In recent years, few technology innovations have created the amount of buzz and hype that currently surrounds ChatGPT, the conversational AI created by OpenAi. And it doesn’t take long to realize that its open-source natural language processing model which utilizes the most recent developments in deep learning and artificial intelligence is more than just an automated chatbot. At first and at second glance, openAI’s latest creation has all the traits of a disruptive technology. A technology that is now commonly referred to as Generative AI.

Using natural language processing (NLP) technology and capable of generating human-like text based on the input it receives and the information it has been trained on, it’s designed to – and I quote ChatGPT’s own answer on how it would describe itself – “respond to a wide range of topics, including questions about various subjects, language translation, and creative writing. [Its] goal is to provide accurate, concise, and helpful responses to users.”

ChatGPT – The Holy Grail for Business Success?

But does ChatGPT really deliver on this promise? Looking online, it certainly seems like it. It’s only been a little over two months since ChatGPT became available to the public and already the web is full of tutorials and how-to guides promising readers insights into leveraging the AI’s mind-blowing capabilities. Just the sheer number of articles could lead one to think that unlocking ChatGPTs full potential is synonymous to finding the Holy Grail of business success: a clear and easy path to driving growth – no matter which industry.

Of course, as is often the case with hyped up technology innovations, such promises are unlikely to be kept one hundred percent. Even openAI’s CEO Sam Altman admitted as much when he recently tweeted:

ChatGPT is incredibly limited, but good enough at some things to create a misleading impression of greatness.

it's a mistake to be relying on it for anything important right now. it’s a preview of progress; we have lots of work to do on robustness and truthfulness.

While some might find Altman's strict assessment surprising, especially given ChatGPT's recent success stories. After all, the AI helped to write a children’s book, generated intricate Python code and not only composed college-level essays, but also passed exams from law and business schools. Still, it's always nice to get an honest take from a (tech) CEO once in a while. Not least because the potential downsides of Generative AI have already been identified. For example, Eric Ulken, product director at the publisher Gannett, has a point when he recently published his opinion on the “disruptive potential of the machine-creative revolution”: 

“For all its promise, generative AI can get more wrong, faster – and with greater apparent certitude and less transparency — than any innovation in recent memory.”

Being aware of ChatGPT’s potential downsides is essential when using it. After all, it doesn't search the internet for solutions as Siri or Alexa do. Instead, it builds a sentence word by word, choosing the most likely "token" to appear after each one, a process based on its prior experience. Educated guess is the term that comes to mind here and it hints at why ChatGPT sometimes argues obviously incorrect responses as if they were entirely accurate.

Can ChatGPT Really Reduce Costs and Improve Customer Experience?

However, rather than spending too much time lamenting these downsides, I thought it would be equally interesting to look at the clear and immediate gains to be had from using ChatGPT. To do that, I decided to focus on a specific industry and examine specifically how ChatGPT can be used to reduce costs and improve customer experience. 

However, to answer that question, I decided on an unusual approach. I Started not by doing my own research. Instead I took a shortcut: I consulted ChatGPT first.

The industry I chose was the automotive industry. It’s a sector that’s currently experiencing its own major technological and digital transformation driven by technological advancements, with electric vehicles being the spearhead and the major driver of that innovation. This shift towards EVs is not only affecting the manufacturing and design of cars, but also the way they are powered, sold, and serviced. Additionally, the rise of connected and autonomous vehicles is changing the way people interact with their cars and the transportation industry as a whole. 

Sounds like there must be some way ChatGPT could be of help here. And unsurprisingly, the AI agreed and quickly spit out a four part response when asked the question “How can ChatGPT be used to reduce costs and improve customer experience in automotive?”. The answer was:

The response sounds convincing. But does it make sense, too? Let’s take a closer look at the four areas of application that ChatGPT identified.

Automating Customer Service 

Leveraging ChatGPT’s high level of conversational skills seems like a no-brainer when it comes to helping businesses streamline customer conversations and automate responses.

Its natural language processing capabilities appear sophisticated enough to enable organizations to deliver speedy and efficient customer assistance while maintaining quality. And it’s a level of quality that I have not yet seen in any of the “regular” chatbots that I encountered on the web in recent years.

Responding to frequently asked customer questions, such as product details, costs, and availability, should be a rather easy task for ChatGPT. But it could be a very valuable one, with 24/7 support substantially improving the customer experience and freeing up staff to concentrate on more complex customer problems. 

Increased efficiency could be another factor, since the AI should be able to process and address a multitude of customer interactions simultaneously, very quickly responding to multiple inquiries at once.

While it seems plausible that ChatGPT could reliably assist with standardized processes like product returns or exchanges, it’s questionable in my opinion how helpful ChatGPT can be in handling detailed complaints or resolving more complex customer issues, like technical problems. In the end, it all depends on how well the AI was trained on the specifics of the respective business and product portfolio. 

Personalizing Marketing and Sales

Leveraging ChatGPT in order to personalize marketing and sales is the most obvious area of application. Of course, marketing and sales are different from industry to industry, but there are enough commonalities that make the use of the AI seem reasonable.

In the automotive sector, ChatGPT may, for example, inquire about the sort of car a buyer is looking for, their budget, and desired features. Based on this data, ChatGPT could provide a selection of automobiles that suit the customer's requirements and give extra information about each one. This can enhance the customer experience by making the purchasing process more efficient and entertaining.

ChatGPT could also assist clients throughout the buying journey, from initial research to final purchase. It may, for example, answer questions about financing choices, help with test drives, and manage any post-purchase issues. ChatGPT could also help to grow customer satisfaction and increase the chance of repeat business by delivering a high degree of customized service.

Like with the automation of customer service, there are concerns to be considered when using ChatGPT for personalized recommendations and support in the automotive industry. Potential challenges include making sure the AI complies with data privacy and security regulations – an issue which is of particular concern in the automotive industry where sensitive information about vehicles and their owners is being collected and processed.

Another concern is the accuracy of personalization. ChatGPT would need access to detailed and up-to-date customer data in order to make accurate recommendations. Because faulty or misleading personalization would make the customer experience worse, not better.

Streamlining Service Processes

 

At first glance, streamlining service processes seems very similar to the automation of customer service. ChatGPT even admitted as much when asked about this similarity, but then pointed to an aspect that it sees as a key difference. Here’s its answer:

“Automating customer service refers to the use of ChatGPT to handle customer inquiries and provide support related to general information about the company, its products and services, and frequently asked questions. This can reduce the need for human customer service staff and provide 24/7 support to customers.

Streamlining service processes, on the other hand, refers to the use of ChatGPT to assist with specific tasks related to servicing vehicles, such as scheduling appointments, processing payments, and gathering information from customers. The goal of this application is to reduce manual work and increase efficiency in the service process.”

In contrast to the initial response, this follow-up explanation underscores that both points address different aspects of the customer journey: customer service and the servicing of vehicles. As a consequence, both have different goals and outcomes.

Based on this response, it’s easy to imagine that with regard to the servicing of vehicles, ChatGPT can support organizations in scheduling tasks by allowing customers to plan their own service appointments, thereby removing the need for phone or in-person interactions.

The “payment processing” aspect was a bit harder to grasp. But hey, it’s a conversational AI. So I simply asked how exactly ChatGPT would assist with processing payments for services. The answer, again, was quite enlightening:

• Providing payment options: ChatGPT can use conversational AI to provide customers with information about the various payment options available for services, such as credit/debit cards, online payment platforms, and others.

• Processing payments: ChatGPT can be integrated with payment processing systems and use conversational AI to guide customers through the payment process. For example, ChatGPT can ask the customer for their payment information, confirm the details, and process the payment.
   

• Handling payment inquiries: ChatGPT can use conversational AI to handle customer inquiries related to payments, such as answering questions about billing, resolving payment issues, and providing receipts.

Sounds legit, even if some of it only seems possible if ChatGPT is integrated with other specialized technology. 

On to information gathering, a task where I had less difficulty seeing ChatGPT succeeding. After all, collecting information from customers concerning car problems in a conversational style shouldn’t be rocket science. What possibly makes it attractive is the fact that it has the potential to lower the time and effort necessary to diagnose problems. That is if the AI is trained well enough on different kinds of automobile malfunctions and how they manifest.

Let’s take a look at the potential concerns to be considered when using ChatGPT or other conversational AI systems to streamline service processes. I’ve already hinted at the integration aspect above. Current systems and procedures, such as customer relationship management (CRM) systems, appointment scheduling software, and payment processing systems, need to be connected with ChatGPT for it to operate effectively. 

In addition, data accuracy and completeness are key. The Ai can only provide comprehensive and efficient support if customer data is properly collected, stored, and maintained.

A third concern is that of error handling. In contrast to personalization of marketing and sales, mistakes in payment processing can get costly fast. So robust error handling and recovery processes need to be in place to ensure that issues can be quickly and effectively addressed.

Enhancing Product Development

The short explanation by ChatGPT in the screenshot above already hints at a very important aspect of this task: the training of ChatGPT on customer feedback. That training, in turn, can enable a three step process:

1. Gathering feedback from customers about the vehicles they own or have used, including opinions on design, performance, and other factors.
2. Analyzing that feedback and extracting insights like common themes and patterns in the feedback. 
3. Generating reports that provide a clear understanding of customer needs and preferences in order to inform product design, engineering, and marketing decisions.

Utilizing ChatGPT to collect and evaluate client feedback enables automotive companies to better understand what their consumers want and need. Offering products that are more in line with consumer preferences can lead to increased sales and a high level of customer satisfaction.

As with the other three tasks, there are also a number of problematic issues when using ChatGPT for product development insights in automotive. The first two that come to mind are old acquaintances: ensuring the data and feedback obtained are accurate and unbiased as well as taking all necessary steps to address privacy and security concerns.

As product development is a evolving process, one would think that if an AI is regularly involved in this process, it also needs to be able to learn and evolve to meet the changing needs of the industry. That, in turn, requires continuous training on the latest data and information to ensure the AI improves as well.

The Verdict

In my opinion, there’s no doubt that ChatGPT can be helpful in reducing costs (by enhancing efficiency) and improving customer experience in the automotive industry. But to accomplish this task effectively, it should be used exclusively according to its strengths. Looking at the four areas discussed above, the automation of customer service and the personalization of marketing and sales activities seem like the best applications for a conversational AI in automotive – at least right now.

Over time, when ways have been found to mitigate the potential problematic issues, ChatGPT has tremendous upside in the streamlining of service processes and the enhancement of product development, too. But for now, certain risks like for example errors in payment processing, or potentially high integration costs, waiting and observing the further development of the technology seems like the right choice. 

And the fact that ChatGPT’s knowledge cutoff is the year 2021 rules out the AI’s relevance for tasks that require it to include knowledge, developments, numbers, statistics, or anything else that is recent and new.