“We need to implement AI asap!” - As an executive, this might be something that you’re thinking about or that has been “top of mind” ever since the “ChatGPT” tsunami hit. Or it might be something that you’ve been tasked to “do”. In either case, let me share (very humbly) some tips with you and hopefully enable you to see through the “hype” and take the next step of your “AI Journey”.
Before I start, I highly recommend you explore and educate yourself on the subject, the technology and its powers. Many resources exist like Google Cloud Skills, Geek Gadgets or Udemy (to name a few) and of course, check out ChatGPT, Grizzly.ai, Godmode.space or any other app you have heard about to see what they have to offer and how they can help you at work or in your day to day life. So to be clear, I encourage everyone to explore as much as they see fit but if you are thinking about implementing AI in your business, I suggest you take a more structured approach.
There are major differences between exploring and playing around with AI, and sustainably deploying AI at scale to solve business problems. Some of the differences and key challenges that need need to be dealt with are:
Business leaders overestimate the impact of AI and underestimate its complexity, leading to costly project failures.
Organizations face challenges related to integration, security, and privacy issues when moving AI practices from prototypes to production.
Marketing hyperbole around AI is currently causing a lot of confusion and making it difficult for enterprises to realistically value this source of innovation.
As stated by Erick Brethenoux, VP Analyst at Gartner (1) ;
“Superheated rhetoric surrounding the potential benefits of artificial intelligence is inflating related expectations. Data and analytics leaders must demystify AI, remove technology jargon and focus conversations on real business problems and achievable use cases.“
So how do we see through the hype and overcome these challenges?
Like the “Flight Management System” in an aeroplane, if you don’t input the right parameters and destination - it’s not that useful. Sure, it can keep the aeroplane level and flying at the heading you give it but that’s about it.
So before you go and try to “implement AI” - you need to identify clearly what is the problem you’re trying to solve or what is the opportunity you’re trying to capture - in other words, what is your destination.
Now, you can make a conscious decision to just “try it” and have a few people “play around” to explore and evaluate the power of it - which is totally fine - and in this case, I recommend you make a little project out of it and at the end, you can prepare a report or do a debrief (interactive session) to show the organisation your findings and insights.
But if your intentions are to solve problems or capture opportunities, you need to be very clear and spend quality time on defining the problem(s) and/or opportunity(ies) before trying to solve the problem with AI. I can hear you say; “Yes but if we explore a little bit the technologies and software/apps out there, we’ll be able to pick the one that can solve our problems and we won’t need to spend all of this time. It will be much faster”. My answer to this is; “No you won’t”.
We have only scratched the surface of what AI can do. Every week, new tools become available to solve problems we did not even know existed! Every week someone comes up with a creative way to solve issues we have. So before you jump the gun and select an AI technology, software or App, make sure you define your problem or opportunity as objectively as possible and with as much data, metrics and facts as possible. Once that is done, you should have a better understanding of the problem and in a much better position to evaluate which tool or technology can help you solve that problem. And who knows, by the time you are done defining the problem and requirements, a new AI tool may be available!
Once this is done, and before you go and acquire the new shinny AI tool, the first questions I would ask myself are:
Can one of the current technologies I use help me solve this problem? Has there been development (new module, new functionality) on some of the technologies that I use today that I could now leverage to help me? The idea here is to not introduce another technology if we don’t have to (for many obvious and less obvious reasons).
If not, do I have the internal resources to conduct a “technology watch” or “market research”. If the answer is yes - go for it. If you’re not sure - get some outside help (like us!)
Depending on the complexity of the problem you’re trying to solve, the next step might be a “proof of concept” to validate that indeed there is a path forward or you could jump directly into an selection process.
Conclusion:
Implementing AI at scale in a business requires a structured approach. Making sure we identify precise and business-relevant use cases is critical. Key requirements should be documented and supported by data. We should perform a good survey of the market and perform a PoC in order to showcase the effectiveness of AI and rapidly reap the benefits of its implementation.
References and Interesting reading:
Comments