The BIG problem with generative AI

You know I’m a huge fan of the public cloud and how it can help you save money, innovate faster, and delight your customers. But there’s another technology that I think is equally game-changing for telco: generative artificial intelligence (or “GenAI” for short).

GenAI can literally help you write anything from code to marketing copy to presentations. But can telcos fire up ChatGPT “as-is” and start using it? Probably not. For example, you probably heard the story from March of this year where a few Samsung engineers used ChatGPT to help fix bugs in their source code, optimize their test sequences, and create some presentations. But in doing so, they also exposed their confidential data to ChatGPT’s external servers, where it couldn’t be retrieved or deleted. Oopsie! 

This story highlights the big issue for enterprises using GenAI tools that are open to the public. Yes, these tools are fantastic for improving productivity, and their responses only get better with more and more data; but when the input you give them becomes part of the public domain, how can you use them throughout your business without risking your proprietary data? That’s a big problem for telco, where the data is highly regulated and one of your most valuable assets. 

Thankfully, Amazon Web Services (AWS) has an answer for you.

What’s all the fuss with GenAI?

Unless you’ve been living under a rock (I mean seriously, you can’t be on Twitter without seeing the impact of this new tech), you know that ChatGPT rocked the tech world when it launched in November 2022. ChatGPT is “a large language model that can be used for various natural language processing tasks such as text generation, summarization, and question answering” as well as language translation and content generation. It has been used to generate creative content such as poetry, and it even generated a fake Drake song that got 8.5 million views on TikTok and YouTube before it got pulled down. If you haven’t already played around with it, what are you waiting for?

I was working on this blog about how telcos can use large language models (LLMs) like ChatGPT, when AWS dropped its latest awesome bomb: the announcement of Bedrock, a tool for enterprises to help them build with GenAI.

AWS Bedrock is a service that allows any enterprise to build and scale GenAI applications without having to manage the infrastructure. AWS Bedrock provides access to LLMs from leading AI startups and Amazon (including Amazon’s new Titan FMs), giving you the ability to privately customize them with your own data. So I asked Microsoft Edge (via Bing) why AWS Bedrock is better than Azure OpenAI or Google Bard.

Here’s what it said:

1. AWS Bedrock provides access to a wider range of foundation models (FMs) from leading AI startups and Amazon, while Azure OpenAI offers only OpenAI’s models and Google Bard offers only Google’s models.
   
2. AWS Bedrock allows enterprises to privately customize FMs with their own data, while Azure OpenAI and Google Bard do not offer this feature.¹ ²
   
3. AWS Bedrock integrates with AWS tools and capabilities such as Amazon SageMaker, AWS Lambda, and AWS CloudFormation, which makes it easier to deploy secure and reliable generative AI applications, while Azure OpenAI and Google Bard do not have such integrations.¹ ²
   
4. AWS Bedrock is designed for B2B use cases, while Azure OpenAI and Google Bard are more focused on consumer use cases.¹ ²

Source: Conversation with Bing, 4/24/2023.

¹ The fast-changing commercial AI landscape – TechTalks. https://bdtechtalks.com/2023/04/24/commercial-ai-amazon-google-microsoft/ Accessed 4/24/2023.

² AWS Builds Generative AI For B2B From The Bedrock Up. https://www.forbes.com/sites/patrickmoorhead/2023/04/18/aws-builds-generative-ai-for-b2b-from-the-bedrock-up/ Accessed 4/24/2023.

³ Compare AWS and Azure services to Google Cloud. https://cloud.google.com/docs/get-started/aws-azure-gcp-service-comparison Accessed 4/24/2023.

I could have googled all of that and written up the four points, but Bing did it for me in about 30 seconds. How crazy is that? It’s crazy awesome, tbh.

How telcos can use LLMs

Now that AWS has solved that huge problem for enterprises, I wanted to ask GPT-4 to give me some ideas on how telcos could use generative AI to improve their business.

Here’s what it said:

ChatGPT can be a valuable tool for telco network operators to improve their business in several ways. By leveraging the capabilities of this advanced AI language model, operators can enhance customer experience, optimize operations, and support decision-making, among other benefits. Here are five ways ChatGPT can help a telco network operator:

1. Enhanced Customer Support:
   ChatGPT can be integrated into customer support channels (e.g., live chat, social media, or messaging apps) to provide instant, accurate, and personalized responses to customer queries. This can reduce waiting times, improve customer satisfaction, and free up human agents to handle complex issues that require a human touch.
2. Automated Network Monitoring and Troubleshooting:
   By incorporating ChatGPT into network monitoring systems, telco operators can better identify potential issues, suggest troubleshooting steps, and even automate certain tasks. This can help minimize downtime, optimize network performance, and lower maintenance costs.
3. Internal Knowledge Base and Training:
   ChatGPT can serve as an interactive knowledge base for employees, providing them with instant access to information about products, services, and processes. This can improve employee productivity, reduce training time, and ensure that staff have accurate, up-to-date information at their fingertips.
4. Data-Driven Decision Support:
   Telco operators can use ChatGPT to analyze large volumes of data and generate insights to inform strategic decisions. For example, the AI model can help identify trends in customer behavior, perform churn analysis, or optimize network capacity planning. This can lead to more informed decisions and better resource allocation.
5. Marketing and Sales Automation:
   ChatGPT can assist in creating personalized marketing and sales campaigns by generating targeted content, predicting customer preferences, and offering smart product recommendations. This can help telco operators increase customer acquisition, retention, and average revenue per user (ARPU).

By leveraging ChatGPT’s capabilities, telco network operators can improve customer experience, optimize network performance, and make more data-driven decisions, ultimately leading to increased efficiency and profitability.

And I got that response from ChatGPT in about 0.2 seconds. Amazing.

Give it a spin (or a prompt!)

Telco execs need to try GenAI out for themselves. Do NOT delegate it! Think of some “prompts” (the sentences you use to generate responses) and use GenAI to write an email or memo, or use it to help create your next presentation. Be sure to play with different LLMs: you’ll find that some are better at certain tasks than others (similar to how early search engines were different from each other, too). I’ve been using the Poe app (an iOS app with a $20/month subscription charge for unlimited searches; there’s a website as well) because it offers seven different LLMs in one app. Pretty nifty.

AWS is making it super easy to incorporate GenAI into your business by providing enterprise-worthy tools that make it safer to use. AWS Bedrock is in a limited preview, but once it’s ready, it’s going to go gangbusters. If you weren’t busy moving your workloads to the public cloud, then I guess I just gave you one more reason to get going. What are you waiting for?

Note: If you can’t tell, we used a shit-ton of GenAI to write this blog. Mad props to Bing, GPT-4, and Claude+ for helping us learn about using LLMs in telco.

Recent Posts

Get my FREE insider newsletter, delivered every two weeks, with curated content to help telco execs across the globe move to the public cloud.

"*" indicates required fields

Email*

Phone

This field is for validation purposes and should be left unchanged.