Chief Customer Rescue Officer The Customer Lifeguard
Share this content
Tags:

Running a chatbot trial: How to test the machines before letting them take over

12th Apr 2019
Chief Customer Rescue Officer The Customer Lifeguard
Share this content
Robot
istock/CSA

If you're launching a chatbot, you may need to win the hearts and minds of both your employees and your customers in order for it to be a success. How can a trial help?

"In the year 5555
Your arms are hanging limp at your side
Your legs got nothing to do
Some machine's doing that for you"     
                             

Taken from the song In the Year 2025, the above lyrics highlight why it's no coincidence that Zager and Evans, who released this song in 1968, were a “one hit wonder”. They warned of the dangers of technology, portraying a future in which the human race was destroyed by its own technological and medical innovations.

Coming so soon after 1967’s “Summer of Love”, it’s likely that the tree hugging, free loving, earth muffins of the time weren’t ready to ditch the flowers in their hair and the floral pungency in the air and hear what real world issues that their follow-up record may have predicted for us.

Earlier, Fritz Lang’s 1927 film Metropolis explored the themes of industrialisation and mass production where, in the futuristic year of 2026 in the city of Metropolis, wealthy industrialists reign from high-rise towers, while underground-dwelling workers toil to operate the underground machines that power the city. I’m hoping that the New York Yankee’s poet laureate, Yogi Berra was right when he said, “The future ain’t what it used to be.”  

While I don’t necessarily believe that the droids are taking over, perhaps we need to heed the prophetic and fiery climax of Metropolis, where Maria, as the android, is burned at the stake as a warning of the potential damage that unfettered robotic intervention can bring.  

In 2019 the daily output on the Twittersphere, LinkedIn blogs, print media and conference keynotes about Artificial Intelligence (AI), Machine Learning, Virtual Assistants and Chatbots named Gladys has reached epidemic proportions. Only 2018’s gold rush-themed, time-bound frenzy of GDPR matched its hold on world media domination. There are many predictions about what this means for the future of work and more broadly, mankind; some of which are probably accurate.  

A study by Narrative Science found that 80% of executives believe artificial intelligence improves worker performance and creates jobs, and I believe that most people want to believe this. 

Can I speak to a bot please?

While there is no shortage of applications for AI, much of the focus has centred on customer service as a prime target for automation. Many surveys suggest that customers would prefer AI interactions, perhaps driven by their experiences with some contact centres, where the level of intelligence fell short of the expectations for normal human contact. This isn’t a knock against contact centre agents as I realise that most really do care. The intelligence gap resides in the companies that tie their agents’ hands, stifle their creativity and drain them of empathy with dumb rules and policies, aged and creaking technology, and other Orwellian themed procedures.

Most of us naturally want to help others and come pre-wired with an attitude and a caring side that is ideally suited to achieving that objective. A character trait noticeably missing in bots.  The best companies enhance their customer engagement by encouraging employees and giving them permission to build on these natural feelings and attitudes and use their emotional intelligence to handle interactions that need that personal, human touch.

The intelligence gap resides in the companies that tie their agents’ hands, stifle their creativity and drain them of empathy with dumb rules and policies, aged and creaking technology, and other Orwellian themed procedures.

This view is shared in an article by Steven Van Belleghem entitled Defining the ‘human touch’ in the customer relationship. While the article focuses primarily on the difference between human and computer-based customer service interactions, Steven’s underlying theme looks at three areas in which humans excel. This is their ability to add empathy, creativity and passion to any interaction. He notes that: “Creativity and innovation are uniquely human characteristics. It is smart for companies to allow human creativity to blossom in all phases of the customer relationship. You must allow all your staff to think creatively about improvements that can benefit the customer.”

However, questions still remain as to how, where and when to deploy this technology. Alan Turing’s famous test was retold in the film The Imitation Game. A machine can be considered intelligent when it becomes impossible to distinguish between human-generated or computer-generated answers. This may still be a challenge for some organisations. But it’s clear that a blend of AI and human capabilities can deliver a powerful end-to-end experience for customers and do it more efficiently and cost effectively.

It’s not just about the money

While much of the rhetoric has to with companies saving money by deploying AI and other devices, a recent study by Forrester suggested that the major drivers were to improve customer satisfaction (51%) and agent satisfaction (46%).

AI is quickly becoming a ubiquitous technology in consumer devices and services, with Alexa and Siri now the first words that many infants learn and speak. What does it mean for the future of the contact centre and how organisations serve customers?

AI-enabled bots and intelligent virtual agents (IVAs) are already transforming post-sales service by enabling better self-service and allowing more forms of customer engagement. But the role of AI is not limited to post-sales interactions. For contact centres, intelligent assistants can significantly improve the customer experience by gathering data, predicting the customer’s needs, learning about the customer’s behaviour and using that to determine who can best respond to any particular request, whether that be via AI or assisted service.

While much of the rhetoric has to with companies saving money by deploying AI and other devices, a recent study by Forrester suggested that the major drivers were to improve customer satisfaction (51%) and agent satisfaction (46%).

As consumers become more comfortable with AI technology, or at least tolerate it, we are seeing more organisations realise that it can be a critical component in improving customer experience by providing faster, better and easier service, while at the same time reducing the company’s cost to serve. But as with any shiny new technology, caution needs to be exercised in terms of the speed and ubiquity of deployment. Otherwise we may risk overwhelming and frustrating the customer in the way that unbridled use of IVR did many years earlier.

In short, identifying those applications that can initially be handled by a virtual assistant or other AI solution, but ensuring that there is a sure fire and swift transition to a real, fully informed and capable live person when needed. Sounds simple, but then so did Brexit. The question on many people’s minds is “where do we start?”

We need to talk about Gladys

Like many other aspects of customer experience and employee engagement, simplicity doesn’t find its way onto the agenda with the frequency it deserves. But for those companies still seeking to find the start line - especially in the case of IVAs and chatbots, initiating a trial based on some basic, commonly understood customer and employee challenges can be a natural and relatively simple entry point.  

I believe that there are three fundamental steps that can point the way and rally the troops, and conversations play an important part in mapping the destination.

  1. Firstly, the conversations that your customers are having, either with you, or without you. The latter happens more than you know. Which of course is the problem. Conversation is one of the richest, largest and cheapest (you already own it) sources of data. Whether it’s found in survey feedback, social and other media, online comments, or even white mail, it generally reflects customers feelings about your organisation and your people and is often charged with positive or negative emotions that have clarity and can be measured and actioned.
     
  2. Conversations with the most important people in your organisations - your front-line people; contact centre agents, shop floor colleagues, field engineers, etc. They know where the bodies are buried and can usually provide a detailed map. A ½ day spent in their company will provide real, honest, accurate data that can inform, inspire and direct your initial AI pilot programs, as well as showing them that their voices matter
     
  3. Cross-functional company conversations that can uncover turf wars, silos, dumb processes and procedures and other organisational misunderstandings that cause untold damage if left unchecked. Agreeing on, and sharing answers, to commonly asked and generally quickly resolved questions by either customers or colleagues can also be an early use case for the AI trial. For many businesses an in-house trial is a safer bet and can pave the way for expansion into the wider world as the kinks are worked out.

The output from these can be collated, crossed-checked and combined into a top 10 (more or less) “things we can use a chatbot/ IVA for” workshop, perhaps with the guidance of an independent AI practitioner who can assist with prioritisation and realisation to finalise the contenders for the trial. Then it’s probably time to engage the solution vendors for a further dose of reality and financial due diligence. As with many modern-day solutions, it’s likely that these suppliers will provide a low cost, no cost Proof of Concept (POC) to demonstrate capability, veracity and value before you pull the rip-chord on your chatbot's parachute.

I firmly believe that these steps can be carried out expeditiously by any organisation serious about making life better for customers and colleagues. Simple? - Yes. Time consuming? - No. Expensive? You’ll have to ask Gladys. But it’s still only the start, albeit a good one, and perhaps this is the time to quote Winston Churchill. “Now this is not the end. It is not even the beginning of the end. But it is, perhaps, the end of the beginning.” And we all know how that turned out.

All in the family

I’m not sure if bringing a bot into your business is quite like “why don’t you take the puppy home for the weekend”, but it’s very possible that once a business starts to test AI in whatever format, Gladys or Granville, will quickly become one of the family. This isn’t a bad thing, in fact it’s probably the best outcome and will (I hope) allow you to demonstrate to colleagues that “she” or “he” is truly one of us and they won’t get the Joan of Arc treatment.

Tags:

Replies (1)

Please login or register to join the discussion.

avatar
By Pawan Pandey
03rd Sep 2019 09:02

Chatbots have become increasingly more important as messaging applications have become the way brands reach consumers. And for marketing and sales purpose, chatbots can offer your audience a personalized experience, gather and analyze customer feedback and data, and make your brand more fun. We at www.engati.com have started the journey.
Do read our collection of blogs at https://blog.engati.com , test our platform and provide us feedback.

ChatBots will be great! They will do what we struggle or do not want to do.

https://blog.engati.com/why-merging-marketing-with-bots-is-the-best-stra...

Experience Engati (https://engati.com/) The Best Chatbot Platform

Thanks (0)