AI Is Not the Cure-All for Customer Service Woes
Artificial intelligence is receiving plenty of attention lately, due in large part to OpenAI’s ChatGPT technology, which can be used for everything from answering customer service requests to writing marketing copy, replying to sales emails, and writing code to create chatbots. But before customer service organizations invest too heavily in the technology all at once, Forrester Research urges a little bit of caution.
“Many customer service teams will be seduced by the containment potential of generative bots and will spend significant time, energy, and resources to be first to this party despite the many risks this technology holds for consumer-facing use cases,” the firm concludes in its “Three Customer Service Megatrends in 2023” report, one of a pair of recent reports discussing AI.
“The next year will be an opportunity for customer service leaders to prove their CX relevance, reframe their omnichannel delivery for a new era, and experiment with generation-defining AI,” Forrester says.
Though turning to automation when the economy slows is nothing new, this year the automation will include large language models (LLMs) and more intensive use of AI, according to Forrester.
In fact, fellow research firm MarketsandMarkets values the global generative AI market at $11.3 billion today and expects it to reach $51.8 billion by 2028, growing at a compound annual rate of 35.6 percent.
But companies would be better served by AI-based solutions designed to drive employee consistency, efficiency, and productivity in the face of the economic downturn, Forrester warns. “LLMs will not make all chatbots better. In the past decade, some brands have learned the hard way that conversational AI isn’t magic. The consequences of a thoughtless deployment of generative AI are much higher. Generative models bring their own risks, as they are prone to hallucinations and inherited bias.”
Though some companies will see “chatbots as a savior” in 2023, this can be a misguided effort, according to Forrester, which advises, “Stop trying to replace your agents with chatbots.”
It’s a conclusion shared in a second Forrester report. With it becoming harder than ever to design quality digital customer experiences, companies should avoid putting too much faith in chatbots, the firm says in the “Five Digital CX and Design Trends Stand Out in 2023” report.
“So-called conversations start with a cheerful welcome message but quickly crumble into disappointing experiences as it becomes obvious the chatbot can’t do much and mostly misunderstands what customers say.”
ChatGPT only compounds the problem, according to the research firm. “Although ChatGPT and its new sibling, Bing Chat (a.k.a. Sydney), sound confident, they’re sometimes inaccurate, illogical, and even delusional,” it says further.
Furthermore, the plateau for optimizing chatbots and natural language interactive voice response systems is 40 percent to 50 percent, according to the research firm. While some companies might be able to drive those percentages higher, those efforts are likely resulting in diminishing returns, Forrester finds.
“In 2023, brands will instead look to solutions that augment contact center employees to drive additional cost saving,” Forrester adds. “Brands will deploy augmented quality monitoring solutions to improve performance visibility and reduce the human effort of manually scoring calls. They will also tap generative AI solutions to drive agent efficiency, a much more appropriate application of this new technology than consumer-facing chatbots.”
And with the success of AI and automation depending so heavily on good knowledge management and data solutions, “in 2023, savvy customer service leaders will stop using shiny chatbots as a replacement for knowledge management and invest in creating a single source of truth,” Forrester says.
However, the compelling case for generative AI can’t be ignored.
Systems based on the technology, such as OpenAI’s ChatGPT and DALL-E-2, demonstrate that across various media with a lot of content, neural networks can do much of the heavy lifting.
Many designers and other creative professionals, therefore, rightly see generative AI as a surprising delight and a scary threat at the same time, the firm maintains.
The technology can supercharge the work of creators who seize its potential, Forrester says. “The content that generative AI systems churn out feels eerily humanlike much of the time and can be surprisingly effective.”
Yet the research firm cautions that the systems often fail, so human oversight is essential. “AI won’t supplant designers but will enable them to do their jobs better. AI will not replace you. A person using AI will.”
As AI enables design flows to be completed in a fraction of the time they used to take, designers can increase their corporate contributions, according to Forrester: “Design is no longer taking instructions from leaders who conceive and then dictate products and artifacts. Instead, it’s focused on applying design methods to influence strategy and help companies innovate better in a quickly moving technology.”
In a blog post discussing several other AI-related predictions for 2023, Forrester analyst Rowan Curran offers the following:
• TuringBots will write 10 percent of worldwide code and tests.
• One in four tech execs will report to their board on AI governance. AI is now an enterprise-essential, and as such, AI governance will join cybersecurity and compliance as a board-level topic. Forrester’s data shows that 46 percent of data and analytics business and technology decision makers seek out partners to implement AI critical to the business. Board reporting responsibilities will cover explainability, fairness audits of high-impact algorithmic decision making, and accounting for environmental impacts of AI (green AI).
• Ten percent of Fortune 500 companies will use AI to generate content. Human-produced content will never come fast enough to address the need for personalized content at scale, and in the next year, at least 10 percent of companies will invest in AI-supported digital content creation. Models for text, image, audio, and video creation will become an intrinsic part of daily workflows in both existing tools and new applications.
• A large majority of companies have already made significant investments in AI and are seeing positive results, which will only increase adoption even further. The pace of advancement in what is possible for enterprise applications is continuing to expand and accelerate, and this rising tide will lift all enterprises.