Recall, if you saw it, Joaquin Phoenix falling in love with his operating system in the movie “Her”. Where “Her” is fictional, intelligent computing is no longer science fiction and the top eCommerce players know it.
When customers visit an eCommerce website, they want to find things easily and they want your replies to be fast – and unfailingly efficient. They wish as well to feel listened to and understood, and then cared for and assisted. What customers want is human empathy matched by machine efficiency, the sort of service that makes successful and enduring business-customer relationships happen. Conversational Artificial Intelligence is getting closer and closer to that goal. Soon enough, a computer will most certainly fool us into thinking it is human.
Kompyte has recently conducted an assessment of conversational AI in eCommerce, generating a benchmark measuring the efficiency of a given conversational AI. The study of ten key eCommerce Conversational AI vendor solutions found that specialized AI start-ups outperform big tech companies. Discover how and why in this article or download the study itself if you want to go straight to the details.
Conversational AI Resolution Rate, the most important customer service KPI for AI in eCommerce
According to Office for National Statistics, eCommerce grew by 46% in 2020 and with it the need to provide a better customer experience. In this highly competitive market, the best players are those able to understand and identify what customers want – behind their emails, chatbot conversations, or social media comments. Success in this environment depends upon reliably getting at the intent of your customers. Simply put, when a user searches for “large industrial desk” at a furniture design eCommerce website, their intent is to retrieve a selection of desks.
The “Resolution Rate” is a key indicator on how efficient a given AI solution vendor’s model is at understanding human intentions. The higher the rate, the better the AI solution, the better the conversation between your customers and your website, the greater their satisfaction.
An ounce of prevention is worth a pound of cure. Measuring your eCommerce resolution rate is a good way to measure and assess customer experience. A resolution rate of less than 80% indicates the strong likelihood that user experience is suffering as your AI model is statistically less likely to accurately understand and translate a query and correspondingly less likely to successfully match it to a system-generated reply.
Master your resolution rate and expand your eCommerce
Drive acquisition and maintain retention levels
Achieving high resolution rates with conversational AI is a key differentiator between successful and unsuccessful adoption of conversational AI in eCommerce. In a competitive and myriad environment such as that of eCommerce, it is becoming harder to grab and retain the attention of customers. Doing so reliably truly depends on how user feels while browsing through your pages or talking to your chatbot.
In the race to provide the best customer experience, having a high resolution rate will ensure:
- Better conversations: A good AI in eCommerce enhances chatbot ability to imitate human agents in conversation. Visitors have the feeling of talking to an actual human.
- Higher conversion rates: Statistics show that from fifty visitors you usually can expect one customer. By improving the quality of the conversations, to better anticipate and fulfill your visitors’ requests, you are more likely to turn your conversations into conversions.
- Improved customer retention: Complex requests can be referred to an agent who can focus on them, while chatbots resolve simpler requests. Quick responses, fast handling, and effective guidance establish trust, boost user experience and therefore customer retention.
Skip the training and get faster time-to-market
Some AI vendors require you to train your model. Take the example of chatbots: just as humans go to school to learn how to talk and how to learn specific skills, certain chatbots need a learning period to improve their communications skills. It takes time and requires building a comprehensive knowledge base, including FAQs, synonyms, compound words, and even some personality elements.
Not all conversational AI vendors require that training phase. It is possible to have an excellent resolution rate with minimal to no training of the model. These have the advantage of faster time-to-market and time-to-value by the following:
- Lower setup costs: Use of advanced pre-trained models skips the training phase of the model, saving time and money.
- Streamlined AI training: Benefits from a quicker optimization of the tool and thus a faster customer adoption.
- Reduced maintenance cost: Get a more stable solution from the start with less monitoring required.
- Higher Scalability: Provide your customers with increasingly complex conversation, and more content, faster.
When evaluating an AI vendor, its technology and service quality, these two criteria will make the difference:
- Resolution rate: How successfully does it understand what a human is asking from it?
- Time to value result: How much training of the model is required?
Specialized conversational AI vendors versus Big Tech companies
A conversational AI benchmark in eCommerce (order taking, shipping, and payments) conducted by Kompyte between December 2020 and January 2021 evaluated the resolution rates of ten Conversational AI vendors on the market, from the biggest tech companies (IBM Watson or Microsoft Luis) to smaller specialized AI start-ups.
Large tech companies – IBM, Google, Microsoft – all underperformed, with resolution rates as low as half those of specialized AI start-ups. First-place Inbenta easily achieves leading accuracy in imitating human interaction and demonstrates the small-firm advantage of focus and core competency in conversational AI.
Inbenta’s Symbolic AI – the real game-changer in the eCommerce industry
Inbenta scored the highest rate (84%) across all topic categories (order taking, shipping and payments), with the best capabilities to detect and translate interactions to modeled intent. They scored consistently above the 80% resolution rate threshold, at minimum 10% ahead of the other AI providers studied. Inbenta is clearly a potential game-changer for the eCommerce industry.
Why does Inbenta outperform big tech companies and other AI start-ups? Not only has Inbenta specialized in conversational AI from its creation, but also uses NLP Technology powered by Symbolic Artificial Intelligence. Where other vendor solutions are based on machine learning, the critical element of Inbenta’s success is Symbolic AI. Symbolic AI embeds human knowledge and behavioral rules into computer programs. This technology enables any tool to understand and anticipate user intent – regardless of their vocabulary – and to provide those users with fast and accurate replies for a more positive, efficient, and successful user experience.
Want to know more about how to boost your resolution rate
and push your customer service to the next level?
Fill the form below and download the
“Conversational AI Benchmark in eCommerce” study by Kompyte:
Check out our similar articles
- What Is Predictive Search? Reasons & Tips to Implement It
- 10 Cool Chatbot Examples to Inspire Your Project
- How to Increase Conversions thanks to E-commerce Search Automation