'Personality Engineering' Puts a Human Face on Telco AI Agents

Amdocs and Nvidia are pioneering "personality engineering" to create AI customer service agents that embody brand identity and adapt to customer needs. This approach aims to make AI agents more relatable and effective representatives of telco brands. AI customer service agents, sometimes humorously called "clankers," have struggled with reputation and acceptance. Amdocs defines personality engineering as the process of tuning AI agents to reflect an organization’s brand guidelines, language, visual identity, and values.

"Your agent is your brand representative, just like a human agent in the call center would be," said Hillel Geiger, Amdocs global VP of corporate marketing, at the DTW Ignite Forum in Copenhagen.

AI agents must also consider the context of interactions. For example, a customer upgrading a plan has a very different conversation than one complaining about a bill or service quality. Additionally, the AI should adapt its tone depending on the customer’s profile — a conversation with an 80-year-old differs from one with an 18-year-old.

Consumer Preferences for AI Agents

Amdocs conducted a study with 7,000 respondents across 14 markets and engaged 130 telecom executives to understand consumer preferences for AI agents. Key findings include:

• Gender: 45% prefer female-presenting AI agents, 25% male, and 30% have no preference. Among female respondents, 67% prefer female-presenting agents.
• Age: Preferences are mixed; 39% have no preference, 35% prefer agents of similar age, 14% prefer younger, and 12% prefer older agents.
• Tone: 58% prefer an empathetic tone, highlighting a desire to humanize interactions despite knowing they are speaking with software.
Chris Penrose, Nvidia global VP of business development, noted that even users tend to use polite language like "please" and "thank you" when interacting with AI agents. Customers can customize AI agents’ gender, age, and tone, creating numerous permutations to optimize user experience. However, guardrails are essential to prevent agents from straying into inappropriate topics like politics.

Proof of Concept Success

Amdocs reported a proof-of-concept trial with a large North American operator where AI agents reduced call handling time by 63%, improved first-call resolution by 50%, and boosted Net Promoter Score (NPS) by 50%.

Collaboration Between Amdocs and Nvidia

The companies collaborate on the Amdocs amAIz platform to develop specialized AI agents for telco functions such as care, sales, marketing, and network optimization. Each agent is designed as an expert in its domain, leveraging Amdocs’ deep telco knowledge. They are also working on orchestrating agents to handle multiple intents in a single interaction, akin to experts from different departments collaborating to provide the best solutions.

Emerging Standards for AI Agent Interoperability

Standards like Cisco’s AGNTCY foundation, the Agent2Agent (A2A) protocol, and Anthropic’s Model Context Protocol (MCP) are being developed to enable interoperability between AI agents. Nvidia’s Riva platform supports customizable AI voice agents with real-time multilingual speech recognition and translation, while Nvidia NeMo optimizes performance and scalability.

Large Telco Models (LTMs) and Advanced AI Applications

At Nvidia’s GTC developer conference, the concept of Large Telco Models (LTMs) was introduced. These specialized large language models can reconfigure networks rapidly, demonstrated by AI agents improving wireless network performance by 30% during a simulated baseball game. Humanlike AI avatars, or "digital humans," are emerging as superior interfaces compared to traditional chatbots, offering higher customer satisfaction and purchase likelihood.

Challenges and Risks: AI Hallucinations

Despite progress, AI hallucinations—where agents provide incorrect or misleading information—pose risks. Examples include Air Canada being ordered to compensate a customer after chatbot misinformation and McDonald’s removing AI ordering tech due to bizarre errors. Such errors can damage customer trust and brand reputation, especially in sensitive situations. Mitigation strategies include grounding AI in real data using retrieval-augmented generation (RAG), integrating company knowledge bases, and maintaining human oversight to correct issues.

Final Thoughts

Amdocs and Nvidia’s personality engineering represents a bold step toward transforming telco customer service with AI. The success of this approach depends on scaling deployments, maintaining trust, and balancing human-like empathy with machine reliability. While some users prefer AI agents with personality, others, like the article’s author, favor straightforward, professional AI interactions without anthropomorphizing.

---

Frequently Asked Questions (FAQ)

What is "personality engineering" for AI agents?

Personality engineering is the process of designing AI agents to reflect a company's brand identity, values, and communication style. This involves customizing aspects like tone, language, and even the agent's persona to create a more relatable and effective customer service experience.

What are the key consumer preferences for AI agents mentioned in the study?

The study highlighted preferences regarding gender (majority favor female-presenting agents), age (mixed preferences, with a slight leaning towards similar age), and tone (empathetic tone is preferred by most).

How does Amdocs collaborate with Nvidia?

Amdocs and Nvidia are collaborating on the Amdocs amAIz platform to develop specialized AI agents for various telecommunications functions. This partnership leverages Amdocs' telco expertise and Nvidia's AI technology.

What are the benefits of personality engineering for AI customer service?

Personality engineering aims to improve customer acceptance and relatability of AI agents, making them more effective brand representatives. It can lead to better customer satisfaction and potentially improved resolution rates.

What are AI hallucinations and how are they addressed?

AI hallucinations are instances where an AI agent provides incorrect or misleading information. Mitigation strategies include grounding AI in real data, integrating company knowledge bases, and maintaining human oversight to correct errors.

What is a Large Telco Model (LTM)?

A Large Telco Model (LTM) refers to specialized large language models designed for the telecommunications industry, capable of tasks like rapid network reconfiguration and performance improvement.

What are digital humans in the context of AI customer service?

Digital humans are humanlike AI avatars that serve as interfaces for customer service, offering a potentially more engaging and satisfying experience than traditional chatbots.

What are the challenges in implementing AI customer service agents?

Challenges include managing AI hallucinations, maintaining customer trust, preventing agents from straying into inappropriate topics, and ensuring effective scaling of deployments.

What were the results of the Amdocs proof-of-concept trial?

The trial showed significant improvements, including a 63% reduction in call handling time, a 50% increase in first-call resolution, and a 50% boost in Net Promoter Score (NPS).

What emerging standards are mentioned for AI agent interoperability?

Standards like Cisco’s AGNTCY foundation, the Agent2Agent (A2A) protocol, and Anthropic’s Model Context Protocol (MCP) are being developed to enable different AI agents to work together.

Crypto Market AI's Take

The development of "personality engineering" for AI agents in telecommunications signifies a crucial advancement in customer interaction technology. By imbuing AI with distinct personalities, companies can foster deeper connections with their customers, moving beyond functional interactions to build brand loyalty. This trend aligns with our broader understanding of how AI is being integrated across various sectors to enhance user experience and operational efficiency. At Crypto Market AI, we are keenly observing how AI agents are being utilized not only in customer service but also in complex financial markets. Our focus on AI agents for trading and market analysis highlights our commitment to leveraging cutting-edge AI for actionable insights and automated strategies in the cryptocurrency space. The success of personality engineering in telcos could pave the way for similar nuanced AI applications in finance, where trust and personalized engagement are paramount.

More to Read:

• AI Agents: The Future of Business Automation and Customer Engagement
• How AI Crypto Coins Drive Innovation as Blockchain and AI Converge
• Understanding AI Agent Washing: Risks and Realities

---

Originally published at Fierce Network on August 7, 2025.