The ability to analyse data, interpret insights, derive intelligent decisions and act on them has become a competitive differentiator for companies across sectors. There can be a misconception, though, that harvesting big data is enough. While quantity does matter, it is only through good analysis and interpretation that it becomes truly valuable, says Vinod Vasudevan, Group CEO, Flytxt. The Netherlands-based Flytxt is a fast-growing big data analytics solutions provider for telecoms and communication service providers (CSPs) across the globe. It has a presence in Thiruvananthapuram.
Vasudevan spoke to BusinessLine on how companies are adopting data analytics to drive digital transformation. Excerpts:
Is analytics getting the attention from the top managements?
Some sectors, such as telecoms and financial services, have grasped the incredible value of data analytics, but others have been slower when it comes to digital transformation.
There is a general understanding, though, of big data’s value. In a big data use cases survey, 69 per cent of respondents said it them the ability to make better strategic decisions.
Some estimates have predicted that the global market for big data analytics could be worth as much as $100 billion by 2020. At Flytxt, we maximise the value of customer data and engagement for 100 businesses, across a global network of over 50 countries.
Do you tap artificial intelligence to enable clients to create autonomous and fully-automated decisions?
Yes, we help businesses engage their customers by learning their behaviour and predicting what they want in real time.
Flytxt has built its own portfolio of AI, marketing automation and customer-engagement technology. We make sure our clients have deeper and longer relationship with customers through advanced analytics, machine learning and AI.
Recently, we launched an add-on capability in our product for interface voice platforms such as Google Assistant and Amazon’s Alexa. AI and machine-learned customer intelligence in our product can be extended on them, enabling enterprises to have meaningful human-like conversation with customers, unlike the pre-recorded conversations in a typical chatbot. So we are focussing on harnessing possibilities of combining AI with other technologies like NLP (neuro-linguistic programming), to help enterprises elevate customer engagement to the next level.
How does machine/deep learning fit into the scheme for Flytxt?
We offer the analytical capabilities of machine and deep learning in the form of analytics which are built and bundled as pre-packaged models.
These models are created by our R&D team in partnership with IIT-Delhi and TNO, The Netherlands (an independent research organisation). The R&D team includes resources with multi-functional skills — data engineers, data scientists, decision scientists, and business analysts — who build, customise, package, publish and maintain a set of these packaged analytics models in the product. These ready-to-use analytical models can be readily plugged to the desired marketing workflows to realise incremental economic benefits such as increasing revenue, reducing churn and improving customer experience.
Typically, adoption starts with simple analytical models and one-off use cases. Gradually, it extends to solving more complex problems.
How do you ensure the quality of data being crunched?
The quality of a business decision is only as good as the quality of the data used to make it. AI and analytics solutions invariably should do data validation and data enrichment to improve the quality of data.
Data quality issues can be accommodated to a greater extent by looking out for redundancies and correlations in the source data. However, we only access data which is available with clients. So the responsibility of ensuring consistency and accuracy of data largely remains with the CSPs (Communications Services Providers).
Leading and innovative operators have realised the enormous value that can be generated from their data; they are investing to explore new data sources and improve consistency and quality of data.