In the future, AXIATA may deploy customer service AI agents capable of human-like interactions for its customers. Additionally, we believe CDB may market customized Generative Artificial Intelligence (Gen AI) models and applications to Malaysian enterprises via its partnerships with global cloud providers. We maintain our forecasts, TP of RM3.00 and OUTPERFORM call.

We came away from AXIATA’s IDEAS@ Axiata 2024 forum with a neutral perspective, as it is still early to gauge the full impact of Gen AI. The forum was themed: Leveraging the AI revolution to unleash a new wave of value creation. The key takeaways are as follows:

1. Enterprises and organizations can build and deploy customized Gen AI models by integrating their respective data lakes with foundation models from various AI providers (e.g. OpenAI, Microsoft, Amazon, Anthropic, Meta) as a starting base. These customized models undergo fine-tuning, and may use retrieval- augmented generation to enhance the accuracy and relevance of outputs. Moving forward, we believe CDB may market these Gen AI applications to Malaysian enterprises via its existing partnerships with cloud providers (e.g. Amazon Web Services (AWS) and Microsoft Azure).
    
2. In the future, AXIATA is optimistic about leveraging Gen AI to develop customer service chatbots capable of human-like interactions for both its telco and ADA (Axiata Digital Analytics)’s customers. Previous attempts using classical AI were unsuccessful due to their rigid rule-based approach. Hence, this limited the chatbots' ability to respond to unanticipated questions not covered by predefined rules. Moving forward, robust large language models (LLM) can be utilized to deploy more effective AI agents. For instance, one LLM with analytical capabilities will be used to determine a customer’s intent, while another LLM with generative capabilities will produce appropriate utterances. The combination of these LLMs is expected to result in a natural, fluid, interactive and personalized chat experience.
    
3. AXIATA recently introduced the application of Gen AI into its AI Transformation Program that was first launched in 2021. As a start, AXIATA has identified top-five use cases for the deployment of Gen AI, comprising: (i) coding assistant, (ii) virtual agent, (iii) personalized care, (iv) customer operation, and (v) network operation. Whilst AXIATA is proactively investing in the deployment of Gen AI, its direct costs are marginal relative to the group’s total opex and capex. The bulk of AXIATA’s AI investments are largely channelled into its AI factory and staff training. There are currently more than a hundred AI models in AXIATA’s AI factory, that is powered by AWS and placed in the group’s major opcos, (ie. XL, Idea and Robi).

Forecasts. Maintained.

Valuations. We also maintain our Sum-of-Parts TP of RM3.00 (refer below).There is no adjustment to our TP based on ESG given a 3-star rating as appraised by us (see Page 5).

Investment case. We continue to like AXIATA for: (i) its plans to deleverage and strengthen its balance sheet, (ii) growth prospects for digital telcos and tower assets at emerging markets, and (iii) strong asset monetization prospects for Edotco and its digital businesses. Maintain OUTPERFORM.

Risks to our call include: (i) a strong USD may weigh on the performance of its digital telcos at frontier markets (e.g. Robi Bangladesh, Dialog Sri Lanka, Smart Cambodia), (ii) gestational earnings and cash flow drag from Link Net’s aggressive expans ion, and (iii) capex up-cycle from looming implementation of 5G in Indonesia.

Source: Kenanga Research - 7 Aug 2024