After ChatGpt became mainstream, some organizations waited to see what had happened. Now we see hyperschools like Google, Nvidia, Openai and more quickly roll out new features and proven potential for a variety of use cases, so that's a race mindset.
Genai is a priority for everyone, but it takes time to really get it right within an organization. The large-scale language model (LLM) introduced over the past two years is from external sources, and therefore could not invade and connect to the “closed garden” communication information systems.
Our top priority was to connect to all the different domains within the telephone company and build a solid foundation for connecting them to each other, but never before. This is necessary to give call center agents sufficient insight, as customer complaints may contain information from the network. A proper architecture should be in place before the phone company can fully assist agents with Genai or interact with customers.
Its architecture offers great benefits. For example, we worked with very large North American operators to deploy POCs as an aid to call centers, reducing call processing times by 63% and improving initial call resolution by 50%. As we move towards a wider commercial deployment, greater benefits are occurring.
This is a bright signal for all operators on the customer-facing aspect, and there is more. We investigated how consumers feel about their interactions with Genai. Furthermore, approximately 80% of respondents already interact with each other in their daily lives, so about 80% of respondents are positive.
Network and other domains
Communications Service Providers (CSPs) have been using AI for a long time in their networks, but they take a long time in their networks and other domains. When ChatGpt hit the mainstream, no one realized that progress was so important. Now we see that the scale of the feature continues to improve and creates an implementation table interest. We are in a good place and it will just be faster.
People realize that if they don't take a positive approach, it will affect their future. They also realize that this is not a technological revolution. It is a holistic, end-to-end revolution that changes organizational setup and scale. In 10 years, will customer care be processed by 20,000 people following the sun? no. The real question is how quickly you can reach the minimum number of people you need. Also, what does that mean for care organizations, CSPS network organizations, or marketing?
Telecom is a human-intensive industry, so there is scope to make the industry even better. No more initial call resolutions that fail, or excuses for customers waiting on the line. In theory, there is no problem as it will be resolved before a complaint is received. Customer care is highly personalized and our products are as personalized.
Our CSP customers will also be very profitable. That's not the cloud revolution. With Genai, you can climb the ladder of customer satisfaction and reach new heights. This is exciting for the industry with the lowest NPS score. In 3-5 years, people should be very pleased with the involvement of CSPs.
The impact of the additional agent AI
AMDOCS demonstrated agent AI to customers at MWC's booth in 2024, generating a lot of interest. We developed a new field called personality engineering, as we learned that we needed to use several random profiles to teach agents how to act along the culture of our brand, reflect their values and use that language.
Human agents adapt their actions to the person they are talking to. I don't want to talk to an 85-year-old man, as the same agent doesn't use the same language or approach as an 85-year-old man. Furthermore, from the studies we commissioned, we learned that preferences differed depending on the location. In most Asia, people prefer to interact with older people, but this is not necessarily a European preference.
These issues have not yet been widely discussed, but are extremely important. Agent AI is not a high-tech upgrade, it's a whole new channel for brands, especially important for CSPs that rely heavily on humans.
Progress through the catalyst project
At DTW Ignite, AMDOCS participates in three catalysts. This is a unique opportunity to partner with CSPs and have a real conversation, addressing real issues.
In the first catalyst, Bind (for intelligence, networks, and digital twin bridging) Amdocs works with Vodafone and Telstra to create digital twins for the network in real time and connect knowledge-based guaranteed stocks. The scenario is that the operator wants to start a new service and make it available to all components to provide the service. Also, if there is a fault, how to correct it accurately.
AI-equipped billing QA and anomaly detection Avoid shock from the bill. Traditionally, communications are to collect information throughout the month and finally create a bill. Things go wrong for many reasons: the operators themselves lose about $40 billion a year from inaccurate billing and revenue leaks. Here, genai is actively applied to ensure quality and address any issues that arise. You can also predict customer complaints and the likelihood that customers will be tumbling.
I'm cultivating B2B using autonomous agents It enables flexible and fast citations for medium-sized businesses. This is difficult because it involves a lot of handguns, information and logistics. Currently, there are errors in 50% of bid requests. I want to use autonomous agents to convert my B2B sales process.
Here, the agent uses the autonomous agent to create and combine the information, and manages the activation of the proposal and ensures fulfillment. This is very important. One of the worst statistics in our industry is the delay between ordering and delivery completion. This also means a long period between concepts and cash.
