2021 OCT 07 (NewsRx) — By a News Reporter-Staff News Editor at Insurance Daily News — A patent by the inventors Dunn, Ryan (Grayslake, IL, US), Lundsgaard, Soren Okay. (Inverness, IL, US), Pham, Anthony (Northbrook, IL, US), filed on November 16, 2020, was revealed on-line on September 21, 2021, in response to information reporting originating from Alexandria, Virginia, by NewsRx correspondents.
Patent quantity 11128754 is assigned to Allstate Insurance Company (Northbrook, Illinois, United States).
The following quote was obtained by the information editors from the background info equipped by the inventors: “Aspects of the disclosure relate to processing methods. In specific, points of the disclosure relate to processing methods having a machine learning engine and machine learning datasets.
“In some situations, people could request to determine voice name classes with staff of an enterprise group (e.g., customer support consultant, brokers, and/or different staff). The enterprise organizations could route these requests to accessible representatives in a generic method based on consultant availability. In many situations, nevertheless, this may occasionally lead to sub-optimal name routing and thus could lead to sub-optimal communication between staff and shoppers.”
In addition to the background info obtained for this patent, NewsRx journalists additionally obtained the inventors’ abstract info for this patent: “Aspects of the disclosure present efficient, environment friendly, scalable, and handy technical options that handle and overcome the technical issues related to name routing.
“In accordance with a number of embodiments, a computing platform comprising no less than one processor, a communication interface communicatively coupled to the no less than one processor, and reminiscence could obtain, from a number of shopper gadgets, a lot of requests to determine a voice name session. The computing platform could determine, based on cellphone numbers equivalent to the a number of shopper gadgets, demographic info equivalent to customers of the a number of shopper gadgets. The computing platform could rating, utilizing a machine learning mannequin and based on the demographic info and performance data for every of a plurality of representatives, potential client-representative combos, the place: 1) the rating for every potential client-representative mixture signifies a chance of a profitable final result ensuing from establishing a voice name session between the respective shopper and consultant, and a pair of) scoring the potential client-representative combos is based on fall off charges for every of the customers of the a number of shopper gadgets, the place the autumn off charges point out modifications in consultant effectiveness with every of the customers of the a number of shopper gadgets as a maintain time will increase. The computing platform could alter the scores based on a historical frequency of interplay between every consultant and shoppers equivalent to the recognized demographic info. The computing platform could choose, based on the adjusted scores, no less than one of many potential client-representative combos. The computing platform could set up a voice name session for the no less than one of many potential client-representative combos.
“In a number of situations, the computing platform could gather consultant performance data indicating gross sales performance of every of the plurality of representatives throughout their communication with shoppers equivalent to completely different demographics data. The computing platform could practice, utilizing the consultant performance data, the machine learning mannequin, which can configure the machine learning mannequin to output the scores.
“In a number of situations, the adjusted scores could point out how possible a consultant of a specific potential client-representative pair is to make a sale to the corresponding shopper of the actual potential client-representative pair. In a number of situations, the demographics info could point out a number of of: gender, earlier accidents, earlier contact with the enterprise group, age, household composition, marital standing, schooling info, earnings info, or geographic info equivalent to the customers of the a number of shopper gadgets.
“In a number of situations, scoring the potential client-representative combos could additional embrace: 1) evaluating the variety of requests to determine a voice name session to a lot of accessible representatives; 2) based on figuring out that the variety of requests to determine a voice name session exceeds the variety of accessible representatives, scoring, utilizing the machine learning mannequin, combos of every accessible consultant with every consumer equivalent to the requests to determine a voice name session to determine an optimum consumer for every consultant to serve; and three) based on figuring out that the variety of accessible representatives exceeds the variety of requests to determine the voice name session, scoring, utilizing the machine learning mannequin, combos of every accessible consultant with every consumer equivalent to the requests to determine a voice name session to determine an optimum consultant to serve every consumer.”
The claims equipped by the inventors are:
“1. A computing platform comprising: no less than one processor; a communication interface communicatively coupled to the no less than one processor; and reminiscence storing computer-readable directions that, when executed by the no less than one processor, trigger the computing platform to: obtain, from a number of shopper gadgets, a lot of requests to determine a voice name session; determine, based on cellphone numbers equivalent to the a number of shopper gadgets, demographic info equivalent to customers of the a number of shopper gadgets; rating, utilizing a machine learning mannequin and based on the demographic info and performance data for every of a plurality of representatives, potential client-representative combos, whereby: the rating for every potential client-representative mixture signifies a chance of a profitable final result ensuing from establishing a voice name session between the respective shopper and consultant, and the rating for every potential client-representative combos is based on fall off charges for every of the customers of the a number of shopper gadgets, whereby the autumn off charges point out modifications within the chance of a profitable final result of a consultant with every of the customers of the a number of shopper gadgets as a time throughout which the shopper is put on maintain will increase; alter the scores based on a historical frequency of interplay between every consultant and shoppers equivalent to the recognized demographic info; choose, based on the adjusted scores, no less than one of many potential client-representative combos; and set up a voice name session for the no less than one of many potential client-representative combos.
“2. The computing platform of declare 1, whereby the reminiscence shops further computer-readable directions that, when executed by the no less than one processor, additional trigger the computing platform to: gather consultant performance data indicating gross sales performance of every of the plurality of representatives throughout their communication with shoppers equivalent to completely different demographics data; and practice, utilizing the consultant performance data, the machine learning mannequin, whereby coaching the machine learning mannequin configures the machine learning mannequin to output the scores.
“3. The computing platform of declare 1, whereby the adjusted scores point out how possible a consultant of a specific potential client-representative pair is to make a sale to the corresponding shopper of the actual potential client-representative pair.
“4. The computing platform of declare 1, whereby the demographics info signifies a number of of: age, household composition, marital standing, schooling info, earnings info, or geographic info equivalent to the customers of the a number of shopper gadgets.
“5. The computing platform of declare 1, whereby scoring the potential client-representative combos additional contains: evaluating the variety of requests to determine a voice name session to a lot of accessible representatives; based on figuring out that the variety of requests to determine a voice name session exceeds the variety of accessible representatives, scoring, utilizing the machine learning mannequin, combos of every accessible consultant with every consumer equivalent to the requests to determine a voice name session to determine an optimum consumer for every consultant to serve; and based on figuring out that the variety of accessible representatives exceeds the variety of requests to determine the voice name session, scoring, utilizing the machine learning mannequin, combos of every accessible consultant with every consumer equivalent to the requests to determine a voice name session to determine an optimum consultant to serve every consumer.
“6. The computing platform of declare 1, whereby the autumn off charges point out a tolerance for being on maintain for the customers of the a number of shopper gadgets.
“7. The computing platform of declare 1, whereby the reminiscence shops further computer-readable directions that, when executed by the no less than one processor, additional trigger the computing platform to: determine the plurality of representatives, whereby figuring out the plurality of representatives contains: figuring out representatives which are presently accessible to interact in a voice name session, and figuring out representatives that aren’t presently accessible, however that can possible be accessible inside a predetermined period of time.
“8. The computing platform of declare 1, whereby adjusting the scores additional contains adjusting, based on time info equivalent to the performance data.
“9. The computing platform of declare 1, whereby scoring the potential client-representative combos comprising: scoring the potential client-representative combos at a primary time and a number of second instances, whereby the a number of second instances are later than the primary time; and computing an general rating for every client-representative mixture by including the scores for the primary time and the a number of second instances.
“10. The computing platform of declare 9, whereby choosing the no less than one of many potential client-representative combos contains choosing the no less than one of many potential client-representative combos based on the general scores.
“11. The computing platform of declare 1, whereby the reminiscence shops further computer-readable directions that, when executed by the no less than one processor, additional trigger the computing platform to: obtain further performance data from the chosen no less than one of many potential client-representative combos; and replace, based on the extra performance data, the machine learning mannequin.
“12. A technique comprising: at a computing platform comprising no less than one processor, a communication interface, and reminiscence: receiving, from a number of shopper gadgets, a lot of requests to determine a voice name session; figuring out, based on cellphone numbers equivalent to the a number of shopper gadgets, demographic info equivalent to customers of the a number of shopper gadgets; scoring, utilizing a machine learning mannequin and based on the demographic info and performance data for every of a plurality of representatives, potential client-representative combos, whereby: the rating for every potential client-representative mixture signifies a chance of a profitable final result ensuing from establishing a voice name session between the respective shopper and consultant, the rating for every of the potential client-representative combos is based on fall off charges for every of the customers of the a number of shopper gadgets, whereby the autumn off charges point out modifications within the chance of a profitable final result of a consultant with every of the customers of the a number of shopper gadgets as a time throughout which the shopper is put on maintain will increase, and the rating for every of the potential client-representative combos is producing by scoring the potential client-representative combos at a primary time and a number of second instances, whereby the a number of second instances are later than the primary time; computing an general rating for every client-representative mixture by including the scores for the primary time and the a number of second instances; adjusting the general scores based on a historical frequency of interplay between every consultant and shoppers equivalent to the recognized demographic info; choosing, based on the adjusted scores, no less than one of many potential client-representative combos; and establishing a voice name session for the no less than one of many potential client-representative combos.
“13. The technique of declare 12, additional comprising: amassing consultant performance data indicating gross sales performance of every of the plurality of representatives throughout their communication with shoppers equivalent to completely different demographics data; and coaching, utilizing the consultant performance data, the machine learning mannequin, whereby coaching the machine learning mannequin configures the machine learning mannequin to output the scores.
“14. The technique of declare 12, whereby the adjusted scores point out how possible a consultant of a specific potential client-representative pair is to make a sale to the corresponding shopper of the actual potential client-representative pair.
“15. The technique of declare 12, whereby the demographics info signifies a number of of: age, household composition, marital standing, schooling info, earnings info, or geographic info equivalent to the customers of the a number of shopper gadgets.
“16. The technique of declare 12, whereby scoring the potential client-representative combos additional contains: evaluating the variety of requests to determine a voice name session to a lot of accessible representatives; based on figuring out that the variety of requests to determine a voice name session exceeds the variety of accessible representatives, scoring, utilizing the machine learning mannequin, combos of every accessible consultant with every consumer equivalent to the requests to determine a voice name session to determine an optimum consumer for every consultant to serve; and based on figuring out that the variety of accessible representatives exceeds the variety of requests to determine the voice name session, scoring, utilizing the machine learning mannequin, combos of every accessible consultant with every consumer equivalent to the requests to determine a voice name session to determine an optimum consultant to serve every consumer.
“17. The technique of declare 12, whereby the autumn off charges point out a tolerance for being on maintain for the customers of the a number of shopper gadgets.”
There are further claims. Please go to full patent to learn additional.
URL and extra info on this patent, see: Dunn, Ryan. Machine learning system for routing optimization based on historical performance data. U.S. Patent Number 11128754, filed November 16, 2020, and revealed on-line on September 21, 2021. Patent URL: http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=%2Fnetahtml%2FPTO%2Fsrchnum.htm&r=1&f=G&l=50&s1=11128754.PN.&OS=PN/11128754RS=PN/11128754
(Our studies ship fact-based information of analysis and discoveries from around the globe.)
