The Emergence And Shortcomings Of Chatbots In Customer Experience

Over the last few months, AI-driven (Artificial Intelligence) LLM (Large Language Model) applications have come into the limelight. These applications have shown massive use and capabilities in industrial, academic and creative fields. Among these and probably the most famous, is ChatGPT, a chatbot that reportedly attained 1 million users within 5 days of its launch.  

According to a report by Juniper Research, LLMs may soon become an integral part of customer interactions. The firm predicts that by 2024, such chatbots will handle around 70% of interactions with customers.

Chatbots are learning to interact in a more human way with quicker and clearer turnaround time. This has made it popular with businesses looking to better their CX (customer experience). Additionally, businesses are using them to improve marketing strategies and enhance personalized services.

While speech recognition software and NLP (Natural Language Processing) have been used in customer experience processes since before, experts believe the new technology chatbots could revolutionize customer experience in the future.

According to Sean Mullaney (CTO, Algolia - a search engine SaaS platform), “LLMs are fundamentally changing the way search algorithms work.” He added to this by mentioning that where older search engines match individual words to indexed content, LLMs essentially understand the meaning of the words to produce better results.


However, sometimes chatbots have trouble understanding complex requests or questions, which makes them seem inflexible in customer experience, resulting in the intervention of humans.

According to a survey conducted by Conversica (an AI company), the experience provided by first-gen chatbots doesn’t match customer expectations. The survey found that four out of five buyers ditch the chat experience when the chatbots can’t cater to unique requirements.

Jim Kaskade, (CEO, Conversica) said, “First-gen chatbots rely on predetermined scripts that are tedious to program and even harder to maintain,” and added, “In addition, they don’t understand simple questions, and limit users to responses posed as prewritten messages.”

Pieter Buteneers (Director, Sinch Labs) said, “The problem with models like ChatGPT is that ChatGPT ‘memorized’ everything it could find on the internet into only 175 billion numbers (5,000 times fewer than the human brain). So, ChatGPT is never 100% sure of the answers it gives you. It is impossible to remember every minute detail, especially if we’re talking about storing all the knowledge on the internet. In every situation, it will just blurt out the first thing that comes to mind.”


Recently, in a promotional video for Google’s Bard (Microsoft-backed ChatGPT’s competitor) the chatbot gave a partly wrong answer to a question (read here). This saw a $100bn+ devaluation and a 9% share dip in Google’s parent company, Alphabet Inc.

The question asked was “What new discoveries from the James Webb Space Telescope (JWST) can I tell my nine-year-old about?” The incorrect part of the answer suggested that the JWST took the very first pictures of a planet (or exoplanets) outside Earth’s solar system.

Despite how far AI-driven applications have come, they still have room for improvement before they can properly mimic human behavior. What do you think of LLMs and their involvement in customer experience? Let us know in the comments below!