‘Hey Alexa, can you play Viva la Vida by Coldplay?’
‘Hey Siri, can you set an alarm for 2AM?’
How about — ‘Hey Buzzo, can you find me a short sleeved blue polo t-shirt with a printed pattern and a matching trouser?’
According to a global study conducted by Wunderman Thompson in the year 2020, 27 percent online shoppers have used or currently use voice assistants to make a purchase. In the US and the UK alone, voice-based shopping is estimated to gross over $40 billion by 2022 — driven primarily by Amazon’s smart speakers. This essentially means consumer spending through voice assistants is expected to account for a sizeable market share of 18 percent.
Wondering how voice search can command such a massive presence in such a short duration? Well, the answer simply lies in human nature. On a typical day, an average person speaks about 16,000 words. Speaking is sometimes as involuntary as breathing. With increasing proliferation of smart speakers and devices such as the Amazon Fire TV Stick, the search for entertainment is already witnessing a massive shift towards voice. But why limit it to entertainment?
People love shopping online… or do they?
Online shopping is already ingrained in modern lives to an extent that groceries, food, apparel, furniture, cosmetics, electronics and almost everything else is being ordered online and no age group is resistant to this change anymore. What started with books has penetrated almost all categories of purchases today. However, this shift from offline to online shopping left one major gap — the trust and confidence a customer gets from the expertise of a knowledgeable salesperson.
A typical user’s online shopping journey:
Today, a typical customer spends several hours and sometimes even days, conducting research about a product to purchase by carefully reading the fine print, tech specs, hundreds of reviews and watching several YouTube videos to arrive at a decision. Moreover, online shopping in its current form often leads to a situation wherein customers get overwhelmed with information and choices, and ultimately abandon or delay the purchase. According to a recent survey, 46 percent shoppers stated that they have failed to complete a purchase just because there were too many options to choose from. While several consumers enjoy this DIY nature of online shopping, many still feel the need for a product expert to guide through their purchase decision.
Enter Buzzo: AI Shopping Assistant to power voice commerce on apps & websites
Imagine online shoppers being able to speak as freely to a voice assistant as they would with an in-store expert to search for relevant products, gain recommendations/guidance, filter and sort results based on their preferences, compare options, get specific product information, and find matching or complementary products to complete their shopping experience. Wouldn’t a shopper just feel awesome to be able to state multiple instructions simultaneously to their e-commerce app and let it find the most relevant products just as a human would physically?
This is exactly what the AI Assistant does! It can transform your customers’ online shopping experience by saving time and effort, and acting as their personalised assistant. It offers buying guidance even for someone with a zero frame of reference; personalised results as per a user’s stated preferences; insights deeper than just product specs; and super quick, multi-layered filtering. In addition, it bears a human-like persona to deliver a shopping experience as close to the real one as possible. Let’s talk about how some of these features work.
Suppose a fitness enthusiast is looking to buy a fresh pair of yoga pants, and tells the AI Assistant, “I’m looking for yoga pants”. This query prompts the Assistant to provide guidance around the key parameters customers consider before buying yoga pants and help set a frame of reference in case the user does not have one. For instance, in this case, it displays the pros and cons of major fabrics used, i.e. cotton and polyester, which essentially serves the purpose of educating the user how she should go about her search.
In parallel, the Assistant prompts the user with a question asking her preference towards fabrics while showing the most commonly selected options based on past data available. This prompt is aimed at having the user apply quick filters to narrow down the search. Further, in order to encourage the user to continue using natural language, it recommends hints as answers to the question, such as “suitable for summers” or “wearable in chilly mornings”.
The buying guidance feature essentially tries to replicate the real offline experience of a shopper, where an in-store expert tries to determine the best suitable product for a user by asking questions around a customer’s preferences and usage.
Another proprietary feature the AI Assistant offers is personalised recommendations through ‘complex feature mapping’. It basically understands the naturally used terms while specifying requirements, and converts them into catalog filters to showcase the most relevant products. Imagine a woman looking to buy a foundation for her makeup needs. She activates the AI Assistant and simply says “I am looking for a matt finish foundation that should be wearable for more than 8 hours and have deep sun protection”. This triggers the Assistant to decode these preferences into industry terms like “SPF value” and “stay-in-place” to filter out relevant products that meet the criteria.
Thus, the AI Assistant is capable of decoding such complex preferences and applying relevant filters automatically, saving the user’s time and effort in having to develop a comprehension of all the terms manually. Moreover, the resulting personalised recommendations prove to be more effective in improving session to transaction conversion rates.
The next key feature of the AI Assistant is its ability to generate deeper product insights that can help a user gain answers to very specific questions regarding a particular product, or perform a feature-wise comparison of multiple products. The Assistant aptly answers such queries by scanning through available product specs and reviews. For example, a customer looking to buy an AC might want to know whether the product is noisy or silent in its operation. In this case, the shopping assistant would look for decibel levels in specs and sift through reviews containing terms such as noise or noisy, silent or silence, and other such combinations, verb forms, synonyms and antonyms to provide a clear answer to the user. These answers are designed to sound like coming from an actual person. In our case of an AC, it could throw up something like “I totally get you! No ‘grrring’ sounds is a must. This one is good at cooling silently, though it might make a slight amount of white noise.”
It can even display ratings on such parameters and most relevant reviews addressing the user’s question. Further, users can compare products in a list of results and drill even deeper by asking the assistant to compare two products on a specific parameter — like when the user asks — which one produces lesser noise. These capabilities enable the Assistant to provide a uniquely delightful experience for online shoppers.
One more way in which Buzzo assists e-commerce players as well as their customers is through its contextual cross-selling feature. Let’s refer to our yoga pants example once again to understand how this works. Assume that a customer searches for and adds to her cart a pair of yoga pants. Post that, that assistant suggests her some ‘frequently bought together item categories’ based on previous data, like sports shoes with yoga pants. To this, the user can simply respond with something like “show me shoes that go with my pants”, and the assistant gets into action by using already developed knowledge models to find relevant products basis the current product attributes. For example, here the assistant find her shoes of a specific colour & pattern, basis the attributes of the yoga pants the customer has selected like the striped pattern and red and black color (refer GIF below).
Similarly, in case of online food ordering, this cross-selling feature suggests the right add-ons at the right time — based on what customers generally order together and combinations of items that have been reviewed positively.
The cross-selling feature essentially helps unlock the potential for growth in average order value per user, driving higher revenues for e-commerce players.
In conclusion, all these capabilities essentially make Buzzo a one-of-its-kind voice-based AI assistant that best replicates the in-store expert that customers look for to streamline their search and selection process. It not only feels human, but also talks and helps the customers like one. With these capabilities, it can delight customers and thus, enhance conversion rates, improve sales figures, and help acquire new users as well. Essentially, this form of effective and efficient voice-based commerce is poised to transform the online shopping experience globally.
To know more about the potential business impact, read here on how Buzzo’s AI Assistant significantly increased the cart addition rate of one of India’s largest shopping apps within a span of 3 months.
If you are looking to leverage voice AI to enhance the shopping experience on your app/website, we would love to hear from you at email@example.com