Abstract: In some implementations, a system may receive a submission associated with an innovation associated with an entity. The system may analyze, using a first machine learning model, the submission to identify a classification associated with the innovation. The system may analyze, using a second machine learning model in association with the classification, content of the submission to identify a characteristic of the innovation. The system may determine, using a third machine learning model and based on the characteristic, a ranking of the innovation relative to individual innovations in the subset of innovations. The system may determine, using a fourth machine learning model, an impact score associated with the innovation. The system may perform, based on the impact score satisfying a threshold, an action associated with a project involving the innovation.

Abstract: A method for disoriented cell configuration includes determining a first number of misaligned sectors in a first set of disoriented cells of a first set of cells, generating data of a second set of disoriented cells based on the first set of disoriented cells or the first number of misaligned sectors for in the first set of disoriented cells, determining a first set of cross-feeder cells, a first set of sector swap cells or a first set of cyclic swap cells, and changing a configuration of an antenna of a disoriented cell of the second set of disoriented cells. Changing the configuration of the antenna includes changing a connection of a set of cables of an antenna of a first cross-feeder cell, or changing a deployed azimuth of an antenna of a sector in the first set of sector swap cells or cyclic swap cells.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating a win prediction for an opportunity. In some implementations, a server receives data representing an opportunity. The server obtains historical data from a data repository based on the received data representing the opportunity. The server generates standardized feature vectors based on non-standardized data, the non-standardized data comprising (i) the received data and (ii) the obtained historical data. The server filters the standardized feature vectors. The server generates a win percentage based on the filtered feature vectors using classifiers. In response to generating the win percentage for the opportunity, the server generates a plurality of data insights using the classifiers and the generated win percentage, wherein the plurality of data insights describe one or more data predictions for pursuing the opportunity.

Abstract: An AI-based process monitoring system access a plurality of data sources having different data formats to collect and analyze KPI data and shortlist KPIs that are to be used for determining the impact of automation of an automated process or sub-process. Information regarding an automated process is received and KPIs associated with the process and sub-processes of the process are identified. The identified KPIs are put through an approval process and the approved KPIs are presented to a user for selection. The user-selected KPIs are evaluated based on classification, ranking and sentiments associated therewith. The evaluations are again presented to the user along with a set of questionnaires wherein each of the questions has a dynamically controlled weight associated therewith. Based at least on the weights and user responses, a subset of the evaluated KPIs are shortlisted for use in evaluating the impact of process automation.

Abstract: In some implementations, a system may receive a submission associated with an innovation associated with an entity. The system may analyze, using a first machine learning model, the submission to identify a classification associated with the innovation. The system may analyze, using a second machine learning model in association with the classification, content of the submission to identify a characteristic of the innovation. The system may determine, using a third machine learning model and based on the characteristic, a ranking of the innovation relative to individual innovations in the subset of innovations. The system may determine, using a fourth machine learning model, an impact score associated with the innovation. The system may perform, based on the impact score satisfying a threshold, an action associated with a project involving the innovation.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating a win prediction for an opportunity. In some implementations, a server receives data representing an opportunity. The server obtains historical data from a data repository based on the received data representing the opportunity. The server generates standardized feature vectors based on non-standardized data, the non-standardized data comprising (i) the received data and (ii) the obtained historical data. The server filters the standardized feature vectors. The server generates a win percentage based on the filtered feature vectors using classifiers. In response to generating the win percentage for the opportunity, the server generates a plurality of data insights using the classifiers and the generated win percentage, wherein the plurality of data insights describe one or more data predictions for pursuing the opportunity.

Abstract: An AI-based process monitoring system access a plurality of data sources having different data formats to collect and analyze KPI data and shortlist KPIs that are to be used for determining the impact of automation of an automated process or sub-process. Information regarding an automated process is received and KPIs associated with the process and sub-processes of the process are identified. The identified KPIs are put through an approval process and the approved KPIs are presented to a user for selection. The user-selected KPIs are evaluated based on classification, ranking and sentiments associated therewith. The evaluations are again presented to the user along with a set of questionnaires wherein each of the questions has a dynamically controlled weight associated therewith. Based at least on the weights and user responses, a subset of the evaluated KPIs are shortlisted for use in evaluating the impact of process automation.