Don’t blame the bot! AI and the human factor


When I talk about marketing with my customers and at events, I discuss how analytics, including using artificial intelligence ( AI ) and machine learning, can drive improved customer experience. I often say that without analytics, marketing departments might as well just go home.

Technology, data and insights

In today’s world, we are surrounded by data. Companies are trying to collect the right data and transform it into meaningful insights using powerful algorithms. But despite increasing use of technology and increasingly demanding consumers, the basic rules of customer engagement have not changed. In marketing, the equation of “relevance + convenience = acceptance” is still valid. The only difference is that over time, the precise definitions of relevance and convenience have expanded to recognise the importance of tighter targeting and the involvement of real-time technology and more channels.

The basic rules of customer engagement have not changed. In marketing, the equation of “relevance + convenience = acceptance” is still valid.

Most companies have a range of technologies they use to (try to) deliver a consistent customer experience to their clients and prospects. Organisations are doing everything possible to understand their customers, who are their most valuable assets. They are trying to learn about them and predict their behaviour using real-time analytics with high-performing models. The technology allows organisations to use the data, discover the best insights and deploy these insights in daily operations to drive optimal customer experiences.

But is this really true? What about the human factor? I think this could turn out to be even more important.

Technology or talk?

Let me tell you a story to illustrate this.

When my son was born, almost 19 years ago, I opened a savings account for him at my bank and put a fixed amount into it every month to build a fund for him. When he was about 6 years old, he opened his own children’s savings account at the same bank – the lure of the piggy bank, with PIN code security and digital money-counter was too cool to resist. My son used the savings account for a while, but when he turned 12, it became a traditional current account. He never used it as such, because his mother opened another account for him at her bank, but it remained there, dormant.

The day my son turned 18 years old, the money in the savings account that I had opened for him was transferred to that dormant current account in his name, and the savings account was closed. I only discovered that when my monthly deposit to the savings account was refused, and I called the bank to ask what had happened. They suggested that my son should contact the bank because I was no longer able to handle the account on his behalf.

This is where the story gets complicated, and indeed, quite astonishing. The adviser at the bank told my son to go online and apply for a debit card, which was free for students. He could then use the debit card number to create an online banking profile, transfer the money to the account of his choice and then close the long-dormant current account.

Don’t blame the bot! AI and the human factor

This young man needs to decide what to do with his long-term savings account. He could be your next loyal customer, depending on the advise he gets.

Yes, really. A long-term customer (albeit not active), son of another long-term customer, and that was the advice: how to close your account and go elsewhere. Not only a ridiculously drawn-out process for something that should have been very simple, but a huge missed opportunity!

The bank adviser could have asked my son whether he needed to access the funds immediately, or if he wanted to open a student’s savings account or short-term deposit. They could have advised him on how to invest the money for the future – it was a substantial amount by this stage, after all. This could have been an opportunity for the bank to build a relationship and grow a customer with a potential good and predictable lifetime value. But instead he was advised to close the account, so he did.

The future could be bots

Two weeks later, I had a business meeting with the team responsible for customer engagement at that very bank. They showed me all the new initiatives they were launching and how they were using analytics technology (provided by the company I work for) to understand customers and predict their behaviour, including using the latest AI technology to drive smart conversations with customers through digital assistants and chatbots. The contrast with my son’s experience struck me forcibly.

Don’t blame the bot! AI and the human factor

A bot would have significantly outperformed a human bank adviser – at least as far as the bank was concerned.

A bot, it seemed, would have recognised my son, collected all the necessary data from his questions, analysed his behaviour, asked the right questions and advised him on what to do with the money in his current account. It would, therefore, have significantly outperformed the human adviser – at least as far as the bank was concerned. My son might even not have been able to tell the difference between a human and a bot on the phone, although face-to-face would probably be a different matter.

And that is the nub of the matter: Would you be prepared to talk to a robot instead of a person? I think the moral of this story is that the human factor is often forgotten in the process of digitisation. The bot might have done better for the bank, but the adviser answered my son’s question, and helped him to achieve what he wanted (the account closed, and the money elsewhere). A combination, though? The human adviser, supported by the bot that recognised my son? Perhaps a very different story, and a better outcome for both bank and customer.

Organisations forget the human factor at their peril. They need to continue to invest in people, get the right personnel to interact with customers, train them well and teach them how technology can help them to get the best out of every conversation. Bots are not the answer to everything, but neither are humans alone.

Organisations need to continue to invest in people. AI and bots are not the answer to everything, but neither are humans alone. #AI Click To Tweet

About Author

Rene van der Laan

Director, Global Practice - Customer Intelligence

Rene van der Laan celebrated his 20th anniversary in the enterprise software business in 2017. Although he can be considered a software veteran, he is still not done learning, thinking and definitely not done talking. Rene closely follows the ever-evolving world of marketing, customer experience and analytics. In his international role at SAS Institute he evangelises the Customer Intelligence topic - in his presentations he does not shy away from statements like: "without analytics, marketing departments lose their right of existence", "most companies underestimate the value of the customer experience" and he challenges organisations that often have no clue who their customers really are and don't know which direction the organisation should take with their customer journeys. The bold statements are based on the many conversations Rene has with marketing, analytics, digital and customer experience professionals as well as with their respective leadership. SAS' software solutions can help organisations to rise to their challenges, with a joint roll-up-your-sleeves mentality to deliver the ultimate experience their customers deserve - while keeping the organisation's objectives in mind. In his spare time, he also does not just shout from the sidelines, but during the weekends he leads many hockey matches to sporting success or he coaches one of the kids to victory. If he is not doing that, you will probably see him working on his handicap on the golf course.


  1. Mathieu Gaouette on

    Good insight. Never looked at AI as a way of avoiding "dumb" mistakes like the one the Bank rep made when in touch with your son.

  2. Christian Lamarche on

    I'm always amazed to see how many people think AI and bots will be the panacea to every short comings a company may have. If the company cannot teach their rep how to correctly handle a customer, I'm not convinced they will be able to train a bot.

    • Rene van der Laan
      Rene van der Laan on

      Training is key - training for employees, but also training (with data) for the bots. If the training (the teacher for the employees and the data for the bot) is not of the right quantity and quality, the trainees cannot be expected to excel in their jobs.

  3. AI and ML are nowadays very prominent in every facet of life, but not much attention has been given to how to enhance quality of the needed data right from the source – structure, type, format, etc. of the data to be collected. These are very vital in determining the accuracy of action and reaction with regards to AI performance.

    • Rene van der Laan
      Rene van der Laan on

      That's where human control is still required - especially with chatbots you have to define the task for the bot, what's its purpose, then train it with the correct data, train it again and train it even more, and make sure to monitor progress and quality of service the bot is providing.

  4. This is revealing. The bank could have definitely done better with some AI assistance together with the human touch.

  5. Yes, but I am worried about the fact that ML utilities are only as good as the data they're trained on. Depending how many of a particular case have come up, I can easily see the "correct" conclusion being missed. And once a human learns to rely on the ML utility, they may stop digging to establish that the utility made a mistake.

    • Rene van der Laan
      Rene van der Laan on

      I agree, once a bot gets lost or is not sure of a given solution, it should be trained to come up with a suggestion for human intervention. Also for bots there an 80/20 rule applies. 20% of the conversations still may require human intervention, especially with delicate subjects. But again, over time the bot will guide the customers in the right direction.

  6. Matthew Landals on

    Indeed we need both artificial intelligence and human intervention. On the one hand, computers don't make mistakes and can perform calculations a lot faster than humans but only humans can adapt to situations not foreseen when artificial intelligence was developed.

  7. It has to be hybrid model until bots (or AI) can completely take over. Technology has to be an enabler first then an alternative.

  8. I went to a talk last night about AI and Healthcare, with very much the same premise. I think its safe to say that the use of AI to support the human brain and not replace it is where we are today.

  9. Jean-François Ducré-Robitaille on

    What if you son had been a bot talking to the bank bot... When bots will be making more and more decisions between themselves, hiw wre we going to react and what will we do?

  10. I just changed cell phone companies. I made my initial decision and switched one of my phones. Then switch the second one based on price. Each cell phone company in Canada has similar prices and I am talking saving 20$ a month over a two year contract. There was something else behind my switching too.
    I like to call phone representatives at all hours of the day. The new company has longer hours. Then the story gets interesting the phone reps at the new company are never rude or in hurry to hang up or send me elsewhere like your human banker did.
    The old phone company did that sort of behaviour telling me they did not have what I needed and instead to go to the brick and mortar store. The new company it was mentioned in the news is using customer intelligence software. The reps even know the company did this to improve their approach to customer relations.
    The old company is trying to gain a monopoly by lobbying the government regulators to make a less free and diverse Internet.

    I actually don't work in a field/business with customers and we do not sell anything. It is interesting to see this marketing but it is not something I know very well.

    • Rene van der Laan
      Rene van der Laan on

      Thanks for sharing your experience! And I fully agree that the use of Customer Intelligence technology helps drive a better Customer Experience.

  11. Bruce Densmore on

    Sometimes we are too smart for our own good. We need to continue to remember the basics ... and what it is that we are REALLY trying to do?

  12. Marc-André Thivierge on

    Interesting to point out that we need to keep our mindon both technologies and human interactions. Often we seem to focus more on one or another. The challenge is to get the right balance to maximise the benefits.

  13. Donald Wildeboer on

    The bot might have even been able to notify you or your son when the first account was closed and money transferred as an opportunity for positive customer engagement.

Leave A Reply

Back to Top