DE102021121943A1 - Method and answering machine for answering calls using artificial intelligence - Google Patents

Abstract

The invention relates to a proposed method for answering a call, a calendar, a contact list and a plurality of voice response modules being provided to a user's answering machine as respective text files and respective audio files with a respective assigned context. At least one context is assigned to a respective contact list entry having an identity and a telephone number of a caller. The voice response modules are spoken by the user. In the case of a currently incoming call, the identity of the caller is determined on the basis of the telephone number. Using an artificial intelligence function, based on the respective identity and the associated at least one context assigned to the respective contact list entry, the voice response module associated with the at least one context is displayed to the user as a text file in a selection menu as the respective entry. A time specification for the end of a currently ongoing appointment is read from the calendar and inserted into the respective entry. After selecting the respective entry, it is played back to the caller as a voice response. In addition, an answering machine and a computer program product are claimed, which allow the method to be carried out.

Description

The present invention relates to a method for answering a call using artificial intelligence. Furthermore, an answering machine is claimed, which allows the method to be carried out.

When a voice call is received, a personalized voice message can be assigned to a caller's identity on a communication device of a user. This can either be carried out automatically by a control unit implemented on the communication device or by the called user at the moment of the voice call.

The pamphlet US 2009/0059897 A1 discloses an assistance system that can link a personalized voice message with a caller ID for an incoming voice call. This can either be triggered automatically or by the user at the moment he receives the call. The voice messages can be combined with an AI function and the assistance system can schedule a call for a later time when the user is available again.

In the pamphlet WO 2014/164304 A1 describes a system that enables a communication device to respond to a caller according to the caller's intent based on an automatic determination of the caller's intent and the caller's permission level.

The US publication US 2018/0139587 A1 provides an artificial intelligence system using a machine learning algorithm that notifies of a call.

However, according to the prior art, the user can only select from ready-made answers or have them selected by the artificial intelligence in order to generate an answer that suits the caller.

Against this background, it is an object of the present invention to provide a method for answering a call, in which the caller is addressed more individually in a response message. Furthermore, an answering machine is to be claimed which allows the method to be carried out.

A method for answering a call is proposed to solve the above-mentioned task, wherein a user's answering machine is provided with a calendar, a contact list and a plurality of voice response modules as a text file and audio file with a respectively assigned context. At least one context is assigned to a respective contact list entry having an identity and a telephone number of a caller. The voice response modules are spoken by the user. In the case of a currently incoming call, the identity of the caller is determined on the basis of the telephone number. Using an artificial intelligence function, based on the respective identity and the associated at least one context assigned to the respective contact list entry, the voice response module associated with the at least one context is displayed to the user as a text file in a selection menu as the respective entry. A time specification for the end of a currently ongoing appointment is read from the calendar and inserted into the respective entry. After selecting the respective entry, it is played back to the caller as a voice response.

The selection can thus advantageously be made by the user directly at the time of the call, without acoustic communication being necessary. This is advantageous, for example, during an ongoing meeting in which the user is participating. The identity of the caller is shown to the user on the answering machine display. The method according to the invention can be carried out, for example, on a mobile terminal device, which thus performs a function as an answering machine and whose touchscreen is used as a display. If the user is also shown several entries with respective voice response modules for selection on the basis of several contextual assignments associated with the caller, the user can make a selection that appears suitable when recognizing the identity of the caller. The selection options can vary between more formal voice response modules that are rooted in a work environment, depending on the context assignment made beforehand. For example, a work colleague who is also a friend can be assigned both more formal language response modules and more jovial-sounding language response modules in the context. Depending on whether it can be assumed that other people in the professional environment are listening in to the call from the work colleague, the appropriate entry can then be selected in the selection menu.

Texts such as "I'm sorry", "I'll be in the next", "Minutes in a seminar", "and then call you back", "Minutes in a meeting and call you back via the video conference system", " Kind regards, Peter”, “See you soon, Peter Schmidt”. Depending on the context assigned to the caller, the voice building blocks together or supplemented by time information from the calendar. For example, if a family member calls, e.g. daughter "Klara", the text would read as follows: "I'm sorry Klara, I'll be in a seminar for the next 30 minutes and then I'll call you back, best regards, Peter" . If, on the other hand, a work colleague calls, e.g. "Mr. Müller", the selection text to be played could be as follows: "I'm sorry Mr. Müller, I'll be in a meeting for the next 30 minutes and will call back via the video conference system. See you soon, Peter Schmidt". The user advantageously also speaks his contact list. If there are unspoken text passages, e.g. with a new contact, the selection text is supplemented with a generated computer announcement.

The artificial intelligence function can, for example, include a machine learning algorithm, for example an artificial neural network, which is also continuously conditioned by the user during operation.

In one embodiment of the method according to the invention, selected numbers are spoken by the user to initialize the answering machine for specifying the time.

In another embodiment of the method according to the invention, further voice response modules are added to the plurality of voice response modules by the user. A context is taught to the artificial intelligence function for a respectively added language module.

In a further embodiment of the method according to the invention, the identity of the caller is formed by at least one piece of information from the following list: surname, first name, social relationship, address, authorization to make an appointment. "Social relationship" means a classification, for example, in categories such as "family", "partner", "circle of friends" or "colleagues". An "appointment permission" is a logical "true/false" value that is requested when uploading the contact list or creating a new contact list entry.

In yet another embodiment of the method according to the invention, the selection of the respective entry is made automatically after a predetermined period of time has elapsed. Five seconds, for example, is selected as the time span, on the one hand to give the user the opportunity to make a selection himself and on the other hand to react promptly to the caller.

In a continued further embodiment of the method according to the invention, a caller's intention is determined by means of the artificial intelligence function and assigned to a context for a voice response module. This can be achieved, for example, by previously agreed signal words on which the artificial intelligence function or the artificial neural network is trained.

In a still further embodiment of the method according to the invention, if the identity of the caller has an authorization to make an appointment, a callback is entered in the calendar for the end of the currently ongoing appointment.

Furthermore, an answering machine for answering a call is claimed, the answering machine being connected to a user's calendar and a user's contact list and comprising a plurality of voice response modules as respective text files and respective audio files with a respective assigned context. At least one context is associated with each contact list entry having an identity and a phone number of a caller. The voice response modules are spoken by the user. The answering machine is configured to determine the identity of the caller for a currently incoming call based on the telephone number. The answering machine is also configured to use an artificial intelligence function to display the voice response module associated with the at least one context as a text file to the user in a selection menu as the respective entry based on the respective identity and the associated at least one context assigned to the respective contact list entry. The answering machine is further configured to read out a time specification for the end of a currently ongoing appointment from the calendar, to insert this into the respective entry and, after selecting the respective entry, to play it back to the caller as a voice answer.

In one embodiment of the answering machine according to the invention, the answering machine is configured to allow the user to add further voice response modules to the plurality of voice response modules and to teach the artificial intelligence function a respective context for a respectively added voice module.

• • Abspeicherung einer Kontaktliste mit jeweiligen Kontaktlisteneinträgen, welche jeweils eine Identität und eine Telefonnummer eines Anrufers aufweisen;
• • Abspeicherung einer Mehrzahl von Sprachantwortbausteinen als jeweilige Textdateien und jeweilige Audiodateien mit einem jeweilig zugewiesenen Kontext;
• • Zuweisung des mindestens eines Kontexts zu dem jeweiligen Kontaktlisteneintrag;
• • Aufnahme eines Einsprechens der jeweiligen Sprachantwortbausteine durch den Nutzer;
• • Feststellen der Identität des Anrufers bei einem aktuell ankommenden Anruf anhand der Telefonnummer;
• • Anzeige des Sprachantwortbausteins, der mit dem mittels einer künstlichen Intelligenzfunktion anhand der jeweiligen Identität und damit einhergehend des dem jeweiligen Kontaktlisteneintrag zugewiesenen mindestens einen Kontext verbunden ist, als Textdatei in einem Auswahlmenü als ein jeweiliger Eintrag;
• • Auslesen einer zeitlichen Angabe für ein Ende eines aktuell andauernden Termins aus dem Kalender und Einfügen in den jeweiligen Eintrag;
• • Abspielen des jeweiligen Eintrags nach dessen Auswahl als eine Sprachantwort gegenüber dem Anrufer;
• • Aufnahme eines Einsprechens ausgewählter Zahlen für die zeitliche Angabe durch den Nutzer bei einer Initialisierung des Anrufbeantworters;
• • Hinzufügen weiterer Sprachantwortbausteine zu der Mehrzahl von Sprachantwortbausteinen und Anlernen eines Kontexts zu einem jeweilig hinzugefügten Sprachbaustein an die künstliche Intelligenzfunktion durch den Nutzer;
• • Automatisches Auswählen des jeweiligen Eintrags im Auswahlmenü nach einem Verstreichen einer vorgegebenen Zeitspanne;
• • Ermitteln einer Anruferintention mittels der künstlichen Intelligenzfunktion und Zuweisung zu einem Sprachantwortbaustein;
• • Eintragen eines Rückrufs in den Kalender für das Ende des aktuell andauernden Termins, falls die Identität des Anrufers eine Terminvereinbarungsberechtigung aufweist.

Furthermore, a computer program product with a computer-readable medium is claimed, on which a program code that can be executed on an answering machine is stored, wherein the program code, when executed on a processing unit, in particular as a component of the answering machine, causes the processing unit to do at least one of the following:

• • Storage of a contact list with respective contact list entries, each of which has an identity and a telephone number of a caller;
• • Storage of a plurality of voice response modules as respective text files and respective audio files with a respective assigned context;
• • assignment of the at least one context to the respective contact list entry;
• • Recording of the user speaking the respective voice response modules;
• • Establishing the identity of the caller for a current incoming call based on the telephone number;
• • display of the voice response module, which is connected to the at least one context assigned by means of an artificial intelligence function based on the respective identity and, as a result, the respective contact list entry, as a text file in a selection menu as a respective entry;
• • Reading out a time specification for an end of a currently ongoing appointment from the calendar and inserting it into the respective entry;
• • playing the respective entry after its selection as a voice response to the caller;
• • Recording a speaking of selected numbers for the time specification by the user when initializing the answering machine;
• • The user adding further speech response modules to the plurality of speech response modules and teaching a context to a respectively added speech module to the artificial intelligence function;
• • Automatic selection of the respective entry in the selection menu after a predetermined period of time has elapsed;
• • Determining a caller's intention using the artificial intelligence function and assigning it to a voice response module;
• • Calendaring a callback at the end of the current appointment if the caller's identity has appointment scheduling permission.

It goes without saying that the features mentioned above can be used not only in the combination specified in each case, but also in other combinations or on their own, without departing from the scope of the present invention.

QUOTES INCLUDED IN DESCRIPTION

This list of documents cited by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent Literature Cited

• US20090059897A1 [0003]
• WO 2014164304 A1 [0004]
• US20180139587A1 [0005]

Claims (10)

A method for answering a call, wherein a user's answering machine is provided with a calendar, a contact list and a plurality of voice response modules as respective text files and respective audio files with a respective assigned context, with a respective contact list entry having an identity and a telephone number of a caller, at least one context is assigned, with the voice response modules being spoken by the user, with the identity of the caller being determined on the basis of the telephone number in the case of a currently incoming call, with an artificial intelligence function using the respective identity and, consequently, the at least assigned to the respective contact list entry a context associated with the at least one context voice response module as a text file is displayed to the user in a selection menu as a respective entry, with a time from the calendar Specification for an end of a currently ongoing appointment is read and inserted into the respective entry, and after a selection of the respective entry it is played back to the caller as a voice response. procedure after claim 1 , wherein selected numbers are spoken by the user to initialize the answering machine for the time indication. Method according to one of the preceding claims, wherein further speech response modules are added to the plurality of speech response modules by the user and a context for a respectively added speech module is taught to the artificial intelligence function. Method according to one of the preceding claims, in which the identity of the caller is formed by at least one piece of information from the following list: surname, first name, social relationship, address, appointment authorization. Method according to one of the preceding claims, in which the selection of the respective entry is made automatically after a predetermined period of time has elapsed. Method according to one of the preceding claims, wherein a caller intention is determined by means of the artificial intelligence function and a context is assigned to a voice response module. A method according to any one of the preceding claims, wherein if the caller's identity includes appointment scheduling authorization, a callback is entered in the calendar for the end of the current appointment. Answering machine for answering a call, the answering machine being connected to a user's calendar and a user's contact list and comprising a plurality of voice response modules as respective text files and respective audio files with a respectively assigned context, with a respective contact list entry having an identity and a Has a caller's telephone number, at least one context is assigned, the voice response modules being spoken by the user, the answering machine being configured to use the telephone number to determine the identity of the caller for a currently incoming call, the answering machine being further configured to use an artificial intelligence function based on the respective identity and, as a result, the at least one context assigned to the respective contact list entry, the voice response module associated with the at least one context as a text file to the user in a selection menu as a respective entry, with the answering machine being further configured to read a time specification for the end of a currently ongoing appointment from the calendar and to insert it into the respective entry, and after selecting the respective entry this as a voice response to play back to the caller. answering machine after claim 8 , wherein the answering machine is configured to add further voice response modules to the plurality of voice response modules by the user and to teach a respective context to a respectively added voice module to the artificial intelligence function. Computer program product with a computer-readable medium on which a program code that can be executed on an answering machine is stored, the program code, when executed on a processing unit of the answering machine, causing the processing unit to carry out at least one of the following steps: • Storage of a contact list with respective contact list entries, which have an identity and telephone number of a caller; • Storage of a plurality of voice response modules as respective text files and respective audio files with a respective assigned context; • assignment of the at least one context to the respective contact list entry; • Recording of the voice response modules being spoken by the user; • Establishing the identity of the caller for a current incoming call based on the telephone number; • display of the voice response module, which is connected to the at least one context assigned by means of an artificial intelligence function based on the respective identity and, as a result, the respective contact list entry, as a text file in a selection menu as a respective entry; • Reading out a time specification for an end of a currently ongoing appointment from the calendar and inserting it into the respective entry; • playing the respective entry after its selection as a voice response to the caller; • Recording a speaking of selected numbers for the time specification by the user when initializing the answering machine; • The user adding further speech response modules to the plurality of speech response modules and teaching a context to a respectively added speech module to the artificial intelligence function; • Automatic selection of the respective entry in the selection menu after a predetermined period of time has elapsed; • Determining a caller's intention using the artificial intelligence function and assigning it to a voice response module; • Calendaring a callback at the end of the current appointment if the caller's identity has appointment scheduling permission.