Artificial Intelligence | News | Insights | AiThority
[bsfp-cryptocurrency style=”widget-18″ align=”marquee” columns=”6″ coins=”selected” coins-count=”6″ coins-selected=”BTC,ETH,XRP,LTC,EOS,ADA,XLM,NEO,LTC,EOS,XEM,DASH,USDT,BNB,QTUM,XVG,ONT,ZEC,STEEM” currency=”USD” title=”Cryptocurrency Widget” show_title=”0″ icon=”” scheme=”light” bs-show-desktop=”1″ bs-show-tablet=”1″ bs-show-phone=”1″ custom-css-class=”” custom-id=”” css=”.vc_custom_1523079266073{margin-bottom: 0px !important;padding-top: 0px !important;padding-bottom: 0px !important;}”]

Turning Chatbots From a Frustrating Failure to a Flourishing Business Function

There are many aspects of life that drive us all crazy. When it comes to dealing with brands, there is a real front-runner in the frustrations steaks – chatbots. As a staple of online brand interactions they are something we’ve all had experiences with. Mostly poor ones. Their inability to naturally communicate can leave users feeling anywhere from frustrated at best, to wanting no further interaction with the brand at worst. In the majority of cases they are as ineffective as they are prevalent and it’s time something was done about it.

Only Four Percent of Businesses Using Chatbots Efficiently

The advances in artificial intelligence (AI) as well as increasing user acceptance have resulted in chatbots gaining tremendous popularity in recent years. According to Drift’s 2020 State of Conversational Marketing report, usage of chatbots as a brand communication channel has increased by a whopping 92% since 2019. It found 24.9% of buyers used chatbots to communicate with businesses in 2020, up from 13% the year before. It’s a trend only going one way. According to Data Bridge, the chatbot category is expected to grow to an estimated $46bn market globally by 2028, while Juniper Research predicts eCommerce transactions via chatbots will reach $112 billion by 2023.

There is undoubted potential in chat bots but, despite promising predictions for future growth, data tells a different story when it comes to how they’re being utilized today. According to Gartner, though 90% of enterprises have invested in chatbots, only 4% are actually using them efficiently. Research [Janssen et al, ICIS 2021] suggests that 53% of chatbot projects were discontinued after 15 months – an extremely high failure rate.  The reason for the discrepancy between expectation and reality is as simple as it is harsh – chatbots have been built to fail.

Recommended: AiThority Interview with Sam Yoshida, EVP & GM of Canon USA

Built to Fail 

First things first, any new or evolving technology is subject to the rigours of a discovery process – something not specific to chatbots. There are many technology projects that don’t work on account of anything from; lack of resources, internal support, financing and other issues. But the reason for the universal failure of chatbots runs much deeper.

The main one is that they have simply not been built to support natural conversation flows. Their rigid design means that, when the chatbot loses track, can’t understand or needs the ability to go ‘off script’, it can’t, taking the consumer on a very frustrating trip as a result.

Virtual Assistants like Alexa or Siri are heralded as pioneers but, in reality, they are the most ubiquitous examples of chatbot failure and acutely emphasize this point because neither converse with the user. They simply provide a response. For true virtual assistants to have any chance of success, they must be designed to reflect how humans speak. This means things like; answering multiple, unrelated questions, contextualizing responses to the situation and the user and the ability to seamlessly pick-up a previous thread.

Read More AIThority Insights: Role of AI in Helping B2B companies that are Missing Out on Buyer Intent Data

Another major source of failure is that chatbots have not been built to scale. Because the technology has been constantly evolving, updates and iterations tend to be made piecemeal and at the code level.

Related Posts
1 of 860

Teams build a chatbot, realize its limitations, write custom fixes through  code to improve the experience and repeat. It’s not that the people designing these chatbots don’t know what a good experience is, it’s just something that is incredibly difficult to create. Often improvements or fixes are made for a very specific use case so when the situation changes again a short while later, the new update becomes old – tailoring chatbots for seasonal peaks and troughs is a classic example. The upshot being that the business is left with a Frankenstein-esque chatbot, that can’t be easily rebuilt, with a CEO reluctant to sign off more investment to right the wrongs and repeat the cycle. Most just leave it as it is, which is why poor chatbots remain such a common occurrence.

 

The third and final reason for failure, especially in large enterprises, is that a chatbot project involves many people. It includes individuals from; marketing, branding, legal, technology and others. All with a slightly different perspective of what it should and shouldn’t look and feel like. Because it cannot move forward without each party being happy, the process tends to be long which exposes the chatbot to changes in technology and leads to the aforementioned scenario. While this is a problem that is common to a lot of enterprise projects it is especially relevant for chatbots because they are a new technology and there is no shared understanding of what can be done, what should be done and, crucially, how it should be done.

For chatbots to realize anything like their potential value and become true virtual assistants, something needs to change.

Creating the True Virtual Assistants for Businesses

Einstein once said, the definition of insanity is doing the same thing and expecting different results. If we are to derive true value from conversational AI we need to completely change our approach. In order to elevate them from awkward chatbots to useful virtual assistants we need to rethink how they are designed, developed, deployed and maintained. This is not just about throwing more technology at the problem or more sophisticated AI. It is about re-framing how we approach the problem from the ground up and being smarter and more efficient about how we exploit the technologies available to us.

That is exactly how we approached the challenge.

Starting from academic models of conversation that capture the fundamental principles of how humans have conversations in social settings, specifically, conversational context, and combining that with existing conversational AI technology for NLP. The rethinking of how to model and reason about conversational context computationally and can move between different threads of a conversation seamlessly gives our virtual assistants superpowers when compared to chatbots that are restricted to rigid flows. The results are virtual assistants that feel much more natural to converse with and can deliver the delightful experiences we want our users to have with much less effort and in a way that is far more scalable.

Next was ensuring that this capability was easily accessible to enterprises and answered the needs of different stakeholders at scale. That is why access to this capability is made simple through a SaaS platform and no-code environment designed to cater for the entire team mitigating the issues of shared understanding.

Don’t Give Up Just Yet

Virtual assistants today can deliver outcomes faster and lower cost and with a far better user experience that chatbots have so far managed. And our message is simple – don’t give up on them just yet. After all, chatbots so far have been built to fail.

Through those failures we’ve learned the value they can bring and the process of discovery has uncovered what was being done wrong. We now have an opportunity to get things right by creating genuine virtual assistants capable of redefining the way machines interact with humans.

[To share your insights with us, please write to sghosh@martechseries.com]

Comments are closed.