Due to the success of ChatGPT and the media coverage it’s received, AI is currently receiving a lot of attention. This is starting to generate a tsunami of new AI solutions and apps which are flooding the CX and digital marketplace. But how many of these solutions and applications will truly provide value versus how many are just riding the wave of hype surrounding Generative AI?
According to Jay Sanderson, Global Product Marketing Manager for Progress who is based in Australia , Generative AI has the potential to be more significant than the introduction of the web browser or Graphical User Interface (GUI), but he advises, “The introduction and impact of technology like this has three phases. Firstly, there’s the short-term or hype phase. Where everyone is excited about what it can do but doesn’t understand its full potential or how it can be applied. Secondly, there’s the mid-term phase. Where applications have been built or where generative AI has been incorporated into existing solutions and applications. Finally, we have the long-term phase or vision of where this technology will lead, which I guess no one really has in their head at present.”
Various industry figures, including Elon Musk have advised that research and development into AI should be paused because of the potential risks it poses to society. It’s hard to see this happening as companies are keen to get products into market to gain advantage over their competitors. In the last three months, AI companies have received over $US 16 billion in funding. A whole herd of new AI startups is predicted to emerge in 2023 as well as existing technology players incorporating AI into their applications and technology platforms.
The customer-facing departments and operations of companies today are likely to look very different in five to ten years’ time. It may prove to be, however, a complex path for organisations to navigate through the various technologies and vendors that are emerging and decide which ones to adopt.
Generative AI versus Conversational AI
Conversational AI has proven its ability to automate and enhance customer interactions. Relying on massive amounts of data, machine learning, and natural language processing, it can engage users in human-like conversation through virtual assistants or chatbots. Generative AI such as ChatGPT, says Sanderson, can greatly enhance the capabilities of chatbots and existing conversational AI tools. “The ability to create content and text on the fly in response to questions from customers can greatly improve the experience and the quality of customer interactions”.
Generative artificial intelligence (AI) has the potential to revolutionise the CX and contact centre industries, offering customer experience enhancements that improve efficiency and effectiveness, while reducing the need for live agents to handle mundane tasks. Sanderson comments, “Building generative AI capabilities in a chatbot or virtual assistant will not only allow it to respond in a human-like manner, but it will also be able to personalise its responses to the specific needs of each customer.”
Generative AI is quite capable of understanding human language and responding to questions on a broad range of topics. Sanderson says, “ChatGPT has the potential to augment what customers are already doing and make customer interactions stronger. ChatGPT has to be plugged into other existing platforms. Marketers and CX practitioners should look at generative AI tools already embedded into existing conversational platforms so it can lend itself better to composability.”
Most chatbots today present information stored somewhere else and can only handle very specific as well as a limited range of tasks before they need to transfer to a live agent. A chatbot that incorporates generative AI has the potential to handle far more complex cases and enquiries. Vanvisa Brown, Head of Marketing at Curious Thing explains, “Current Conversational AI is often rule-based, making it hard to understand and respond to human needs. With Generative AI, however, AI assistants can better understand the nuances of human situations and give personalised responses. This means Generative AI has the potential to transform how businesses interact with customers by providing more helpful, human-like responses to customers’ needs.”
The challenges
To take full advantage of generative AI in terms of marketing and customer experience, organisations need to address a range of challenges and limitations. Sanderson advises,” AI bias, privacy and security need to remain top of mind because they will be challenged more than ever. Marketers and CX professionals need to be mindful of how AI-generated content will be received – is the content actually talking to the right audience? Is it understandable? Could someone find it offensive?”[2]
According to a recent research report released by Progress, Data Bias: The Hidden Risk of AI research, conducted by independent research firm Insight Avenue, 66% of Australian organisations suffer from data bias. 84% of the 640 business and IT professionals interviewed believe they need to be doing more to understand and address data bias in their organisation.
In the past, many organisations have been reluctant to adopt AI in their contact centres, for fear of putting a machine between them and their customers. Interactions with chatbots and AI were less than satisfactory. But over the years, as chatbot technology matured, the situation has improved.
According to recent research, 87% of consumers currently have neutral or positive experiences with chatbots. Brown says, “Technology has become smarter in recent years, with advancements in speech-to-text, text-to-speech, and artificial intelligence that can understand human conversation and generate responses. This evolution will likely continue and further improve.”
“We are currently seeing a whole lot more curiosity from the public and from businesses who previously have gone; no, we don’t trust this, we’re not interested. Their initial experiences with AI and chatbots were really bad. I think the tone of the conversation has shifted where people are opening up to, maybe, robots don’t have to be a terrible thing”,
Relying solely on AI-generated content without human control and supervision can lead to inaccuracies, irrelevant content and some very unusual situations and interactions. Brown adds, ‘while chatbot technology has improved, it’s important to build them with appropriate business rules and data sources to avoid irrelevant responses.”
Like with the first introduction of conversational AI and chatbots, the poor implementations of generative AI may have significant consequences for brands and the experiences they provide customers. Sanderson says, ”Are brands trying to replace human interactions? This is dangerous territory. The human side of customer experience is important and we need to keep connections and empathy alive, including accessibility and social impact”.
The potential of generative AI is huge but so are the consequences if you get it wrong. Sanderson highlights “attackers may generate new and complex types of malware, phishing schemes and other cyber dangers that can avoid conventional protection measures. Such assaults may have significant repercussions like data breaches, financial losses and reputational risks”.
He also raises the concern of an overabundance of content and information that will place increasing pressure on marketing departments to perform. “Creating text, images, hyperlinks on the fly in only a few minutes will put more pressure on marketers. If ChatGPT can create a blog in 5 min, how much acceleration of the pace marketers will be able to cope with?”
“Marketers and CX experts have to become “prompt engineers” – AI is as good as the cues it’s being given. AI will save marketers’ time on one end, but on the other they need to invest time in learning how to be good ‘prompters’ – how to ask AI questions, how to interact with it, how to correct it, etc.”
At present, the use of generative AI in the context of customer experience is in its infancy but most CX technology providers are integrating generative AI functionality into their software and platforms. The potential impact of these innovations is likely to be enormous. Most people, however, are still investigating and experimenting with what it can do.
Regarding Sanderson’s three phases, the technology is somewhere between phases one and two. You need to beware of the potential dangers and risks associated with generative AI and as do as much as you can to reduce the likelihood of unintended consequences. Don’t try to do too much at once and be really clear on the CX and business goal you are trying to achieve.
