what it is they want.
17 atur
2.1 da
● iprenasion: This word is either
6.2 y, M
54
.6 ay 7
incomprehensible or in a foreign
by ,
Th 2022
yV
● james: This query could refer to any
u h from
language. Since it is not clear that this is
oa
ng IP
number of individuals, locations,
r o
IP
m anh or other
f g
an
a foreign language query, the flag is not
h things and is too general
, 2 022 hoato n
rate.
7 u y V
appropriate. y , Ma Thy
da b y
a tur 254.6
● video: Query is too vague to identify an
S 6.
.1
172
● flondip: This query is nonsensical in any intent, even after research. Does the user
language. want a product, information, or media?
We have no way of knowing.
TELUS International confidential 6
 Tips
● Language is important! Take the entire query into account - the user entered it that
Sa
tu
way for a reason. Whether something is singular or plural,172or
.16 rincludes
da
.25 y, M
any additional
4.6 ay
information such as brand or location can be very useful in narrowing by 7, 2
Th 022
down the
yV
likeliest user
IP hintent(s). u h from
o I
m an P
2 fro ng an ga
2 a nh
7 , 20 u ho
ay hy V
● Always use the search results provided to get a complete picture of the intent. You
M
, T
r d ay 6 by
.
don’t need to become an expert, but do enough research to understand which
tu
Sa 6.25
1
4
.
172
possible intents are or aren’t likely.
S
17 atur
2.1 da
6.2 y, M
● If a query doesn’t load, is in the wrong language, or is unexpectedly pornographic, it
54
.6 ay 7
by ,
Th 2022
should be flagged as such using the relevant flag.
yV
u h from IP
oa h
ng IP f rom g an
an
h , 2 022 hoan
● Each question on the worksheet has its own instructions and examples, so if you are
M
7
ay hy V
T
u
,
u r day .6 by
unsure of how to answer a question start by reviewing these. If you are still unsure
t 4
Sa 6.25
.1
after checking the instructions and applying your experience, you can always reach
172
out to us at adsquality@telusinternational.ai with screenshots of any relevant
information and we will be happy to help!
TELUS International confidential 7
 S
17 atur
2.1 da
6.2 y, M
54
.6 ay 7
by ,
Th 2022
yV
IP h u h from
oa
m
fr ng an
o ng IP
2 2 a an
7 , 20 u ho h
M ay hy V
, T
ay 6 by
tu
Sa 6.25
. 1
r d
4 .
Thank You!
172
S adsquality@telusinternational.ai
17 atur
2.1 da
6.2 y, M
54
.6 ay 7
by ,
Th 2022
yV
u h from IP
oa
ng IP r o m anh
f
an
h 2 0 22 oang
, h
a y 7 y Vu
a y , M y Th
d b
a tur 254.6
S 6.
.1
172

TELUS International confidential