CW Innovation Awards: Gleaning data insights with AI

Like most large organisations, Hong Kong-based Citic Telecom CPC manages vast amounts of customer, product and order data that can provide valuable insights into market trends and the competitiveness of its products.

However, the data was scattered across different systems and the analytics reports generated from the data not only contained discrepancies, but were also complex to implement and maintain. These challenges led to delays in generating reports and high server resource consumption.

Given that the reports are instrumental in helping the company’s management team to assess current operating conditions, analyse market trends and make informed decisions regarding product marketing strategies, it was important to address the challenges sooner than later.

In December 2023, Citic Telecom CPC, a subsidiary of Citic Telecom International that specialises in providing tailored and scalable ICT solutions for multinational enterprises, launched AI Databank, a platform that consolidates data from various sources into a central location.

By enabling intelligent analysis of the data, users can now delve deeper into data trends, uncovering hidden patterns and valuable insights to drive better business outcomes through sales performance analysis and risk identification.

Thanks to the use of large language models (LLMs), users can also interact with an intelligent report assistant using natural language to ask questions and retrieve data from the database. This has improved user experience, reduced complexity and increased the efficiency of data retrieval and analysis.

Just months into its launch, AI Databank has already had a positive impact on Citic Telecom CPC’s business. For example, it has reduced report generation time from one hour per report to less than 10 seconds.

With approximately 50 existing business reports, including daily and monthly reports, the estimated annual time saving in server resource consumption is more than 4,000 hours, reducing server resource usage and datacentre operating costs.

The time required for developing new report pages has also been reduced from 30 man-days to 15. Considering the current demand for developing new business reports, with around five to eight new report requirements per year, it is estimated that 105 man-days can be saved annually. This translates to a 50% increase in efficiency of report development and manpower cost savings amounting to about HK$167, 000.

By automatically analysing chart data and summarising data trends, the intelligent report assistant has made it easier for users to understand charts, enhancing user experience. It has also reduced the workloads of IT teams who receive at least two to three raw data queries per report by 700 man-days per year.

The AI Databank platform was built in-house from the ground-up, leveraging the company’s capabilities in AI, big data, and the internet-of-things. It has also obtained computer software copyright for the platform in China and applied for relevant patents on using AI to enhance the platform.

Change management was key to the project’s success. Change requests were evaluated by designated team members who assessed each request’s needs, goals, and potential impact on project duration, costs, and deliverables. Approval or endorsement from management was required before accepting a change request.

It also carefully selected team members with the necessary skills, knowledge, and experience to drive change. Additionally, having team members who were open to change and receptive to new ideas was advantageous to handing changes with efficiency.

Finally, agile project management principles were applied in the project, allowing for flexibility and adaptation to changes. The project was divided into six sprints, with each sprint involving requirements confirmation, design, implementation and user acceptance testing. This approach significantly reduced the likelihood of change requests.