It’s an undeniable fact that Artificial Intelligence (AI) is changing the way we live. From facial recognition and voice technology assistants to self-driving cars, AI has crept into our lives and as consumers, we utilise it without a second thought.
But its impact across a wide range of business sectors is perhaps the hottest topic in tech right now. AI has developed and matured to the stage where, for some functions and operations, the levels of accuracy have overtaken human skills.
Yet with stories of mind-boggling complexity, escalating project timescales and spiralling costs, the much-hyped technology is still regarded by many business owners with confusion and perceived as a risk. Justin Price, Data Scientist and AI lead at Logicalis believes that knowing what you want to achieve and setting realistic expectations are the best guarantees of a successful first AI adoption.
Choose the right AI tool for your business need
During my meetings with clients, and from talking to CIOs and business users, it has become apparent that confusion reigns over the terminology used to describe Artificial Intelligence.
There are three key terms at play. All three fall under the umbrella term of AI, and are often used interchangeably but each has a different meaning. AI is a technology that retrains or ‘learns’ patterns and other specified behaviours to achieve a set goal. Crucially, it is about producing something which didn’t exist before.
Other types of AI behaviour include:
- Robotic Process Automation or RPA, a software designed to reduce the burden of simple but repetitive tasks done by humans.
- Machine Learning – essentially probability mathematics used to spot patterns in very large samples of data.
- Deep Learning – a Neural Network which mimics the way the human brain works to examine large data sets, HD images and video.
Being aware of the subtle differences and uses of these terms allows a greater understanding of which tool will best-support your business’ data insight needs.
Make sure your data is up to the job
When delivering an AI project, around 80% of the total effort and time will go into making sure your data is correct. Underestimating the importance of top quality data is a common pitfall for organisations because, just like any other IT tool, AI will perform poorly if you have low quality data. So the focus in every instance must start here.
Data must be well structured and it must be in format that’s consistent and compatible with the AI model. Don’t forget that AI is a process which must be regularly re-trained to ensure accuracy, so ongoing maintenance is essential.
It’s important to have an idea where your AI solution will be hosted and how many people are going to use it. If you don’t know, get some advice. Once your data is organised, it helps to have an understanding of what you have and where you can get it from. This is where traditional business analytics come in, giving you a good idea of what value you can get from your data before you start to use AI to drive recommendations.
Brace yourself for complexity
Never underestimate the breadth and complexity of what is involved in building and delivering an AI project. For many CIOs this will be their first experience of AI and, even with the right data, there are many variables at play that can add to both the costs and timescale of implementation.
Working with the right partner is essential to guide you through the first project. We recommend undertaking an initial project with a fixed fee whereby you can deliver a functioning result while you establish trust and credibility with your solution provider.
As well as the importance of good data, another critical factor in delivering a successful AI project is finding a solution that is scalable. It’s one thing to write an AI model on a laptop. But it’s a completely different thing to write a model in a scalable way that will survive a deployment across a business. This is where getting expert advice will help you decide on the correct infrastructure to support your project.
We advise clients to consider pre-built services. All the major IT players have been quick to offer a range of on-premise and cloud solutions, and we’d highly recommend looking at some of these instead of trying to build your own.
Know when AI is (and isn’t) the right tool for the job
The concept of AI has been around since the 1950s but now, suddenly, everyone is talking about it. That’s because finally it has become commercially viable. We have the data, the processing capabilities and the skilled people to harness its power. With a plethora of impressive use cases available to businesses spanning most sectors, it’s no wonder AI is the tech tool of the moment.
But AI is just another technology and won’t be the right choice for every business looking to gain insight from their data. The importance of prioritising people, process and culture in any AI project has already been discussed by my colleague Fanni Vig in a previous blog, and this is absolutely essential to ensure your business isn’t trying to use AI where a different, but still hugely valuable, tool could deliver the desired results.
At the highest level, AI allows you to work through far larger data sets than previously possible. It can be used to help automate your data workflows, redirecting low-difficulty but high-repetition task to bots, which allows people previously engaged in these tasks to work more efficiently.
This process creates a new way of working that may have greater implications across the business as roles change and skills need to be channelled in different directions. Introducing AI is a truly cross-business decision. And let’s not forget that, at the most fundamental level, using AI to harness your data is an investment that must show a return.
Finally, get advice from the experts
While the impetus to adopt AI may come from the IT department, the results generated can help drive cross-company productivity; help differentiate businesses from their competitors; and delight your customers through a more tailored services. The impact cannot be underestimated. But neither can the complexity. So if you are a business considering whether AI could help you get more from your data, my advice would be to work with a trusted solution provider to guide you through your first successful deployment.