Four simple principles to help you get started with AI
How to scale AI within your organisation when you have limited skills and resources
Business leaders have been hearing for a while that it’s time to deploy artificial intelligence across the entire organisation. I speak to leaders of all types of business every day who are trying to figure out “what does good look like?” for scaling AI in their organisations – and the urgency is growing along with the need to cut costs.
You probably already know using AI across your organisation will lead to fundamental changes in your business. AI has the potential to change everything we do. In the future, when people look back they’ll compare it with other step changes like the introduction of the telephone, the personal computer and the internet.
Here’s the rub though. It’s easy to point at the rapid trajectory of AI deployment across organisations like Amazon, Google and Facebook. Tech giants can scale AI because they have millions to spend on research and development, and it’s literally part of their job description to be on the bleeding edge of innovation. The way these companies already leverage AI is absolutely something to marvel at, but let’s not kid ourselves that the best way for most businesses to use AI across the organisation is to try to rip and repeat what the tech leaders have done. Most don’t have billions to spend, or the best and brightest minds in technology to make it happen.
Here are four failsafe principles that will help you use AI and machine learning at scale across your organisation.
Principle 1: agree what you want to achieve
Goal setting is a stalwart of resilient businesses for a reason. But even though it seems obvious, many people overlook this crucial task. Machine learning alone isn’t an end goal. Ask yourself what purpose will be served, what outcomes you need to achieve and what the timescales are. This will ensure you stay on track and keep everyone accountable for delivering a positive business transformation.
Principle 2: do your research
To know what is achievable, it helps to know what others have achieved. There is a lot of hype around machine learning in general, so take some time to do a bit of research and understand what other similar businesses have gained from scaling AI within their organisations. This will be invaluable in adopting and scaling ML successfully.
Principle 3: learn to walk before you can run
It doesn’t matter how old your business is, if you are bringing in machine learning where you haven’t had it before and you don’t review your business processes, it will limit what ML can achieve. There are two types of ML you need to know about – supervised and unsupervised ML. Kjell Carlsson has defined these in his January, 2020 report “Shatter the seven myths of Machine Learning”:
- Supervised machine learning: ML approaches where the variable to be predicted or analyzed (e.g., how much a customer purchased) is present in the data
- Unsupervised machine learning: ML approaches that typically require more human judgment, because the variable to be predicted or analyzed is not present in the data
It might seem academic, but this is important. Your ML algorithm is heavily dependent on the quality and integrity of the data you feed it. So if your data contains lots of errors or is incomplete, your algorithm won’t be reliable and won’t be able to make the predictions you need.
If you take the time to learn to walk before you start running, you’ll be able to get in place the necessary processes so you can feed the best possible data into the algorithm you are using.
Principle 4: don’t reinvent the wheel
Understand multimodal and automation-focused ML? Know how to interpret variable importance scores or partial dependency plots? Congratulations, you must be a data scientist.
For everyone else, there is AutoML. AutoML tools can be used even if you don’t have a data scientist and don’t have time to reinvent the wheel. These can help you use AI throughout your organisation to get new, valuable insights faster. Avora’s offering was included in this report among Augmented Business Intelligence Solutions.
Using tried and tested tools will also help with this – now is not the time to try to reinvent the wheel.
As a CTO I’m always looking at how the latest technological advances can help me gain a competitive edge. All businesses need to include AI in the mix when considering any new initiative to stay ahead. Use the guideposts above and wherever possible learn from companies who have already been through the learning curve.