Some months ago I wrote a blog about Beyond the hype in 5G, AI and IoT (LinkedIn article). The world around us has changed due to the pandemic. For several companies Covid-19 has worked as an accelerator to drive digitalization and new ways of working, but for some it has delayed investments.
We at Silo AI have worked closely with our customers for the past three years to understand how AI can make a significant impact. There is no silver bullet, but I believe we have identified four principles that set the foundation for success.
Four principles for successful implementation of AI
1. Ask why
2. Build bonds
3. Keep learning
4. Be good
AI is everywhere, and hopefully soon reaching the top of its hype curve and moving into actual implementation. There are so many possibilities that could be done. There have been countless explorations, numerous proof-of-concepts, but still limited amounts of productized solutions.
Many successful companies have embraced the ability to fail fast and celebrate the learning from mistakes. But too often the approach has turned into a pilot factory, where nothing gets productized. So, there must be something to be improved?
100+ AI projects in production
At Silo AI, we have helped more than 100 AI products and solutions get “out there”, into real use. Based on our experience, we strongly emphasize picking the most important projects for the incorporation of AI.
As you might remember, at the turn of the millennium everyone was building websites. In the 2010s, everyone was building apps and now, everyone is making AI. Just doing something because everyone else is doing it, doesn’t make sense. Building another AI gimmick is unlikely to change the future – therefore you should use these powerful technologies in cases that actually matter for your business.
We want to help our customers ask why. What would be the single most important thing that you’d like to change if you had a magic wand? Is there something where your industry has struggled for decades and you’ve given up hope trying to solve it?
Another acid test we tend to use is the business case. If there is a need for a business case requiring tens of Excel sheets with multiple variables, perhaps it is time to go back and ask why you have this problem. Would you be able to find something where the benefits are so clear that the calculations are clear and easy.
The remote working model makes building relationships and trust much harder. Nevertheless, those are more important than ever. The traditional industry boundaries have broken down and incumbents have fallen. Staying competitive requires fostering innovations and sharing knowledge across the traditional business lines.
Identifying the right problems to tackle with the help of AI
We have seen amazing AI talent struggling to deliver results at several companies, because they were not able to identify the right problems. On the other hand, people understanding the business or processes may not understand the possibilities that AI enables.
Although understanding the business and domain is essential, the wealth of experience external consultants can bring to the table from various other assignments will help make the right choices, avoid the worst pitfalls and speed up the delivery. Building the algorithms is just part of the equation. One needs to collect, validate, and annotate the data. Depending on the solution, you need to find the optimal mix between embedded-, edge- and cloud computing as well as make sure data flows optimally through sensors, with suitable connectivity and integrations to the other systems. Someone also needs to tackle the privacy, regulative and security issues and multiple other aspects.
There is no single company, no single hardware or software provider or integrator to make this happen alone. Therefore it is crucial to build bonds and build trust between different actors as this expertise will benefit you when scaling your solutions.
Unfortunately, no single technology or product alone can create business benefits, nor solve a challenge. The value will only come from the ability to look beyond individual technologies and understand the dependencies not only with new and exciting things, but also with existing legacy systems.
When solving a problem, choose the right areas to ensure you have an adequate business case. It doesn’t mean you should eat the entire elephant at once. Get started, solve a piece, and improve.
Even though AI as a science or technology are not new, the speed of the development we are seeing is unbelievable. What is seen as cutting edge today, may be completely outdated in six months. Things that were not considered possible earlier, are being implemented at ease.
Start early and ensure flexibility in your solution architecture
The fast evolution doesn’t mean that you should wait, but actually the opposite. If you never start, you never learn. It is also much easier to adjust and manoeuvre the course of a moving vessel than one that is sitting still. When defining your architecture, keep in mind that things will keep changing and improving. Often an experienced solution architect can ensure that components can be added, removed or replaced without the need to rewrite everything.
Learning naturally applies beyond technology. Therefore you need to identify the activities required to drive the cultural change, the actions needed to ensure increasing the awareness of new possibilities and enabling multidimensional communications. Several of our customers have seen these challenges and we have together designed programs to drive AI readiness.
We should ensure that we understand how algorithms work, and how they learn from data. Technology as such isn’t good or bad, but without taking proper attention AI can become biased and thus result in unwanted behaviour.
Our society is also facing unseen challenges in terms of sustainability. United Nations’ Sustainability Goals are a good start on addressing those challenges. When planning the next solution or innovation, it is always good to think on whether there is a more sustainable way of doing it. Not just because it is the right thing to do and the decisions we are making are impacting the future generations, but it is also good business
Being good doesn’t only reflect the battle of good and evil and the global challenges, but for everyone on a personal level. As mentioned, trust, bonds and learning are vital. One kind word, one advice or one question asking why can make a difference in the people around us and make life better for all of us.
Asking why, build bonds, continuous learning and being good is really essential can make the difference. I don’t think that it is a coincidence that those are also our values at Silo AI. These values guide us every day and help us identify and solve the most important challenges our customers are facing.
Roope has 15+ years of experience in building global digital businesses Fujitsu, BBC, Honda, National Geographic and Nokia. Throughout his career, Roope has strived to enable smarter and more efficient business through advanced technologies.