AHMED MIO (JP)
SCHMITZ MICHAEL (DE)
HOSSAIN MOHAMMAD FAYADAN (BD)
MOSTAFA NAZIBA (BD)
| US9984377B2 | 2018-05-29 | |||
| US20160080567A1 | 2016-03-17 |
| Claims: 1. A method implemented in a computing system for delivering user specific content via a cellular or data network comprising: a. receiving, by an IVR system, inputs from a user using an arbitrary digital device; b. Collecting user data history generated from all previous interactions with the system, said history being from registration to accessing different IVR based services; c. generating, by the IVR system, individualized data-driven services to the user, said services comprising personalized advertisements; and d. Rendering these speech based contents comprising at least one of services and adverstisements through the IVR system. 2. The method of claim 1, wherein the digital is any device capable of rendering and recording speech data, connected to either a data or cellular network and comprises at least one of mobile phones (smartphones, feature phones, stationary phones), stationary phones, personal computer, laptops, tablets and IoT devices. 3. The method of claim 1, wherein the user specific data is collected as at least one of: transcribed contents of speech based interactions in machine readable format, analysis of the auditory properties of the speech itself, meta-data of the interaction (such as time and coordinates) and person related data (age, gender etc.) 4. The method of claim 3, wherein the user specific data is collected at at-least one of: a. a location data determined through the end user device b. coarse location information, caller history determined through the cellular or data network c. speech related data and historical data determined through the server system (all) 5. The method of claim 1, wherein the speech based contents of services and advertisements are delivered to the user through at least one of: a. the IVR prepending or appending the speech content to an IVR session initiated by the user b. the IVR pro-actively connecting to the user by calling over cellular or data network and playing the speech-based message once the user picks up the call. c. selection of corresponding data-driven service by the user during an IVR session 6. The method of claim 5, wherein the advertisers manage the user group to push speech based contents of services and the advertisements are delivered to the user based on at least one of: a. an electronic -platform for enabling the advertisers to select user group based on area, age, gender, income level, purchase nature, spending nature, frequency of call, duration of call, sentiment analysis, number of family members etc. b. an electronic -platform for enabling the advertisers to purchase number of minutes to advertise directly from the content providers or Telecom operator. c. an electronic-platform for enabling the advertisers to select the month, week, day, and time to push speech based contents of services and advertisements d. an electronic-platform for enabling the advertisers to purchase number of minutes 7. A system for delivering user specific content via a cellular or data network comprising: an arbitrary digital device; and an IVR system for: receiving inputs from a user using an arbitrary digital device; collecting user data history generated from all previous interactions with the system, said history being from registration to accessing different IVR based services; generating individualized data-driven services to the user, said services comprising personalized advertisements; and rendering these speech based contents comprising at least one of the services and the advertisements 8. The system as claimed in claim in claim 7, further comprising: an electronic platform for enabling the advertisers to manage the user group to push speech based contents of services and delivering the advertisements based on at least one of: a. the advertisers selecting user group based on area, age, gender, income level, purchase nature, spending nature, frequency of call, duration of call, sentiment analysis, number of family members etc. b. the advertisers purchasing number of minutes to advertise directly from the content providers or Telecom operator. c. the advertisers selecting the month, week, day, and time to push speech based contents of services and advertisements. 9. The system as claimed in claim in claim 7, wherein the digital device is at least one of mobile phones, stationary phones, personal computer, laptops, tablets and IoT devices. 10. The system of claim 3, wherein the user specific data could is collected through at- least one of: a. an end user device for determining location data b. cellular or data network for determining coarse location information and caller history) c. a server system for determining speech related data and the historical data. |
FIELD OF THE INVENTION
The present invention discloses a method of generating voice based, highly customized content for a mobile device user. More particularly the present invention relates to a model and framework for generating and exploiting mobile user contexts through their voice based input in a voice enabled solution for context sensitive advertising or automatic feedback generation.
BACKGROUND OF THE INVENTION
Traditional voice advertising on mobile wireless communication devices may be similar to more traditional advertising such as Television, Radio where a pre-recorded message is delivered.
Advertising that are tailored for specific user need are typically delivered via various platforms such as Web API, Mobile Application, SMS, USSD message by doing contextual analysis of user’s usage behavior and identifying their characteristics.
Similarly voice advertising is also delivered by generating a phone call from a server to the end user. Users became more reluctant to respond to a advertising that is generated via an inbound phone call from a server.
The computer implemented method and system disclosed herein, in general, relates to generation of revenue by a telecommunication Value Added Service (TVAS). More particularly, the computer implemented method and system disclosed herein relates to generating revenue by placing advertisements in phone calls and messaging service when the telecom consumers are subscribed to a VAS (Value Added Service) service sharing the generated revenue with the telecom consumers.
The existing methods of advertising on a mobile communication device, for example, a mobile phone typically involve in ways such as unsolicited phone calls or text messages from advertisers without the consent of the consumers. Furthermore, these modes of advertising offer no monetary benefits to the consumers and often conceived as breaching privacy. SUMMARY OF THE INVENTION
The invention relies on the analysis of user data collected during his/her activities through various means, including voice enabled applications. This data is generated by all transactions of that particular user and also by further analysis of the complete (anonymized) data of the whole user base. Furthermore, data can be gathered through surveys: A user’s reply to an IVR based survey will be stored and are available for further user specific analysis. User generated voice data will also undergo sentiment analysis to enrich the dataset with information about the user’s emotional response.
This analysis is based on artificial intelligence solutions, such as machine learning and allows highly customized information services to the user.
These services can be delivered to the user in two ways by the system, either on demand or pro-actively. In the on-demand scenario, the user would request a certain information via voice user interface, e.g. asking for individual sales predictions of a certain product for the coming month. If the system acts pro-actively, the user will receive contents without a specific request, for instance in the beginning of or during an ongoing voice interaction session or another purpose. In this case, the information could for instance be personalized advertisement or a customized reward/discount due to the usage pattern of that individual.
The invention enables third-parties to send highly customized information to a user or user group, based on contextual / machine learning based analysis of the aggregated data about the specific user through his previous usage, survey results and also related data of other users. This information can be presented pro-actively by the system and can consist of customized targeted advertising or individual rewards/discounts, which will be rendered through voice to the user. Other information can be directly inquired by the user, for instance sales predictions based on an machine learning analysis on all relevant data.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1, Fig. 2 and Fig. 3 are schematic overviews of system components of an exemplary embodiment of the presented invention; and FIG. 4 and Fig. 5 are flow charts with action sequences involved in the presented inventions
DETAIL DESCRIPTION OF PREFERRED EMBODIMENT(S)
Referring to FIG. 1, a diagrammatic view of the system organization for an embodiment of the presented invention is shown. A user (can be multiple) 101 can initiate a call to the open ended IVR system 104. The caller 101 can make the call using any device 102, which could be a smartphone, feature phone or stationary phone and the call will land on the system, which is running on cloud servers. Once the call has landed, a Natural Language Processing and Generation module 105 will translate spoken utterances of the caller into machine readable data and vice versa. The translations from human voice into structured data are used to connect to a service 106, execute some of its functions and send feedback to the user via generating speech through 105 and rendering it through the IVR system 104 to the user 101.
FIG. 2 shows a more detailed representation of the system in FIG 1 to include the flow of information (203, 204, 205), which is stored for later use on databases (201, 202). The components 102, 103 are omitted for brevity. The open ended IVR system generates Meta Data 203, which includes for instance, how often and at what times the user called up the system, the audio quality of the call, and call durations. The NLP 105.1 module further collects data, such as the interpreted input of the user 204 and the results of an analysis of this data, for example the inferred sentiment of the utterance. The active service 106 generates information 205 on a more abstract level, mostly related to the corresponding functions of the service, such as the goals/intents of the user and the entities of the specific context of the service. If, for instance, the user is connecting to a service that keeps records of a shop, in order to submit a transaction which states that she has sold a particular amount of a certain good, the system would e.g. be able to store how many of that item have been sold by this user for what price.
The system distributes the above mentioned types of data between a Global Knowledge Database 202 (GDB) and a User Specific Knowledge Database 201 (UDB), typically hosted centrally on a cloud server . The GDB would for instance store which amounts of certain items have been sold in total in that city, while the UDB would contain, when the user has made which exact transaction. FIG. 3 shows how the user 101 can access user customized services 302 tailored to his needs. The Open Ended IVR System 104 includes a Service Selector 301, which allows to chose a Service from the current Service Respository 106. Among them, User Customized Services 302 are offered, depending on User Specific Knowledge Database 201, which contains previously collected data about/from that particular user. For example, the user intends to access his virtual shop inventory to find out which amounts of a certain good she has on stock. The User Customized Service 302 will access the UDB to extract that information based on previously recorded sales and purchases to generate a reply through voice (105.2) via the IVR 104.
FIG. 4 visualizes the events and actions of the system when Ads are injected into the regular interaction flow with the IVR system. After identification of the user (401) the system decides whether the current session will be enriched with a personalized Ad 402. If this is the case, the system selects an Ad renders it (403) through the IVR, either as an pre recorded audio clip or through a TTS system. After that the suspended regular selection of a service (404) takes place. After the interaction with the service (405) completes, the user can decide (406) whether she wants to end the interaction or to restart, upon which it will again be evaluated whether an Ad will be presented. At this point (402) the system may always conclude not to present an Ad, which will directly lead to the Service Selection (404).
FIG. 5 presents the last option of delivering Ads to the user, through push-calls. In this case, the interaction session (504) is initiated by the system, by selecting a user that qualifies for being called (501) in order to present an Ad. The system initiates the call pro actively (503), and directly renders the Ad (505) to the user, once she is connected (504) to the system’s IVR. After that, the session is automatically closed (507).