In today’s era of digital connectivity, any enterprise cannot afford to have blind spots and operational bottlenecks.

For Singaporean telco M1, catering to over 1.9 million customers is no easy feat. In order to ensure uninterrupted service delivery, it is crucial for their internal systems and processes to operate seamlessly.

Thus, when an opportunity arose to implement a centralised enterprise data lake, the company embraced it without hesitation.

According to Danny Thien, Director and Head of Data, M1, each of their departments previously had their own data warehouse or databases to support various business demands and user requirements.

“This resulted in data fragmentation between source systems and reporting systems, making it increasingly difficult to manage over 200 data repositories,” he said.

Additionally, ensuring the proper security, patching, and updating of each data repository consumed substantial time, effort, and financial resources.

Catching a snowflake

Recognising that a solid data management foundation is one of the keys to digital transformation, M1 sought a technology partner to help untangle its complex data processes.

Consequently, the telco chose Snowflake, a data cloud company, to establish effective data governance and robust safeguards for securing customer data.

“With a robust and scalable data architecture, we can readily support innovation in various areas like sales, marketing, customer service, risk and fraud management, finance, and billing operations,” explained M1’s Danny Thien.

M1 leveraged Snowflake’s Telecom Data Cloud, a fully managed, cloud-based platform that offers prepackaged data sets. This platform, together with streamlined onboarding processes and advanced data management and analytics tools, supported M1 in optimising its operations.

Specifically, the Telecom Data Cloud unites Snowflake’s data platform, collaborative solutions, and industry-specific datasets to eliminate data silos. With data being able to flow freely and securely across the entire business ecosystem, M1 can enrich its ML models, and then share and analyse data to drive better decisions.

“More than ever, the telecom industry can leverage data to become a critical partner to their customers. Customer expectations are growing because of increased competition driven by developments such as 5G, and as a result, more data — both first- and third-party data — is crucial to delivering the level of personalisation that subscribers demand,” said Sanjay Deshmukh, Senior Regional Vice President, ASEAN and India at Snowflake.

The Snowflake executive also emphasised that service providers should have an agile technology platform that simplifies the preparation of data for ML- and AI-driven data science applications, without the burden of complex integrations or related expenses.

“This enables the monetisation of new services, driving revenue and growth by overcoming complexity,” Deshmukh continued.

Integration strategy

Presently, Snowflake holds a central position in M1’s system landscape. Its platform acts as the hub where raw data from source systems is first replicated into the data lake. The data undergoes cleansing, transformation, and loading processes to create data schemas and models required by each business function.

“This integration enables self-service analytics for users across the organisation. The Snowflake data lake also integrates with critical application infrastructure including our business support systems and operation support systems to facilitate smooth and efficient overall business processes and operations,” M1’s Danny Thien said.

However, the adoption of Snowflake’s platform posed challenges for M1. The telco had to upskill its employees and familiarise them with the features, functions, and capabilities of the new data platform. They needed to understand how each component contributed to the revamped data architecture.

“People transformation is an important component of our overall digital transformation journey. To ensure a successful transformation, we organised training sessions for our employees to help them go through a process of not only mastering a new cloud application and associated technologies. They had to unlearn and relearn processes, workflows, and operations. Thankfully, we received support from our vendor and system integrator in this journey,” the M1 executive noted.

In a significant way, Snowflake has empowered M1 to process and analyse data on a large scale. As a result, M1’s data analysts and data scientists now possess the necessary tools and access to data sets to build analytical assets and models. This capability has enabled M1 to enhance customer engagement and drive hyper-personalisation.

Moreover, the scalable data architecture facilitated by Snowflake has enabled M1 to innovate and improve risk and fraud management.

“Snowflake’s fine-grained, role-based access control and column-level security has been instrumental in safeguarding sensitive customer data while maintaining necessary compliance. Their data loss prevention measures instil confidence in data recovery in the event of any unforeseen circumstances,” Thien remarked.

“Also, Snowflake offers readily available anonymising techniques like dynamic masking, and the data sets are protected with AES-256 encryption at rest and TLS in transit,” he added.

Ultimately, M1 is able to comply with regulatory requirements like PDPA and GDPR because of Snowflake’s data protection, the M1 Head of Data said.

Eyes on the prize

With its renewed agility and scalability, all lights are green for M1’s other data initiatives.

“In addition to leveraging AI and ML to enable hyper-personalisation for customers, we are increasing collaboration with other Keppel entities and trusted partners across industries. We have been working with regulators and partners on data sharing initiatives that leverage privacy-enhancing technologies to co-create synergistic products that bring real value to consumers,” stated M1’s Danny Thien.

Meanwhile, Snowflake is determined to take advantage of generative AI and large language models (LLMs) moving forward.

“Snowflake users are already taking advantage of LLMs to build really cool apps with integrations to web-hosted LLM APIs using external functions, and using Streamlit as an interactive front end for LLM-powered apps such as AI plagiarism detection, AI assistant, and MathGPT,” highlighted Snowflake’s Sanjay Deshmukh.

Deshmukh further emphasised that industry-specific versions of platforms, such as Snowflake’s Telecom Data Cloud and Manufacturing Data Cloud, ensure that businesses can leverage tailored data models, as well as connectors to common applications and data sources to monetise their information.

“In the end, these benefits free data experts to create greater value for the business,” the Snowflake executive concluded.