How should organisations prepare for implementing AI?

AI has the potential to change the world. That’s all very grand, but how does that come about in practice in your organisation?

Does it arrive as a single big bang? I believe it is likely to be more nuanced than this. New technology always brings new threats and new opportunities. The question is: how do you respond to them?

You can bury your head in the sand, a path history has not been kind to even when you own the market (see Sony Music and Spotify or Blockbuster and Netflix). A better path is to examine new tools and techniques and implement them in a way where you don’t “bet the farm” on an outcome but apply time-tested techniques when applying any change.

Start small in a non-critical environment, test it to see if it brings benefit, run it a while to ratify and demonstrate those benefits, then begin the iterative process of rolling this out in ever larger pilots – increasing in size, complexity and importance until it becomes mainstream. This obviously depends on your appetite for risk and the potential benefits you can change the step sizes between the complexity of each iteration.

We believe this is the low-risk, high-value approach to AI delivery, augmenting your teams and the work they do by automating much of the mundane, repeatable tasks and processes they undertake where appropriate in the right way is the key to unlocking the potential of AI in your organisation.

So, how do you do this? Where do you start?

The need for AI talent

You need people who are prepared to get behind the potential of AI and not just dismiss it as nonsense, but also people who are pragmatic about what’s achievable and look beyond the hype.

You need individuals who know enough about how the engine works under the bonnet so they know snake oil when they see it (tools overpromising, badged and marketed as using AI but not in practice). Most importantly you need individuals undertaking pilot projects who are mindful of the safety of AI – knowing the restrictions of its use and how to keep your IP and data safe, and not releasing it into the internet so it becomes the training dataset for the next-gen AI tools.

Take our Ten10 Academy training as an example. All Academy Ti10s are trained in the basics of AI, including its background, limitations, applications, and various ethical considerations such as how to use it safely and how to measure the value it creates. Most importantly, in this ever-evolving environment, they are trained to look to the horizon and to evaluate, test, and demonstrate the value of new AI technologies, tools and models in an effective yet safe manner, enabling you to take advantage of AI now and in the future.

We have three levels to our training:

• AI Aware: Everyone across our organisation has at least foundational knowledge of how AI technology works and what work it can complete. Knowing this means they can spot instances where results or images have been generated (via deepfakes or voice replication) and know when a tool/platform is over-promising its AI results or capabilities.
• AI User: All Academy Ti10s are trained to this level, meaning they know how to safely evaluate AI tools, how to pilot them, how to test them safely without exposing customer information to the internet, and how to use AI as a means of increasing productivity.
• AI Creator: Ti10s who complete our Artificial Intelligence specialist pathway attain this level, meaning they can work with Machine Learning within existing ML/Data teams. They are aware of which tools to use and approaches to take, building testing and running models to produce insights. There are two distinct areas: tactical AI (building the answers) and strategic AI (using their engineering skills to build, deploy and maintain models in production).

All members of your organisation should be at ‘AI Aware’ level so they are aware of what AI can do and how it should be used. When you know the business process you are going to utilise AI with, consider who should be ‘AI Creators’ who build the technology you need and who will be the ‘AI Users’ who use the software every day and can provide feedback on its improvement.