In our daily lives, interactions with machines have become more pervasive than we sometimes come to realise. Chatting with a BOT disguised as a friendly customer service agent to get information or help to complete simple tasks is commonplace.
This week for instance while considering options for a family holiday and unable to zero in on a location, I turned to the messenger option of a popular travel website. Maybe I had overestimated the capability as the chat fell short on my expectations. With no ill-intention, I gave somewhat vague and open-ended answers in response to where I wanted to go. While I didn’t have a preferred location in mind, I asked for ideal spots to visit during the summer season in Cornwall. In return, my personal adviser aka BOT provided information that was incoherent and unhelpful. Programmed to provide information to only structured conversations, the BOT failed to pick up the anomalies and the context of this particular chat, sub optimally using its natural language processing power. As a result, I was left with no other option but to continue with my search through the endless sea of information to find the perfect holiday spot for the summer.
Was I wrong to expect the messenger to give me useful advice to help with at least narrowing down the options? Should the BOT have better used its processing power to work out customised responses? Or perhaps handed over the query to a human when unable to decipher the request itself?
The answer depends on what the BOT was set-up to do and its intelligence quotient. While advancements in AI and in particular conversational AI have led to a higher adoption rate and chatbots are commonly deployed in customer services, its capability has remained somewhat underdeveloped. An important factor that has contributed to this is a shortage of skilled AI and data science practitioners. As, ironically, fully functioning and intelligent machines can be built only by intelligent and trained human minds.
It is estimated that the demand for AI professionals in the UK tripled in 2017. From an article published in The Innovation Enterprise, ‘According to the job search website Indeed.com, June 2015 to June 2017 saw a 500% rise in the number of job postings in the field of AI. Of these job postings, 61% were for machine learning engineers, 10% were for data scientists and just 3% were for software developers. And according to a survey by Tech Pro Research, just 28% of companies have some AI experience, and more than 40% organisations have enterprise IT personnel with no AI/ML skills for implementation or support.’
While a debate is still ongoing about the impact of AI on human jobs, a more pressing challenge in front of the industry presently is to upskill employees and train younger professionals to fully harness the power of machine intelligence. In this context, budding entrepreneurs have made it their business to develop a niche and AI tech-startups are blooming globally to tap into this opportunity. According to Gartner*, by the end of 2019, AI start-ups are set to overtake big tech firms namely Microsoft, Google, and IBM with disruptive business solutions.
So, for organisations looking to build digital tools hinged on machine intelligence, exploring niche outfits to tap into AI talent is a sensible first step. While it can take months to build an in-house specialist team, some start-ups can help bridge the skills gap quickly by tapping into their carefully screened database of AI professionals. AtomX Digital, for instance, offers a unique end-to-end solution for resourcing AI experts (MSc / PhDs with specific industry backgrounds) in a short timeframe (10-12 days) from understanding project requirements to onboarding fully ramped up resources.
* Gartner Predicts 2017: Artificial Intelligence