The problem with AI is that it’s devilishly logical where emotions need to be engaged. To turn an unhappy customer into a satisfied one, you first need to calm them down, and then sell them something 😉 But for that, AI needs to be taught human emotions. And a good starting point could be something simple – like this empathetic and emotional customer support chatbot.
The Essence of the Project
Siena AI has developed an AI chatbot for customer service in online stores. The unique feature of this chatbot is that, according to its creators, it possesses empathy.
Formally, empathy is the ability to sympathize and empathize. However, in this case, it refers to the chatbot’s attempt to engage in an emotional dialogue with the interlocutor, closely resembling the style of communication of a living person – not just a machine operating exclusively with logical concepts and arguments.
Online store owners can set the chatbot’s communication style in a special interface, choosing the desired style from a menu – and seeing examples of the same phrases expressed in different styles. Eventually, the chatbot will likely learn to adjust its communication style to match that of its interlocutor.
The chatbot can answer customers’ questions about products offered in the store – suggesting similar items, explaining their maintenance, and providing other additional information.
It can even be connected to the store’s social media page, where it will respond to questions and comments from page visitors.
Interestingly, the chatbot can be configured to answer questions on the website in a more formal and professional style, while in social networks, it can adopt a “friendly and cute” style, complete with lots of emojis 😉
The chatbot can be connected not only to websites and social networks but also to SMS and WhatsApp.
With the proper integration, the chatbot can not only answer customer questions but also perform necessary actions – for example, changing the address and other delivery parameters.
The chatbot can also perform other popular requests – exchanging purchased goods for others, accepting returns and refunding the money, changing subscription delivery schedules, pausing or completely canceling subscriptions.
The startup claims that their chatbot can independently resolve over 80% of the requests that customers make to online stores, simultaneously reducing the time to receive the first response by 98%, and the time to resolve the customer’s problem by 90%.
Customers remain so satisfied with the quality of problem resolution and the chatbot’s communication style that the average customer service satisfaction rating for the startup’s client stores is 4.81 out of 5.
The first version of the chatbot was launched six months ago, and the startup promises to release a fully autonomous version in the second quarter of the next year.
Over these six months, even the first version of the Siena AI chatbot managed to attract 65 clients, and the startup has now raised its first investment of $4.7 million.
What’s Interesting
Interestingly, the founders of today’s startup previously worked on another project – Cartloop. That startup attempted to bring back visitors to online stores to purchase items they had left in their abandoned carts, for which live people engaged them in dialogues via SMS.
The main reason Cartloop employed live people was precisely because of the emotions they could express when communicating with clients – and this was what increased efficiency compared to standard letters and reminder messages about abandoned carts.
In 2022, the founders decided to shift their focus to the broader problem of online stores communicating with their clients on various issues, not just abandoned carts.
Importantly, they continue to believe that emotions are a key part of customer communication. However, by that time, they realized that AI could also be taught to express the necessary emotions. This led to the idea of today’s Siena AI.
A couple of weeks ago, I wrote about the startup Connectly, which launched an AI chatbot named Sofia, trying to bring back visitors to complete purchases of abandoned carts in online stores. However, Sofia, as usual, focuses on logic – offering individual discounts or replacements for abandoned goods that are more suitable in price.
So, Siena AI’s focus on emotions is something new 😉 Although very natural – because most decisions and relationships of people are based not on logic but on emotions. And when it comes to customer service – even more so.
The founders of Siena AI’s desire to switch to the broader theme of automating general customer service is quite understandable. This topic has begun to grow rapidly due to the ongoing refinement of AI technologies.
Currently, companies automate only 1.6% of their interactions with customers – but by 2026, at least 10% of interactions will be automated. This is a very good rate of natural growth and a very appropriate time to enter this market.
The startup Octo as an example, I described three key properties that modern AI chatbots must have to be able to automatically handle at least 80% of customer requests:
- Ability to not only answer customer questions but also take independent actions to solve their problems. For example, changing delivery time, canceling subscriptions, refunding money, etc.
- Learning from past successful customer interactions to give more accurate answers to questions and use correct problem-solving methods.
- Ability to independently initiate communication with company employees and contractors who can help solve the customer’s problem. For example, independently contacting employees (or a chatbot) of the delivery service to find out the reason for the delay in delivering an order to a customer. It seems that this list should include a fourth property – the ability to express emotions, which can smooth over a customer’s negative reaction to a problem and/or prompt them to make an action leading to a new purchase.
Another interesting point is that as part of promoting their platform, the startup plans to open the “Siena AI Academy,” where it intends to teach online store owners how to create and use AI in communications with clients. One of the academy’s courses will be dedicated specifically to “emotional intelligence in AI” – a quality that online store owners should start considering essential in “proper” chatbots after training 😉
Where to Run
The main general direction of movement is the automation of companies’ interactions with potential consumers and clients, i.e., creating platforms for dialog marketing and customer service.
This task has always been important because companies spend a lot of money on such interactions. B2B services spend an average of 10% of their revenue on customer support. It feels like the percentage for B2C companies might even be higher.
But only now can companies start spending less money on this – by beginning to automate these interactions with AI technologies. So such automation has suddenly become a very relevant direction. However, as a result, many startups have rushed into this area, and one needs to stand out against their background to attract the attention of potential clients and investors.
Focusing on the emotional component of communication is a cool option that few startups are actively promoting right now. So this can be used while it hasn’t yet turned into one of those functionalities that are usually listed in a comma-separated list in the general product properties 😉
If generalized, the discussion might not only be about creating AI chatbots capable of expressing emotions but about AI chatbots with a distinctly individual personality. I’ve already expressed this idea in one of the recent reviews, suggesting creating digital employees not only with certain competencies but also with a certain character – like “aggressive salesperson” or “emotional support staff member”. An emotional support staff member, as we see, has even been created 😉
About the Company
Siena AI
Website: siena.cx
Last Round: $4.7M, November 16, 2023
Total Investments: $4.7M, Rounds: 1