Искусственный интеллект в финансовой сфере: 5 практических кейсов
Artificial Intelligence is critical for optimizing the testing process, aiding automation, and ultimately designing software that is self-healing. Read on to learn about key use cases on how AI can be leveraged for testing in the financial services world.
“Our intelligence is what makes us human, and AI is an extension of that quality.” – Yann LeCun
The global financial services industry is at an inflection point. Customer experience across all channels has become a must due to which customers are increasingly adopting digital products and services.
However, digitization has increased the complications for the testing professionals. Open banking, lightweight architectures, legacy systems integrated with a plethora of applications are making testing extremely time consuming.
Our recent World Quality Report, a joint publication of Capgemini, Microfocus and Sogeti, indicates that artificial intelligence (AI) in testing is an upcoming trend to improve speed to test particularly for digital initiatives in financial services. It is aimed at optimizing the testing process, aiding automation, and ultimately designing software that is self-healing.
59% of our financial services respondents stated that they would use AI techniques to optimize their QA processes. Our survey identified five use cases that are the top choice for using AI in testing. These are indicated by the figure below:
Intelligent automation: 50% of our financial services respondents stated that they would use AI for intelligent automation. This implies, deciding what to automate by using machine learning algorithms, algorithms such as co-relation and random forest algorithms
Predictive analytics: 41% respondents stated that they would use AI for predictive analytics. As an example, machine learning algorithms such as regression and time series algorithms can be used for defect prediction and release prediction
Prescriptive analytics: 39% of the respondents stated that they would use AI for predictive analytics. This implies, deriving various insights such as, what to test to drive intelligent decision making. For example, NLP algorithms such as cosine similarity can identify duplicate test cases and hence can be used for optimizing test suites
Cross application dashboards: 32% of the financial services respondents stated that they would use AI for cross application dashboards. This means using machine learning algorithms to create cross application dashboards. It is critical to determine application dependencies in terms of requirements, test assets and environments, which provide a single view to plan and govern testing activities
Self-learning cognitive platforms: 36% of respondents said that they would use AI for self-learning cognitive platforms. A use case for this is adopting machine learning algorithms to automate test environments and test case remediation across complex IT environments.
My recommendation to start on this journey, is to firstly understand above mentioned use cases. Secondly evaluate the pain areas in testing where these use cases would bring in value. Reliable data is a pre-requisite for any use case. A great place to start would be to build on data from application life cycle management tools, for example defect data. The next step would be to define metrics to measure success for each use case. Finally create a small team comprising of professionals possessing data analytics, statistics and machine learning algorithms that can create proof of concepts and drive this transformation.