Request a call
  • Hidden

Business Overview

Our client is a large oil and gas production company. They manage numerous oil wells throughout the US that are serviced by thousands of employees. Also, the company works with many third-party vendors.

The company came to NIX with an issue related to monitoring service quality provided by employees and third-party vendors. They looked for a mobile solution that could enable managers to leave feedback about the performance and quality of the work and quickly share it.

We analyzed their business needs and requirements and offered them a comprehensive solution, including: 

  • A speech-to-text iOS mobile app for local managers to quickly record feedback as a voice message and email it.
  • A web application for management with a dashboard to monitor overall statistics and feedback about specific work. 

Challenge

challenge icon

Speech recognition services should work without access to the internet. In the oil and gas industry it’s common to work far from reliable internet connections, so users need the ability to record feedback in different parts of the country where connections are poor.

challenge icon

None of the existing solutions make it possible to highlight action points in the transcribed text automatically. We needed to build a machine learning model to deal with this issue.

Solution

Mobile App

Our experts developed a speech-to-text mobile app for iOS, accompanied by a web application.

Users can record their feedback in a voice message, and the app instantly converts it into text, generating overall feedback assessment and recognizing call-to-action phrases. Also, users can then email it to the recipient directly through the app.

Thus, managers can quickly and conveniently exchange their feedback. Also, all records are stored in the system for statistics monitoring and further analysis.

To solve potentially poor connections, we implemented the following: recorded feedback is saved in the device. When an internet connection appears, the app processes the files according to the standard algorithm of actions.

Group 48096326

How it works:

  • We used Apple and Google’s speech-to-text frameworks to recognize speech and generate a text message of the spoken feedback.

  • Using the Amazon Comprehend natural language processing service, we extracted key phrases that appear in the text and define an overall feedback sentiment such as positive, negative, or neutral.

  • We built and taught a machine learning algorithm using framework spaCy, the library for natural language processing in Python, to identify and highlight the action points from the feedback text.

Web Platform

At the same time, NIX web developers created a platform that provides access to view and analyzes all feedback recorded by managers. Through this platform, top management can view overall statistics across the system:

  • Top 5 best and worst employees and vendors
  • Total number of positive, negative, and neutral types of feedback
  • Total number of managers, employees, vendors, and services
  • New feedback
  • All feedback assigned to the employee or vendor
  • Feedback details such as text, date, and time of the creation

 

notebook (1)

Outcome

As a result, our client implemented the speech-to-text mobile app and the web platform processing feedback into the company’s workflow. This innovation significantly simplifies communication among local managers. 

Quick feedback exchange and a global database allow the company to enhance the analytics quality and increase employee performance.

The feedback app is also an excellent base to build a central system for management of the company with powerful analytical and processing options.

Team:

7 experts (Project Manager, Business Analyst, 2 IOS Developers, 2 Data Engineers, Designer)

Tech Stack:

JavaScript, Swift, Python, AWS, IBM Speech to Text

Contact Us