It has become impossible to avoid the media coverage of ChatGPT. The topic has gravitated from scholarly articles, to traditional media, and is now the darling of the Dinner Party crowd. Few people would express themselves as experts in AI, but attractive opportunities are emerging from the debate and the discourse.

A recurring theme with AI discussions is the seeming inevitability of robots replacing humans in our jobs. The Guardian published a podcast with a similar tone in February. They progressed the debate, suggesting my focus on customer service and experience (CX) is finished because the dreaded “bots are about to take over our beloved industry”.

I believe that generative AI, such as ChatGPT and Google Bard, will undoubtedly  make a significant (and positive) change to the way many industries function – including my own. The evidence does not support the conclusion that we are at, or close to,  the point where humans are no longer required to deliver exceptional CX, however.

ChatGPT, for example, is a tool that is likely to have the greatest impact in highly complex activities. In much the same way that the introduction of the spreadsheet simplified challenging calculations and forecasts, the ‘commoditisation of the complex’ is on the horizon. ChatGPT is a demonstrably advanced step-change in our tool box, but it is still a tool which requires human sensibilities to realise its potential. The ability to imagine, to leap beyond logic, is not yet in reach for the latest technological wunderkind.

Let’s consider my industry first. Customer service requires a lot of people in contact centres answering customer calls and chat messages. Many companies have tried to introduce chatbots over the years to reduce the number of calls by answering simple questions. The awkward truth is, however, the bots have never really performed satisfactorily. The first thing most people try is skipping past the bot so they can talk to a human.

ChatGPT may represent a genuine opportunity to introduce customer service bots that are helpful in a way that we have not seen before. This will help customers to get the answer they need faster. Similarly, it will improve the work of the customer service agent, leaving the simpler queries to the new Chat’GPT’bots (remember where you read that first), providing latitude to focus on the tougher problems that need real -and human– investigative skills. 

Prima facie, this seems quite a positive outcome. At last, a job in customer service may finally become a respected and skilled profession. Like an engineer, only someone with deep product knowledge and great interpersonal skills could engage in this varied style of customer support. Humans will be seen as experts…not faceless, emotionless, insensitive, tech.

It’s worth explaining a little about how ChatGPT works, without diving too deep into the AI and complex algorithms. The easiest analogy is to compare a generative AI system to Google. When you search Google for information you get a page full of the most likely websites where you can find the answer. You still need to go to those pages to look around for more information to find your answer. A generative AI system scans all those pages, finds the most useful information and then creates your answer.

Here is a practical example to make this really clear. A Google search using the question ‘what is the secret recipe for Coca Cola?’ will deliver  these results; pages from Wikipedia, the Guardian, YouTube, Quora, and even the Coca Cola company. 

Ask the same question of version 4 of ChatGPT, and the answer you get is much more engaging;

“The exact formula for Coca-Cola is a closely guarded secret, known as “Merchandise 7x.” The recipe is kept in a vault at the World of Coca-Cola museum in Atlanta, Georgia. While the complete list of ingredients and their exact proportions are not public knowledge, some of the primary ingredients in Coca-Cola include carbonated water, high fructose corn syrup (or sugar in some countries), caramel colour, phosphoric acid, natural flavours, and caffeine.”

Now lets be human, an leap beyond the logic of ChatGPT. Imagine how this can transform the way a customer searches for help when they have a problem with a new purchase. The customer should be able to ask a clear and direct question and then get an answer that is generated from product information – not just a link to a page.

The industry analyst Gartner recently published practical examples of ChatGPT across various industries:

• Parts design: allowing manufacturers to test parts by using simulations – such as testing new car bodies without needing to build models and optimising the design.
  
• Chip design: helping to design the floor planning for chips and therefore moving product design from weeks to month.
• Drug design: drug discovery can take years and cost billions. If AI can help to change years into months then the value is obvious.

What connects all of these examples…and conversely, expresses the disconnective qualities of this new technology in practice?

ChatGPT and other generative AI tools are productivity tools. They can generate new content, but it is based on a body of existing content. It has to learn from something. So the best way to think of these tools is like a smart search engine that not only finds where the answer to your problem is, it can summarise it too.

Not everyone will be immune from the influence of ChatGPT. Some professions that are focused on searching and summarising information, such as paralegals asked to locate legal precedents, may be more directly affected by these tools. 

Generally, however, most professions still need a human as an integral component of the mechanism. This is clear from the customer service example I gave earlier. Most customers would love better chatbots, but many still need to speak to a human when they have a problem.

AI is changing how many of us do our job, but that’s been a constant evolution throughout my career (fax machines, anyone?). Evolution is not over. We can expect more ‘great leaps forward’ from ChatGPT and its technological cousins. 

Robots are becoming more useful, more capable, more prolific. ChatGPT will do what tools have been doing for humans since the first prehistoric flint was fashioned; they will  allow people to reach new goals, new heights, more effectively than before.

‘The Day of The Robots’ is not here just yet…