ChatGPT and the future of AI at work

Having worked in the field of conversational AI for many years (including co-founding a startup that was eventually acquired), I wanted to start off by sharing my excitement around the latest hype on ChatGPT from OpenAi. If you haven’t heard of it, I’m sure it’s only a matter of time before one of your friends or colleagues gives it a mention.

For background, ChatGPT is a type of LLM (Large Language Model) that has been trained on massive amounts of textual data. The current version of ChatGPT has been fine-tuned to respond with a variety of text-based formats, such as translations, Q&As, and more. The exciting part of this for many, including myself, is the model’s ability to produce “human-like” responses like we’ve never seen before.

As of right now, ChatGPT can respond to questions about everything and anything, whether these are topics related to work, translations, recipes—basically anything that you can think of up until 2021 (this is when the model was last trained or updated with information).

It’s been invigorating to see how this has captured the world’s attention, because it validates our approach here at Dialpad when it comes to AI. Leveraging and building with AI-related technologies is a core piece of Dialpad, and we’ve seamlessly infused this directly into every product we’ve created, whether it’s Ai Voice, Ai Contact Center, or Ai Meetings. Imagine—the power of generative AI, harnessed in your business’ customer support team to help answer everyday questions.

That might sound exciting, but I’d venture to say that the true potential of AI goes much farther.

Seeing ChatGPT in action

To give you a better idea of ChatGPT’s capabilities, I tried playing around with different questions to see the types of responses it would generate and whether this tool was truly revolutionary.

I started by asking, “How do you say ‘What is your name in Arabic?’” And the response was straightforward, it had the transliteral text, Arabic script, and a definition of each of the words (with also a caution of the phrase due to different dialects spoken around the world):

And if you haven’t noticed in the past month, ChatGPT can apparently code as well. I wanted to test what a response would look like if I asked for some coding to build out a webpage:

Not only was there a response with the code in Java, but it also gave a brief explanation on how to build out a webpage and where to find documentation and tutorials:

Pretty exciting stuff.

It’s cohesive—but is it coherent?

From my experience, ChatGPT-generated answers and responses to simple questions were easier to decipher if they were right or wrong.

For example, “What’s 2+2?” should always yield an answer of 4. But it’s possible to refute the obviously correct answer, which leads to ChatGPT backtracking on its previous response. Here, I couldn’t be sure if ChatGPT was providing the correct response or not:

Because ChatGPT generates text responses similar to the text that the model was trained on, depending on differing schools of thought, competitors, and different organizations, sometimes the responses may not be accurate.

This is something that we’re cognizant of as we develop Dialpad Ai further. What it means is that we’re making sure that our model will fact-check generated responses based on citations and third-party sources.

(For example, if your customer asks your website chatbot what your pricing is for a certain plan, you’d want that answer to be objectively correct!)

And what’s truly exciting for our team is that Dialpad has a significant amount of business communications-specific data, which will enable us to produce an AI model with even more accuracy and deep learning abilities for our customers. Imagine a generative AI tool like ChatGPT, except it’s 100% trained on—and specializes in—producing information or content about your specific industry or niche. Not basic or generic cooking instructions, not general translations, not writing poetry in different styles.

This is where we envision Dialpad as a customer engagement solution, one that provides accurate and consistent responses so your customers can get the best service and support possible. (More on this in just a bit.)

Where AI can help at work

In the customer engagement space where agent turnover is high, onboarding and contact center training is costly, and the need to improve first contact resolution (FCR) is a priority, tools leveraging advanced AI tools like ChatGPT can improve customer experiences and increase satisfaction.

With traditional basic chatbots, for example, customers may not be getting the responses they are looking for due to the limited technology or lack of AI. Dialpad’s conversational AI functionality is designed to improve on this—it can do a deep search, in real time, of all connected knowledge bases to automatically surface helpful information agents and sales reps while they’re talking to customers:

Not only that, Dialpad’s Ai CSAT feature can also accurately infer CSAT scores for customer interactions (on both voice and digital channels)—without needing to rely on customers filling out survey responses. This has always been one of the biggest challenges for businesses that want to leverage customer satisfaction scores, and again, it is solved by contact center AI:

Long story short: Companies that are investing in AI-powered contact center technology will have improved customer experiences and increased customer satisfaction. No doubt about that.

Dialpad Ai at Work

Here at Dialpad, all of our offerings have our proprietary Dialpad Ai capabilities embedded at their core. We’ve been in the generative AI space for years and have been making our customers’ lives easier with the advanced AI capabilities that we are constantly fine-tuning, improving, and innovating on.

With the launch of our Ai-Powered Customer Intelligence Platform, the industry-leading intelligent customer engagement and collaboration solution bringing together Ai Contact Center, Ai Sales Center, Ai Voice, Ai Meetings and Ai Messaging with built-in advanced Ai capabilities and Dialpad Ai Labs last month, our stance on AI has remained firm. We strongly believe that AI is the future of work—so much so that we have committed $50M towards research and development of AI to accelerate the development of cutting-edge features to automate business processes and provide predictive insights. (Check out the press release!)

There are many ways AI has already improved our lives for the better, from providing better driving experiences, to recommending to us the next best movie or sitcom to watch, to even saving lives in healthcare. Why not see for yourself how Dialpad Ai can improve customer experiences and increase CSAT scores at your company?

With the launch of ChatGPT and other AI models, who knows where AI is headed? The future of AI is exciting and we can’t wait to share more of our work in the LLM and generative AI space with you.