Chatbots are the future they said, but in reality they never quite met the industry expectations for some reason. Failed experiments like Facebook’s highly-anticipated virtual assistant “M” only seemed to further establish that the industry was a fad rather than a trend.

But was it fair to say that chatbots were dead?

In reality, there were numerous reasons for the slow take up of chatbots in the industry. Technology hadn’t caught up to expectations and the implications and issues surrounding AI were just being explored, are but a few speed bumps that derailed the fast growing chatbot wave.

However, as technology took a few aggressive leaps forward, the adoption of chatbots as sales, marketing and engagement tools have also moved forward quickly.

To find out more, we spoke to Matthew Low, Co-Founder and CMO of AiChat, a leading chatbot builder based in Singapore. Established in 2016, this startup has built an impressive resume of clients such as Petron, Mitsubishi Motors and Philips Lighting.

Matthew was kind enough to take some time out of his schedule to share more about AiChat’s journey and the chatbot industry in Southeast Asia.

The Philips lighting chatbot

Chatbots have been part of the consumer journey narrative for some time and were lauded as the next ‘big thing’ a few years ago. Over time, we have seen limited utilisation of that service, but it seems to be growing in popularity once again. Why do you think that is the case?

We started our chat journey in 2015 and witnessed this trend unfold. Back then, major chat players opened up their channels for a bot framework, including Telegram, Facebook Messenger and Line, granting multiple entry points to automating the rising volume of chat. Hence, what seemed like a tech rush as brands jumped on the chatbot bandwagon.

The chat journey is however more than simply setting up some rules and applying it on the channel of choice. It required apt gathering, preparation, conceptualisation and structure for data to work with strategically applied machine learning algorithms. Access points were also crucial to the journey – identifying the most eminent need and necessary data access. With a combination of lapses across these, the customer experience would be significantly jeopardised.

Increasingly now, brands are becoming aware of this and identifying that the chat journey is a process, instead of a quick tech fix. More importantly, it’s the recognition that chats are not meant to replicate preceding tech implementations but to create a system to scale user analytics, visualisation and ultimately personalisation. With that, we’ve seen our enterprise clients roll out chat experiences that are market sensitive and progressive with each user interaction.  

Do you see this as being a disruptor in some industries, maybe impacting sales or customer service to make them less relevant for some brands?

We’ve definitely seen an impact on our clients’ customer experience, lead generation and sales standards. This ranges from helping brands automate 91% of enquiries, generating 5 times more leads to driving 3 times more lead conversions. We are confident this will only get better as we grow with these brands in scope of chat, product range and regional markets.

That said, the disruption does not come in the form of replacement but that of empowerment. Chatbots are meant to empower service or sales agents with its ability to execute data communication 24/7 and at scale. However, the human touch and intuition in such communications can never be fully replaced. In fact, we envision that with the collaborative application of chatbots and human agents, new job roles and opportunities will arise such as Conversational Designers, A.I Chatbot Trainer, Linguists etc.

What has been the most significant technological leap for the industry in the last five years?

The emphasis on big data management, strategies around scaling customer experience and chatbot accuracy have been some of the most significant shifts we’ve seen.

Chatbots enhance the customer experience. They are now able to detect the emotions and sentiments of the customer they are speaking to; they can respond with a personality that represents the voice of the brand; and the chatbots can escalate an issue to human agents when necessary. With chatbots, it is now also possible to tailor-make the conversational experience for the customer. This advancement in personalization is possible as chatbots can profile and segment customers, to provide them with more personalized experience in terms of the content to be consumed, and how products are offered to customers. 

Chatbots are also smarter: they can understand the context of the inquiries clients make, and can thus provide more accurate results. This is made possible through: higher computing power, which allows us to train the A.I. model faster; data availability, through which there are more sources to collect data, allowing us to build up domain specific knowledge bases and native language packages; and better tools than before to build chatbots, with better interface for business users to build a chatbot without any programming skills, hence speeding up the deployment and roll-out of chatbots.

Can you elaborate on how chatbots improve conversion rate and lead to increased sales? What are some of the factors that lead to a better sales conversion rate?

Similar to an actual salesperson, effective chatbots pace the ‘relationship’ through the customer journey, at scale – almost like how that same salesperson reminds what you came into the shop for, your preferences and eventually how to keep you buying. 

The most basic step would be to qualify user needs and segment them accordingly. From there, machine learning kicks in to form a 360-degree view of you as a customer that evolves and adapts with their interaction. Based on these insights, chatbots help to make customers feel incentivised to convert through the personalisation of content, curated recommendations and targeted promotions/deals. Furthermore, with chat, previously masked elements like intention and sentiments help to make lead scoring more detailed. Analysing these structured data and constantly testing and updating the predictions through the bot’s content eventually leads to clearer visualisation of segment groups and buying patterns which then significantly propel conversion rates.

Is there a risk of the uncanny valley issue when it comes to improving the nuances of conversational bots?

It’s a common dilemma – balancing the humanizing of the bot while managing user expectations.

With our clients predominantly using chat to engage and acquire leads, the persona and ‘voice’ of the bot often becomes a centrepiece. Hence, we’ve built bots that play the role of ladies’ buddies for contraceptive advice or a Japanese car sales agent. Free-form conversations are also key to ensuring the conversations do not revolve around static flows and clicking of buttons.

For AiChat, alongside these experiential elements, the core of human-likeness starts from the functions and conveniences a human agent could bring to the user, but now with higher immediacy, connectivity, available 24/7 and within a few chats. This varies across brands but fundamentally the experience should not make the user feel like they’ve been shortchanged when speaking with a virtual agent. Ensuring a proper fallback and handover to human agents for more complex issues is also essential to the conversational journey.

What have been the challenges in getting some of your more traditional brands to use chatbots?

Debunking the myths as mentioned in earlier points as well as aligning the chat solution to legacy models of communication and data collection channels are some of the more prominent issues. While the latter might take a more macro and longer process of digitisation, chatbots sit as a layer on top of existing systems and can be implemented in parallel to aid in the structuring of incoming data in line with digitalisation efforts moving forward.

Another challenge would also be user behavior – especially for more traditional brands where digital efforts would be quite a step forward from existing initiatives. Framing a relevant and, in a way, palatable chat experience in line with the most pertinent needs that were surfaced from analysis of historical and current chat data would be key and some brands might require a few iterations to get it right. 

What do you think is next for the chatbot industry in Southeast Asia?

With the increasing growth of business chat and online shopping in the region, the chatbot industry will only get more crucial to having that edge in such a competitive environment. As with the projects we are working on and incoming ones, we see a management shift to embracing chat as a pivotal part of digital strategies, thus affording it richer data points and deeper integrations – to then move information faster to and from their user base.