In 2018, our client wanted to introduce a machine learning and natural language Processing solution to improve the efficiency of its survey writing team by automatically generating incorrect options for the survey questions. To support this, it also needed to be able to validate data faster and more efficiently, as well as improve SLA compliance through software engineering. The company needed a vendor that could match its ambition to lead its industry in platform automation and automated data delivery with the expertise and capabilities to turn its vision into reality.
While surveys are an effective tool to understand the impact that a brand, service, or a product makes, their success relies on how the questions are written. It was taking one of the company’s inhouse writers around ten minutes to create a single question for each program. Much of this time is spent creating the incorrect or false answer options, creating ongoing inefficiencies in the company’s operations.
The team at ValueLabs suggested an innovative solution that would not only automates the process of writing incorrect and fake survey responses, but would have an active learning component that could learn from user feedback. This would significantly reduce the time taken to produce surveys. Alongside this, an automated tool would reduce the time spent validating data and software development and implementation would result in faster resolution times for service tickets.
The ValueLabs solution automatically generates options that are similar in structure to the right option, but not so similar as to cause ambiguity for the survey takers. It does this through analyzing the semantics and premise of the question and historical data – allowing it to generate unique options for each question by emulating the pattern learnt from previous data. In this way for the same question and answer pair it generates different sets of incorrect/fake options. The writers can then choose from the various options presented to them and use them to publish the surveys, creating huge time savings.
Through integration with the video viewer, Google Docs, and the survey questionnaire tool, the solution enables the user to view videos, take notes, select the question, correct response, generate incorrect options, and move to the final survey distribution step seamlessly. The active learning component enables it to learn from each user’s feedback, which means the system automatically learns and generates better options as time passes.
Natural Language Processing sits at the heart of this opportunity and it was awarded “Best Innovation in NLP” at 2020’s AI Summit in New York, which recognizes those that stand at the head of the curve by adopting NLP into their organization.
Awarded ‘Best Innovation in NLP’ in the AIconics awards at the AI Summit, New York
Migrating to the cloud is on the strategic agenda of most big corporates these days, however, upgrading from one on-premise database to one in the cloud requires all data to be validated between sources. Our client used the Fitnesse automated validation tool to compare data between tables and files.
The tool is applied only to sample data rather than all records. ValueLabs created a regression test suite, validated the data, and easily identified the differences between the data from the legacy table to the new technology table. ValueLabs then prepared the Fixtures and Beans, which are compatible with any type of database.
Using Fitnesse Wiki information enabled ValueLabs to develop an Automation Framework UI, where data sources, such as tables and fields, can be compared, validated, and presented via the UI and as an Excel validation report.
Our client now enjoys a generic framework for validating data across different sources with reduced manual effort and improved accuracy. It has already validated over 500k records, and can validate 1,000 records per minute. Multiple table data comparisons can be performed in Fitnesse, further reducing the need for human intervention. Finally, an audit trail provides visibility of past results and failed records.
Our client wanted to incorporate end-to-end software engineering practices, including development, production application support, database development, database administration, and DevOps support. This required the on-boarding of resources with appropriate skills and acquiring client domain knowledge with defined support delivery processes.
To fulfil these requirements, ValueLabs provided:
As a result, the company now meets 95% of its SLA commitments when it comes to service tickets, enabling faster resolution. In addition, post-delivery defects have reduced to less than five percent, ensuring high quality is maintained. At the same time, AI has been introduced for survey response generation and predictive analytics.
Our client’s strategic partnership with ValueLabs is now well established. The underlying machine learning and development framework will enable it to evolve and refine its operations further, while the software and cloud solutions allow it to work faster and more efficiently. This project is allowing the company to offer increasingly intelligent research services, and further establish itself as a leader in its market.
We’d love to help!
By continuing to use the site, you agree to the use of cookies. more information
The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.