For the last couple of years, we’ve been talking about artificial intelligence. And today AI is doing a lot. However, we are exaggerating its abilities. We, marketers, are playing the game of promoting the concepts, not the realities. And today AI has already an amazing promise.
At least we have our silly Siri.
I grew with the fantastic Star Wars and The Jetson Family in which there are robots clever and practical enough to ease our daily lives. I got frightened by the Terminator. I went to Japan at the beginning of the 2K and saw many astonishing technological advances using AI. [Wow they are still pretty far before us. And now China is rising.]
Despite those visionary movies, the huge bulk of money invested in the technology, no-thing happened. Yet…
What we have today is compute power
AI is a moving target for us. We are not the master of technology. It’s evolving day by day.
AI is already amazing today. Once we were introduced to MS Excel, we were awed with its ability to make calculations in seconds. We can create formulas for complex results. Today AI is just like the processor behind Excel spreadsheets. The only difference is the volume of data.
When you have data,
When you analyse it effectively,
When you create a right model,
When you train your datasets and feed it to AI,
AI does a great job in analysing the information out of data.
Today facial recognition is an output of AI. We willingly upload our photos on social media and tag them precisely. Technology companies train their AI with our data. Then we are awed with the tagging suggestion from Facebook, Apple, Google. In fact, we do the work for AI. We need to be awed with our willingness.
So what is the point here? The only point is that we need to rely on DATA, not the tools generating results over that data.
AI is sneaky and greedy
AI is really magic, it does exactly what you tell it to do. But AI is very difficult to train. It has a huge appetite to process. It demands big datasets to give you back the right results. And yet no big dataset is enough; it is only fruitful if you create the right model.
There are many limitations in having a successful AI:
We do not have key datasets.
We do not have the ability to label them correctly.
We do not have the people with the vision to create data models.
We do not have consolidated data across the corporate, rather we have siloed divisions which holds their data ungenerously.
We do not have the executives who understand the real power of the company is the data they have.
But we like the concepts, we like technology, we like the tools and we like to spend money on ICT.
Do not miss the learning curve
Although many articles and news emerged every day, nobody is in a rush in advancing AI. Even your competitors which declare their results in AI have no real ability to catch the hype.
Other than Amazon, Google, and some couple of companies — which are still working on AI in very narrow applications — , no companies master artificial intelligence. So you have still the opportunity to be the early adopter.
As a business executive, first, you need to understand what AI really is. You need to learn its abilities and limitations. Then you need to start thinking about how you apply it into your company to create economic value and competitive advantage. You need to figure out where to start: find the area of what is already automated + data people are already processing.
You need to build an architecture layout and data plan, how you can bring those distributed data from the divisions of your company.
AI is the ability to analyse data and use it more effectively. It’s that simple.
So, data is everything. Your job as an executive is to come up with a vision of how data would be used more effectively. However, I have bad news. It is not that easy, because your data is mostly garbage today.
First, you need to deep dive in your data to find out the most precious and relatable for your business goals. [Please skip the data gathered about your customer’s hobbies and interests. Nobody tells you the truth.] That precious data enables you to have better decision-making. Make a call for gathering data over your siloed divisions and get their involvement by comments, and even arguments.
Then, you need to get some help. No, no, no… not from big consultancy companies. Just choose the right data scientists. The right candidate would be the one who is generous to be “willing to help” the users. Is that person who you have today in your organisation? The one with no ego to prove you you are doing wrong with your intuitions and experiences. If so, you are lucky. The ideal candidate would communicate with you in simple terms, iterate the model by working closely with the users.
And do not rush into getting high-end, complex technologies. Technology will evolve soon, turn to be easy-to-use, plug-and-play.
What if we do it wrong?
We made our civilisations by the generosity of Prometheus, by gifting us the fire and use of metals. Our civilisation will leap by our humanity, in the use of — our new fire — artificial intelligence. We will have failures, and most probably learn the best out of our misery. Let’s keep trying, experimenting, and by listening to our conscience.