ChatGPT is very much the technological talking point of the year so far, making headlines for its realistic, human-like conversational experience.
The prototype – developed by OpenAI, an artificial intelligence research laboratory co-founded by Elon Musk – has recently been linked with Microsoft in a multi-billion dollar investment.
This technology has the power to transform our lives – not least our world of work. It has been used to perform a range of tasks from writing resumes and suggesting recipes to helping plan holidays and explaining complex topics.
But how will ChatGPT influence the customer service domain? We explore the benefits and risks of ChatGPT in the sector.
ChatGPT can demonstrably save you time, supplement your knowledge on a subject, and provide the heavy lifting for research tasks during customer service outsourcing.
Blair Strachan, from customer service outsourcing experts Kura, says: “We believe the advisor of the future is highly skilled with the expertise to resolve any complex problems.
“People love the ability to get the information they need, when they need it using automation. When things go wrong, that’s why the highly trained adviser will play its part, using empathy and knowledge to put the customers mind at ease.
“Technology will play a crucial part to remove less complex tasks.”
Increase worker efficiency and productivity
ChatGPT can quickly and effectively handle large volumes of repetitive customer service questions, which can enable human customer support to focus on more complex cases. AI tools can respond instantaneously, so users can avoid lengthy queues.
Couple that with ChatGPT’s 24/7 availability, where responses can be sent even outside of business hours. Should a query fall outside of a human agent’s sphere of knowledge, ChatGPT can provide a handy prompt.
Doing the heavy lifting
Following on from this, ChatGPT is effective in providing a summary of complicated information. Its in-built AI was trained using words from the internet to automatically generate responses which can then be relayed to customers.
The model includes an impressive 570GB of data obtained from books, webtexts and other internet articles. Customer service teams can expend a lot of energy categorising incoming questions into relevant groups which ChatGPT can do automatically.
Perhaps surprisingly, ChatGPT can provide more of a personalised response to customers. The tool is programmed to detect and analyse sentiment.
ChatGPT can use customer-specific data about their past interactions to generate responses tailored to each individual customer.
ChatGPT is certainly a handy tool but not without its limitations and shouldn’t be considered a complete replacement for human customer service agents.
And as Blair adds: “While we’re naturally curious about the changing nature of automation and AI, we’re constantly thinking about the people inside our business, who are helping our customers every day.”
ChatGPT is not always factually correct
Arguably the biggest risk to an organisation using ChatGPT in customer servces presently is that, as a chatbot, it can provide answers or information that is not actually 100 percent helpful or even accurate.
With customer services often being the chance for a company to repair any damage to its reputation in the eyes of a customer, the last thing the organisation needs is to provide a negative experience.
Despite generating responses with complete confidence, we cannot always gauge what ChatGPT knows and what it doesn’t, and the tool is not immune to false information.
ChatGPT’s database was collected from countless sources on the internet, including pages that include false or harmful information, and does not contain any information provided beyond 2021.
Since ChatGPT is fundamentally an AI language model, it cannot always filter information competently and has even been known to generate violent, sexist and racist content. This could be detrimental to your customer base who may be offended by its responses.
ChatGPT could pose a security risk
Two potential risks of ChatGPT, as with any automated chat service, is that it could be manipulated by prospective scammers, and it could also leave itself open to possible data breaches.
As the tool is designed to follow the user’s commands, this could leave it vulnerable to security risks when used in a business setting.
Scammers for instance may try and trick it into issuing refunds or providing credit. This could have negative financial implications for your business or could undermine customer trust.
It is important to note, however, that in its current form this is not a risk for a business to be particularly concerned about.
With ChatGPT being at present, just a chatbot, the risks regarding scammers and potential data breaches are only likely to come to the surface should companies choose to implement it as part of their automated customer services – this would then open the door for the possibility for customers to be lured into a phishing scam where ChatGPT can convincingly impersonate the organisation.
Disruptor of the decade
While it is certainly in vogue, whether ChatGPT proves to be the grand disruptor of the decade remains to be seen.
Further down the line, particularly should organisations choose to use this as part of their automated customer services, there may be issues that arise regarding security and the potential for phishing. Presently there is the issue that the database is just that – a database.
However, despite the issues, it would be foolish to ignore its potential benefits for the customer service arena when used responsibly.
The ability to personalise responses, to allow more productivity elsewhere in the workforce, and to respond 24/7 are all benefits of using the technology in the customer service sector.
Fear of AI and generative AI behind EU’s record $1.3 billion Meta fine, says GlobalData
Qualcomm and Microsoft align efforts to scale on-device AI at company event
AI won’t replace you, engineers using AI will, says Monolith CEO