Enhancing telco revenue streams and customer loyalty with GenAi

In an increasingly commoditized market, providers are seeking new revenue opportunities by offering differentiated services to enhance customer experience. With GenAI, telcos now have more flexibility to go beyond traditional offerings and drive innovation and excellence through hyper-personalization.

Chatbot enhancements currently constitute the most common GenAI use cases in telcos. Applications that generate business value, such as improving sales, churn modelling and marketing effectiveness, are also gaining ground. A report by Altman Solon, sponsored by AWS, reports that Communication Service Providers (CSPs) with a higher degree of data capability are pursuing revenue-generating use cases such as personalization and new product feature generation. 64% of CSPs surveyed agreed that many of the GenAI use cases being considered were novel applications.

Unlocking new value streams

Results from GenAI pilots by telcos are encouraging. A McKinsey study reports that a Latin American telco expects costs to fall by 15-20 percent with GenAI enhancements in chatbots and agent support. By using Gen AI’s summarization capability for voice and written customer interactions, the telco expects to reduce associated costs by up to 80 percent.

Telcos are building capabilities around conversational AI at the contact center to capture the customers' voice and intent. Feeding this instance further into GenAI tools will not just transcribe voice into text but also analyze it for context and sentiment. The contact center agent is equipped with a powerful knowledge base built upon a Large Language Model (LLM), which enables quicker resolution of customer inquiries.

GenAI can detect the perfect moments to interact with a customer. For instance, if the customer is using more than their monthly plan, GenAI can suggest an upgrade. Additionally, using data from voice interactions, GenAI can assist the agent in guiding the customer through the upgrade process.

Hyper-personalization based on deep customer insights is driving sales and marketing effectiveness, opening opportunities for higher revenue generation. GenAI models continuously analyze customer data, enabling the creation of personalized messages that adapt to the unique characteristics of each microsegment. A McKinsey study highlights that a European telco is achieving a conversion rate of more than 10% by generating new sales leads from customer interactions.

With the vast amount of customer data, providers can now offer personalized ‘next-best actions’ and ‘next-best offers’ to drive customer engagement and satisfaction.

Deployment considerations for GenAI

1. Data quality: Does your organization have sufficient high-quality data to train and maintain effective GenAI models? Quality data is the baseline, without which GenAI may produce erroneous or biased results. Institute data management practices to ensure input data is accurate, consistent, and complete. Identify all the sources of unstructured data and establish data lineage.
   
2. Network optimization: How does GenAI optimize networks for enhanced performance, reliability, and cost-effectiveness? Telcos can leverage AI algorithms to optimize network performance, predict traffic patterns, and proactively address connectivity issues for improved customer experience. By analyzing real-time network data and user behavior patterns, telcos can proactively identify and address network congestion, service outages and bottlenecks to better the overall network efficiency. Also, AI-powered network optimization can support the deployment of emerging technologies such as 5G and edge computing to enhance core telco service offerings.
   
3. Implementation costs: A question that begs careful assessment is the total cost of ownership of a GenAI deployment. This McKinsey estimate shows how costs could vary depending on different archetypes. The number of variable components—cost toward inference, prompt engineering, fine-tuning, or running the operations—make pricing strategies critical for realizing business value. For example, should an organization consider a token-based pricing model or a per-user subscription model for inferencing an LLM? Should the organization consider a hyperscaler such as Bard or an open source LLM such as BLOOM? Quantifying the intended benefits and monitoring progress are equally important to define an optimal pricing model that balances cost with returns.
   
4. Existing capabilities: Does your team have the skills and infrastructure to build, deploy, and manage GenAI solutions? Companies need to apply fine-tuning techniques to create custom models from LLMs that cater to their business, select the right tools for optimum results, integrate them into workflows, assess the cost implications, and train their resources. Many telcos are making the smart choice of working with a technology partner with expertise in data management, integration and industry domain knowledge to achieve their GenAI goals.

GenAI holds immense promise for customer engagement in the telecom industry. However, success will hinge on responsible data management, adherence to evolving regulations, and a strategic adoption approach. While data quality will be paramount for a trustworthy output, addressing the complexities of privacy laws and ethical AI will ensure customer confidence. An equally important factor will be agility—how an organization finds use cases that offer optimal benefits, deploys them quickly for an early-mover advantage, and continues to adapt as new frontiers open up such as text to video conversion which is slated to be the next big thing.