AI is focusing the attention of CTOs, CDOs and CIOs across enterprise. At Dootrix, we’ve noticed that the key question being asked by our clients has shifted from ‘What can we do with our data?’ to ‘What can we do with AI’? AI hype is currently significantly outpacing its application in practise, and with good reason; with the landscape shifting so quickly, organisations are wanting to get some clarity about how it will settle before they make significant investments. It’s a very hard arena in which to make predictions, but looking ahead is vital for enterprise to leverage new technologies significantly early to win competitive advantage, without being everyone else’s guinea pigs. That’s why I am sticking my neck out with my predictions for AI over the next couple of years.
AI will be a huge topic for at least the next year or two, comparisons have been made to the advent of the smart phone, although I think the better analogy is to when search engines made the internet truly useful.
I predict AI will have a similar effect, making a range of tools and skills available to people with parallel expertise – business managers might get access to advanced coding skills for instance. In turn, it will herald gains in productivity in much the same way as the IT revolution did.
These two things alone mean we should be prepared for the AI revolution to be as significant as those that preceded it. The companies that use AI most effectively will be dominant, in the way that the first companies to master the power of the internet dominated their sectors for years afterwards.
Another observation that’s worth making is this: if you were a company starting out now, you’d absolutely be building everything around AI. The chances are, however, that you’ve already got an established IT estate. For most organisations then, it’s a matter of how you take what you’ve already got and bring it into the age of GenAI.
Whenever a new technology has the power to transform the world of work, people either adapt and thrive or don’t and struggle, but society overall becomes better off. While professions requiring manual skills, such as hairdressers and electricians, are unlikely to be in the vanguard of change, any profession that rests on being able to command and apply a large body of knowledge probably will be.
For many organisations the essential part of the process of integrating AI will be asking: what tasks would be done more easily and better, with the help of an assistant able to find out and apply almost anything you need to know?
To demonstrate this, I often turn to the example of the HR team. Many HR professionals are great at managing people but less comfortable with technology, yet they’re trying to operate in a very complex environment and handle a large amount of data.
With AI, rather than having an admin assistant to help ensure everything runs smoothly, an HR executive will be able to ask questions or instruct the system in plain English: “What is the company’s current policy on paternity leave?”; “Fill in this form for me based on the answers that have been sent to me via this email.”
So, what we’re going to see, in many areas of work, is that digital helpers are going to become smarter, interact with us more intuitively and more naturally and generally become more helpful.
AI will not necessarily take your job, but your role may be at risk of going to someone who knows how to make better use of AI than you do.
The real question is what happens when AI allows someone with a particular area of expertise to expand their portfolio of skills to another area where they have less experience? Subject-matter experts, may be able to complete a range of tasks far more quickly and without help, but they will also become a source of vital support to engineers. So, far from becoming redundant, they become even more valuable.
There may also be a subtle shift in dynamics. Hitherto software engineers have often acted as gatekeepers. Now those business analysts who can articulate and frame the problem will start to play a more pivotal role thanks to better AI. That’s likely to make it easier for them to work with solution architects to solve those problems optimally both from a business and engineering perspective. Right now, the technology is still a bit rough around the edges, but we will soon be seeing situations where you can draw a flowchart, take a picture of it, feed that to GPT4 and watch while it creates TerraForm scripts, the script which lays out the architecture in Azure.
In a way, AI empowers the ideas people, much as the world of 200 years ago was shaped by engineers with both imagination and an entrepreneurial streak. Given access to good tools and a wider skill set, creative thinkers can go from conception to execution much faster.
If your organisation is looking to introduce GenAI, you’ll need to consider what’s often called data hygiene: controlling what can get into your system, what’s in it and what can get out of it. A key element of this is data preparation and cleansing.
The first thing you need to do is to get your data house in order. This means making sure that your own version of ‘the truth’ is correct.
It may be time consuming, but it’s important – and where it’s not done properly, it can lead to project failure. Think of it as the data foundations on which you plan to build.
Among the key tasks in the cleansing process are:
When you have cleansed and pre-processed your data, you’ll need to address other aspects of data hygiene, such as controlling what can get into and out of your system.
When smartphones were starting to become commonplace, some organisations found themselves with security issues. The new devices were – by and large – great, but the cultural shift was so fast that it happened before the security implications were fully addressed.
There’s a danger of the same thing happening with the arrival of GenAI. We tend to forget that the public versions of AI are just that – public. There will be organisations where engineers are submitting the code they’ve just written to the likes of OpenAI or uploading it to Co-pilot on GitHub. It’s easy to think of such an upload as a private connection between the user and the AI, but it’s not.
Equally, there’s no point sorting out your data if you can’t ensure that bad or dangerous data can’t be added to a cleansed set.
Ensuring that the compliance and security layers of your systems are in place is essential. Building a solution that is both secure and convenient is another must. However, there is also a cultural dimension to this issue that’s best dealt with through education: establishing strong, clear protocols and a security-orientated mindset.
If this all sounds like a lot of work, I’m afraid you’ve hit upon a truth of business leadership. Most worthwhile things require effort, not least ensuring your people are onboard. But if you’re prepared to take a methodical approach to adopting GenAI, the potential benefits are huge.
If you'd like to get ready for AI, get in touch