System and method for reducing wait-time in a multichannel contact center

1. A computerized-method for reducing wait-time in a multichannel contact center, said computerized-method comprising: in a system that is supporting inbound-interactions via a plurality of digital-channel-types, said system comprising: one or more processors, a plurality of queue-databases having a queue-database for each digital-channel-type and a memory to store the plurality of queue-databases thereon, for each query received by a customer via a digital-channel-type and not resolved by a bot, said one or more processors perform the steps of:

a) selecting a digital-channel-type from the plurality of digital-channel-types in a predetermined order:

b) operating an aggregator module on the selected digital-channel-type to: (i) identify one or more similar pending queries in a queue-database related to the digital-channel-type; and (ii) aggregate the received query and the identified one or more similar pending queries into a group-of-queries

c) operating a tracker module for a calculated parking period on the group-of-queries to identify a query-resolution and to send a notification as to the query-resolution to the customer via the digital-channel-type that the query has been received and to related customers of the identified similar pending queries each via a related digital-channel-type; and

d) when a total parking period of the plurality of digital-channel-types is lower than an expected wait-time of the received query and the received query hasn't been resolved, repeating steps a)-c) until a query-resolution has been identified or all digital-channel-types have been selected.

2. The computerized-method of claim 1 , wherein the predetermined order of the selecting of the digital-channel-type from the plurality of digital-channel-types is configured to have a digital-channel-type that the query has been received as first and then other digital-channel types in ascending order based on a calculated parking period related to each digital-channel-type.

3. The computerized-method of claim 1 , wherein the similar pending queries are queries which are having a Query Similarity Score (QSS) above a predefined threshold.

4. The computerized-method of claim 2 , wherein the total parking period is determined by aggregating the calculated parking period of each digital-channel-type in the plurality of digital-channel-types.

5. The computerized-method of claim 1 , wherein the identifying of the query resolution by the tracker module comprising: checking when at least one query of the group-of-queries has been resolved.

6. The computerized-method of claim 2 , wherein the received query is resolved via a digital-channel-type different from the first digital-channel-type by: (i) an agent of the digital-channel-type different from the first digital-channel-type; or (ii) a query of the group-of-queries in the digital-channel-tape different from the first digital-channel-type has been resolved.

7. The computerized-method of claim 3 , wherein the QSS of each pending query is calculated based on feature vectors which are generated from processing sentences in the received query and in the pending query, and wherein the processing is operated by using at least one of Neuro Linguistic Programming (NLP) techniques to convert each query into a feature vector: (i) Term Frequency—Inverse Document Frequency (TF-IDF) methods; or (ii) word2Vec technique; or (iii) Smooth Inverse Frequency (SIF).

8. The computerized-method of claim 7 , wherein when the NLP technique is TF-IDF, the QSS of each pending query is calculated by formula I:

s

i

=

s

LT

·

s

ST



s

LT



⁢



s

ST



(

I

)

whereby:

S i is Query Similarity Score,

S LT is a feature vector of the received query,

S ST is a feature vector of a pending query, and

||S LT || and ||S ST || are calculated based on formula II:



S



=

∑

i

=

1

i

=

n

x

i

2

where x i is an element from feature vector.

9. The computerized-method of claim 1 , wherein the expected wait-time of the received query is determined by formula III:

wait

⁢

‐

⁢

time

=

∑

i

=

1

N

-

1

Q

⋆

Ti

(

III

)

whereby:

T is a waiting period per query that is retrieved from the queue-database of the received query,

N-1 is a number of total queries which is ahead of queue from the received query, and

Q is a queue size which is a number of queries in the queue-database of the received query.

10. The computerized-method of claim 2 , wherein the parking period in each digital-channel-type is calculated by formula (IV):

parking period=wait-time/( C*Q )   (IV)

whereby:

C is a number of digital-channel types, and

Q is queue size of a queue-database related to the digital-channel-type, and

wait-time is the expected wait-time of the received query.

11. The computerized-method of claim 1 , wherein when the total parking period of the plurality of digital-channel-types is greater than the expected wait-time of the received query, the query is resolved via the digital-channel-type that the query has been received.

12. The computerized-method of claim 1 , wherein the identified one or more similar pending queries are queries of different customers.

13. The computerized-method of claim 1 , wherein the group-of-queries of each digital-channel-type is stored in a group database related to the digital-channel-type.