I work in AI, where innovative algorithms and automation are the core of what I do. However, I have tried in my career to work with traditional companies to help them integrate AI into their businesses with varying degrees of success.
My success is predicated on many factors that I have lectured about before; available data, the right use cases, current IT systems and processes, etc etc, but over time I’ve learnt that the number one reason for success is always people, and more specifically the people in charge at that traditional company.
The most common phrase you hear is that you need CEO/ board level buy-in for your AI project to stick. Obviously, that helps but there is something deeper at play here. Maybe I can simplify this by the following equation:
SUCCESS = Trust in the AI way > Trust in the old way
What holds back acceptance in AI at all levels, from a data scientist trying to explain their model to their non technical manager, to getting society as a whole to accept driverless cars, is people’s need to understand what the AI is, what it does, and what the risks are. You can’t expect people to give away control to something they don’t trust.
So when I start working at a new company I need to go through the process of understanding what made the company successful in the past. Eg. If I was working for a SAAS company whose success was based on deploying cloud based solutions to replace legacy office based tools, than my life is going to be a lot easier introducing a new AI algorithm than at a company whose success has been based on 1 to 1 human relationships lubricated by phone calls and corporate lunches.
Most algorithms we have in our toolbox use historic data to predict the future, and humans are no different. For example, political discourse can be broadly categorised in the dichotomy of doubling down on keeping things the same vs trying new things to solve current problems. In the business context however, the companies that don’t use the latest technology to offer better customer solutions usually fail.
Therefore, a lot of my job is increasingly spent first listening and then talking to people about how I feel AI can help them, as opposed to getting stuck in code and tweaking Epilson coefficients.
It feels like what I am saying is not very futuristic, but simply listening, understanding and ultimately respecting a person’s working story is the best foundation for them to listen to you about how you want to help their business grow.
Obviously, things will not always work out this way. A lot of time, some people will try to kill your project, as they feel it might kill their job in the future. Sometimes it’s more subtle, like how a corporate culture respects risk aversion more than leaps of faith when making decisions. At worst, the only path to success is for some people to get isolated and moved out of the way.
But I think intention at the start is important. If we intend to respect the past as well as pushing the future, if we intend to spend that extra time to listen, understand and explain, then you have the opportunity to keep the best of what a traditional company does, and help them evolve into the AI age.
The heart of successful AI solutions should always be human.