top of page
  • Writer's pictureLaurence Rau | CEO & Co-founder of Emerge ML

If you knew what was going to happen, what would you do differently?


Hands forming a frame over a sunset

Recently, after 74 years in business, Comair announced that it was filing for business rescue. For those, that don’t know, Comair is a regional South African airline that operates the British Airways South African legs and the low cost airline. I first became aware of this company early in my career when I worked for Discovery Vitality (affectionately known as just Vitality, which Discovery has more recently used for Discovery’s UK Life and Health insurance offering).

In 2006, Vitality launched a partnership with Comair to offer discounted flights for South Africa’s Vitality members based on their Vitality status. This was not the first time the two companies had partnered - to that point a similar offering was in the market for the British Airways brand. But this new offering was really exciting – the brand was just 5 years old at the time, but it had made a spectacular splash in the local airline market with funny adverts and a quirky onboard experience. was the cool, new brand.

I was tasked on Vitality’s side with designing and implementing the partnership. The deal was signed. The customer offering was compelling. Now we just needed to execute.

I remember the first meeting we had with our counterparts at kulula with fondness. We started discussing the various business process decisions and very quickly most things were decided. Just one decision remained – where should we direct the Vitality members to make their booking. There were two compelling options:

  • Discovery Vitality already had a travel booking engine on its website. Its members knew to go there from previous experience. Why change something that was working?

  • But was quickly becoming the first place to check for local flights. Should Vitality members not be expected to follow the trend?

After a lot of debate, one quiet voice suggested, “Why not do both? Let those that would naturally go to Vitality’s website book from there and those that expect to get services at, go there.” It seemed like a crazy idea at the time. Isn’t that double work? Does the cost make sense? Can’t we guide the members to whichever process we choose to use? But we did it anyway – we offered the process from both entry points. And the results were amazing – there were 40,000 bookings made in the first month after go-live and we received just 40 customer complaints. I believe that decision was a watershed moment for the subsequent success of the Vitality-Comair relationship which included various new offerings after that.

I learned two very important things from that project:

  1. Business process is one of the key things in strategy execution

  2. Understanding the way people like to behave is key to any good business process

These lessons have stuck with me ever since.

Today, in my work at Emerge, where we build AI solutions that create exponential value, we always start with the business process. The predictive models we build offer immense potential and when building a new solution, the question we like to start with is, “If you knew what is going to happen, what would you do differently?” Understanding the business process and its relationship to a great AI-based predictive model makes all the difference. 90% accuracy is not enough, without the appropriate business process design.

And effective business process design needs thought about human behaviour. Just achieving efficiency from an AI solution means we are missing a trick. We don’t just want our clients to make the same sales with less cost, we want them to make more sales. We don’t just want our clients to identify which clients are going to cancel / lapse / attrite / terminate (pick the right term for your company or industry) – we need to initiate the right levers or interventions to retain these customers. Good models and even good business process are not enough. You also need a good understanding of human behaviour. Behavioural Economics (a concept I first learned about from Dan Ariely’s book, Predictably Irrational which I highly recommend for those that haven’t yet read it) is something we are increasingly adding into our offering to enhance the value of our AI solutions. To achieve this, we often work with leading experts in this field.

AI models are not enough on their own. They need to work in the context of good business process and more commonly behavioural economics. AI’s success lies in its execution. Accurate predictions, just like good ideas, are not enough. Steve Jobs said it so succinctly, “To me, ideas are worth nothing unless executed. They are just a multiplier. Execution is worth millions.”


This article is article four in a series of six articles created through a collaboration between Crowe and Emerge. The series discusses the significant potential that insights extracted from data can have, as well as ideas to get you started on your innovation journey and key considerations to help achieve maximum impact. Contact us for more information.



Commenting has been turned off.
bottom of page