Back to results
US 12,400,238 B2Granted
Mobile intelligent outside sales assistant
AssigneeAccenture Global Solutions Limited
InventorShannon Copeland
Life of this patent
Estimating the term…
Abstract
Systems, methods, and applications for mobile intelligent outside sales assistance are provided. Embodiments include receiving speech for recognition of an outside sales call; converting the speech for recognition to text; parsing the converted text into outside sales triples; storing the outside sales triples in an enterprise knowledge graph of a semantic graph database; generating real-time outside sales insights in dependence upon the speech of the outside sales call and the stored outside triples in the enterprise knowledge graph; and presenting the real-time outside sales insights to an outside sales agent.
Claims16
A system, comprising:
- a non-transitory memory device storing machine-readable instructions;
- and
- one or more processing devices operably coupled to the non-transitory memory device, the one or more processing devices configured to, upon execution of the machine-readable instructions, perform operations comprising:
- receiving, by a speech-enabled device of the system, speech for recognition associated with an outside sales call, wherein the speech for recognition includes speech of a plurality of persons including a specific person, wherein the speech for recognition comprises one or more of:
- speech from the specific person obtained through a microphone on a headset, or
- speech that responds to a prompt executing on the speech-enabled device,
- digitizing, by an audio amplifier and a coder of the system, the speech for recognition,
- packaging the digitized speech in recognition request messages according to a voice over internet protocol (VOIP),
- transmitting the digitized speech for recognition to a voice server through a VOIP connection on a network, based on the recognition request messages,
- operating the voice server to return converted text representing the speech for recognition, wherein the digitized speech is converted to the text by passing the digitized speech to an Automated Speech Recognition (ASR) engine that is configured to:
- generate a speech feature vector (SFV) corresponding to each digitized word in the digitized speech by using frequency domain components of corresponding digitized word, wherein the SFV is defined by a subset of the frequency domain components,
- infer phonemes for each digitized word in the digitized speech based on the corresponding SFV using a language-specific acoustic model, and
- identify, using the inferred phonemes, a word in a lexicon, corresponding to each digitized word in the digitized speech,
- parsing, using a triple parser, the converted text to identify outside sales triples, wherein parsing the converted text into the outside sales triples comprises applying a pre-defined outside sales taxonomy comprising sets of words with defined semantics and an ontology comprising descriptions of concepts and properties in a domain, relationships between the concepts, and constraints corresponding to use of the relationships,
- storing the outside sales triples in an enterprise knowledge graph of a semantic graph database,
- generating, based on the outside sales triples stored in the enterprise knowledge graph and in response to a query, real-time sales insights associated with the speech of the outside sales call, and
- presenting the real-time sales insights on a display device of an agent associated with the outside sales call.
14 more elements
A method, comprising:
- receiving, by a speech-enabled device, speech for recognition associated with an outside sales call, wherein the speech for recognition includes speech of a plurality of persons including a specific person, wherein the speech for recognition comprises one or more of:
- speech from the specific person obtained through a microphone on a headset, or
- speech that responds to a prompt executing on the speech-enabled device;
- digitizing, by an audio amplifier and a coder, the speech for recognition;
- packaging the digitized speech in recognition request messages according to a voice over internet protocol (VOIP);
- transmitting the digitized speech for recognition to a voice server through a VOIP connection on a network, based on the recognition request messages;
- operating the voice server to return converted text representing the speech for recognition, wherein the digitized speech is converted to the text by passing the digitized speech to an Automated Speech Recognition (ASR) engine that is configured to:
- generate a speech feature vector (SFV) corresponding to each digitized word in the digitized speech by using frequency domain components of corresponding digitized word, wherein the SFV is defined by a subset of the frequency domain components,
- infer phonemes for each digitized word in the digitized speech based on the corresponding SFV using a language-specific acoustic model, and
- identify, using the inferred phonemes, a word in a lexicon, corresponding to each digitized word in the digitized speech;
- parsing, using a triple parser, the converted text to identify outside sales triples, wherein parsing the converted text into the outside sales triples comprises applying a pre-defined outside sales taxonomy comprising sets of words with defined semantics and an ontology comprising descriptions of concepts and properties in a domain, relationships between the concepts, and constraints corresponding to use of the relationships;
- storing the outside sales triples in an enterprise knowledge graph of a semantic graph database;
- generating, based on the outside sales triples stored in the enterprise knowledge graph and in response to a query, real-time sales insights associated with the speech of the outside sales call;
- and
- presenting the real-time sales insights on a display device of an agent associated with the outside sales call.
12 more elements
Non-transitory machine-readable storage devices storing machine-readable instructions, which upon execution cause one or more processing devices to perform operations comprising:
- receiving, by a speech-enabled device, speech for recognition associated with an outside sales call, wherein the speech for recognition includes speech of a plurality of persons including a specific person, wherein the speech for recognition comprises one or more of:
- speech from the specific person obtained through a microphone on a headset, or
- speech that responds to a prompt executing on the speech-enabled device,
- digitizing, by an audio amplifier and a coder, the speech for recognition,
- packaging the digitized speech in recognition request messages according to a voice over internet protocol (VOIP),
- transmitting the digitized speech for recognition to a voice server through a VOIP connection on a network, based on the recognition request messages,
- operating the voice server to return converted text representing the speech for recognition, wherein the digitized speech is converted to the text by passing the digitized speech to an Automated Speech Recognition (ASR) engine that is configured to:
- generate a speech feature vector (SFV) corresponding to each digitized word in the digitized speech by using frequency domain components of corresponding digitized word, wherein the SFV is defined by a subset of the frequency domain components,
- infer phonemes for each digitized word in the digitized speech based on the corresponding SFV using a language-specific acoustic model, and
- identify, using the inferred phonemes, a word in a lexicon, corresponding to each digitized word in the digitized speech,
- parsing, using a triple parser, the converted text to identify outside sales triples, wherein parsing the converted text into the outside sales triples comprises applying a pre-defined outside sales taxonomy comprising sets of words with defined semantics and an ontology comprising descriptions of concepts and properties in a domain, relationships between the concepts, and constraints corresponding to use of the relationships,
- storing the outside sales triples in an enterprise knowledge graph of a semantic graph database,
- generating, based on the outside sales triples stored in the enterprise knowledge graph and in response to a query, real-time sales insights associated with the speech of the outside sales call, and
- presenting the real-time sales insights on a display device of an agent associated with the outside sales call.
11 more elements
Dependencies are inferred from the claim text.