The Chatbot market has reached an unbelievable state. It is both over-hyped and under-hyped at the same time.

There is no doubt that conversation-as-a-service will become the primary means for humans and machines to interact. The changes will eventually be profound impacting everything.

As with any monumental change, the early days are wild as the beginnings of momentum gather pace. Understanding the underlying dynamics determines success or failure.  

We are well versed with 'leading' analysts using methods to explain these types of situations. For example, models such as the ‘hype-cycle’, containing the 'peak of hype' followed by a ‘period of disillusionment’, followed by 'sustainable and transformational growth' are well established. But what are the hidden behavioural economics underpinning these models?  

The other day, someone said that they were finding several executives from tier 1 financial firms saying they are only interested in Chatbot Artificial Intelligence, if it can automate regulatory compliance, with minimum effort from people. For some reason they thought that Natural Language Processing was able to achieve this. To make matters worse, any alternative could not be seriously considered at this stage. It was explained, they were already sponsoring several projects of this type and no longer had the bandwidth to consider alternative options. In other words, anything that puts the spotlight on some of the existing initiatives being flawed were not going to be considered until “further learnings” had been gained.    

Sounds like some people believe Natural Language Processing has become a Singularity – all knowing, god like. Unbelievable. This distorted belief is of biblical proportions, enough to want to break some tablets!

It is now become clearer that over-hype needs two parties to be involved – the buyer and the seller. In these type of cases, the profile of the buyer needs to be a combination of arrogant and ignorant with deep pockets. The seller’s profile is to adapt the perceived capability to meet the needs of the buyer. Once several high-profile clients back proof of concepts then the hype moment gathers traction. This attracts the media and the spin stories follow. The analysts jump on board as the high-profile clients are their clients also. As further momentum gathers then investors jump in.

First round seed capital in the past year has enabled some Chatbot firms to raise between $1.5m to $10m.To raise such money, which requires no capex, is very high for a first round of seed capital. This is a sign that the over-hype is gaining a momentum.   

Once evidence starts to emerge to show that some of the original claims were ‘unscientific’, typically the smarter original sponsors have already switch roles or jobs, naturally for higher salaries as their newly acquired ‘knowledge’ is in demand.    

In summary, hype needs both buyers and sellers to gain the momentum. As we reach the end of 2017, the Chatbot over-hype continues to grow, especially in sectors with deep pockets or where firms are attracting grants and investments. There are genuine Chatbot initiatives happening in those sectors with deeper pockets, but even these firms are frustrated with the levels of misinformation. 

Other sectors with shallow pockets to invest tend to be far more cautious in their engagement with Chatbots. In particularly, the public sector, which is pioneering Chatbots is under immense multiple pressures of greater regulatory complexity, increasing costs, limited budgets, greater levels of transparency whilst struggling with improving productivity, which is a national problem. It is the type of sector profile where some of the Chatbot activities are potentially game changing. In these type of market profiles, the learnings are often being under-hyped.

Maybe the Chatbot market is more like the hare (over-hype) and tortoise (under-hype) race.