Introduction - Understanding Generative AI's Capabilities and Limits
Before diving into the applications, it's important to ground the discussion of generative AI with a clear understanding of what it can and cannot do. Put simply, generative AI excels at creating novel, human-like content conditioned on the data or prompts it receives during training. However, it has no sentience or deeper understanding of what it generates.
While remarkably good at tasks like language generation, image creation, and voice synthesis, current generative AI cannot match human cognition. Any claims of models having achieved general intelligence or consciousness are scientifically unfounded. Generative models may fail or have problematic biases if not developed carefully with human oversight. Still, used properly, they unlock tremendous new capabilities for enterprises across industries.
1. Enhancing Customer Service and Support
Generative AI allows communications companies to create natural customer interactions using advanced chatbots. As an example, T-Mobile is piloting avatar-based customer service powered by generative AI in a collaboration with startup Anthropic. The Claude chatbot interface draws on abroad dataset to have conversations indistinguishable from human agents for common support issues. If needed, it seamlessly brings human representatives into the discussion for more complex or sensitive queries.
Verizon is taking a different approach, using generative AI to assist its live human agents respond faster and more consistently to customer questions. Its Chief Customer Officer described generative models acting as a "coach in the pocket" for representatives, suggesting knowledge base articles, product specs, troubleshooting steps, and answers in real-time during calls. This positions human expertise at the center while AI handles the tedious information retrieval activities.
2. Automating Network Operations
Communications providers operate vast complex networks built up from multiple generations of technologies. Keeping these infrastructures humming along to serve customer needs requires extensive monitoring, predictive maintenance, and capacity optimizations. Generative AI is proving uniquely capable for many such data-intensive but rules-based tasks.
For example, AT&T has successfully developed several generative AI proofs-of-concepts for automating its network operations, management, and infrastructure monitoring. Use cases demonstrated so far include AI-assisted forecasting of user demand, traffic load balancing, proactive hardware issue detection, and automated responses for outage events.By handling many repetitive low-level tasks, AT&Ts AIs allow its human engineers to focus their expertise on higher value network optimization and planning.
Meanwhile, DISH is exploring the use of generative AI conversation interfaces between its network management system, technicians, and automated field robots. Human technicians can describe issues in natural language to the AI, which translates findings into technical tickets, predicts likely fixes, and even coordinates firmware updates and component replacements via robot technicians. This allows DISH to scale rapidly while keeping infrastructure resilient and its limited human technicians more productive.
3. Personalizing Marketing and Sales
Generative AI offers communications companies transformative ways to understand customers and deliver hyper-personalized promotions, recommendations, and customer experiences. As an example, Google-owned AT&T Fiber increased conversions by over 10% by using natural language models to tailor emails to individual customer journeys. Other carriers like Comcast leverage generative AI to continuously A/B test and refine their website content including personalized product recommendations for every visitor. Still more analyze usage patterns, demographics, and service issues to automatically generate custom retention offers for subscribers considering leaving.
In sales, Verizon media's demand-gen unit already uses generative AI to create targeted outreach emails matching the tone, messaging, and offerings most relevant to each marketing or sales lead based on their demographic data and past campaign interactions. They have slashed the time their human marketers previously spent manual personalizing outreach while improving results. Other communications firms will soon adopt similar applications in sales and marketing grounded in their troves of customer data.
4. Developing Innovative New Products
Thus far we have discussed established use cases where generative AI enhances or accelerates existing processes. However, completely new product categories are emerging that can exist only because of advances in generative models.
For example, in late 2022 Verizon launched its BlueJeans AIMeeting Assistant service alongside tech provider NVIDIA. The offering leverages large language models to generate automated meeting notes, highlights, timestamps, smart transcripts, participant summaries and more from the contents of enterprise video calls made over Verizon's BlueJeans platform.Because it works by listening in on natural conversations in meetings just as a human assistant would, none of this detailed capturing and summarizing of meeting contents would be possible without generative AI.
In another innovative offering, smart compose startup Quill has partnered with Canadian telecom giant TELUS to allow subscribers to easily create professional, customized marketing and sales content using Quill's AI.Generative Quill lets TELUS customers generate tailored advertisements, webpages, brochures, and more promoting their individual brands and offerings by simply describing their needs to the AI. Small businesses on a budget can now access agency-quality sales and marketing collateral on demand.
Key Takeaways
- Leading carriers like AT&T, Verizon, T-Mobile, andComcast are actively piloting a myriad of generative AI use cases to serve customers, optimize networks, and enhance sales and marketing.
- Emerging partnerships between communications companies andAI startups illustrate how new innovative products are being made possible entirely because of advances in generative models.
- Still in early stages, generative AI adoption isaccelerating across the communications industry's core functions from customerservice and network engineering to product development and marketing. Realizing the full capabilities while setting proper expectations around limits remains vital as applications scale.
Sources:
1.https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/five-use-cases-for-generative-artificial-intelligence-in-telco
2.https://venturebeat.com/ai/anthropic-t-mobile-customer-service/
3.https://www.capgemini.com/service/why-data-is-key-to-unlocking-the-full-potential-of-ai-in-telecom/
4.https://www.forbes.com/sites/junwu1/2020/05/05/ai-meets-5g-how-telcos-are-ushering-in-new-ai-use-cases/
5.https://berkeley.app.box.com/v/berkeley-ai-research-generative-ai
