AI is much more than a technology: it is a new approach to both how systems are built and how they are used. As a result, being ready to get the most out of it requires more of an organization than previous waves of technology adoption which only concerned the technologists. While enterprise applications have evolved from green screens to gray fat client apps to colorful browser experiences, the underlying functions and capabilities have remained largely unchanged. And most importantly, predictable. Users only needed a basic introduction to move between these styles of user experience and keep on working, oblivious to any underlying technological or infrastructure changes. AI changes all that. And an AI-ready organization needs both an adaptable culture and sound business foundations.
While most people are talking about models, NVIDIA, risk and other implementation artifacts, the starting point is ensuring that your organization has the right culture to apply AI successfully, and that is an adaptable culture. Adaptability is built on three key attributes: openness, safety, and education. Let us look at each in turn.
Openness is a willingness to not just accept change, but to welcome it with open arms. Organizations are all too often ossified at several levels, and these strata of resistance have to be gently brought to realize that it is possible to combine the best of the old and the new. Not everyone will make that journey, and they need to be sorted into those that simply will not change because they believe they are right, while those that struggle to adapt will need training and help. Often the former group are found in the ranks of senior management who have also become detached from the everyday operations of their businesses. The latter category are more likely to be the people doing the actual work, who may lack the required education, are used to having new approaches and rules just dumped on them, and are probably concerned for their jobs.
Safety is an essential part of openness. People need to feel safe in their workplace, not just in the standard health and safety way although that is important too, but in being able to express their opinions, raise issues and generally being able to be vulnerable. Aggressive, dominating managers and the inevitable micromanagers stop people expressing their views through fear. This is why most digital transformation programs fail: both the users and the implementers are all too often afraid to speak out about issues. We cannot allow the same to happen with AI: we need to bring everyone along with us, we are learning and discovering. A humble approach is unquestionably the best, infusing the workplace with sufficient trust that it allows people to display confident vulnerability.
Education is, clearly, essential at all levels of the organization. Education is critical to eliminating fear of change as it brings understanding and awareness. Where an unknown capability is forced on a group of people, be they senior managers or factory floor workers, the reaction is bound to be one of anxiety. Whether the anxiety is over redundancy, loss of status, incompetence or being doomed to irrelevance, the appropriate training can eliminate or reduce it. Many people find the unknown intensely scary, so an appropriate introduction goes a long way to resolving concerns. Neither superficial explanations of the basics nor technical deep dives are appropriate, what is required is a training targeted to specific business roles, something that we are actively building at Data Kinetic.
Your organization will not be ready to Apply AI effectively without these three pillars. And getting them right is much more difficult and time consuming than the technical aspects.
While we are seeking adaptability from our people, AI needs reliability and solidity from other parts of the business to ensure success. Again, I am not talking about building out expensive compute infrastructure, or signing up a basket of SaaS providers, but looking at some of the softer aspects of the business. Specifically, we need to know we can rely on the data, the processes and the portfolio of applications on which everything runs. Again, let’s look at each in turn.
Data has been called the new oil, and much has been written about data scientists and other roles, yet most businesses do not truly care about or look after their data. 80% of business data is what is called unstructured, meaning such things as PDF and Word documents, and spreadsheets. Most of these are dumped into enterprise repositories such as SharePoint, Dropbox, and Google Drive which become part lumber room, part dumpster. Rare is the organization that truly curates and manages this often random assortment of stuff, leaving it duplicated, out of date and all too often erroneous. Humans can usually tell if the documents they are looking at are past their best or plain wrong, but AIs ingesting them for training do not. The result is a resurgence of the old adage of GIGO — Garbage In, Garbage Out. A key step towards being AI-ready is to recognise that if data is indeed oil, it needs to be refined. Organizations need to start investing in data quality as well as uncovering other information resources that have not been previously considered data such as images, videos and streams of data from IoT devices. It is going to be immensely important in the future.
Processes often get a bad rap, but they form an important part of how businesses operate. Part of our AI Strategy work is helping people identify which processes should be AI-enabled first, based on business priorities. Often the objective is based on a negative view: to speed up processes by automating steps, either because it was too slow or too expensive. Less commonly the goal is positive, aiming to find novel ways of working, or identifying new processes that were too difficult or expensive to do without Applied AI. Where a project relates to an existing process, the first step is to fix its existing problems. Automating or accelerating a broken process is a terrible idea: delivering the wrong results faster does not help anyone. Curating and correcting business processes prior to applying AI to them is an essential step that is all too often forgotten. Things are likely to work better even without the AI.
Portfolios of applications build up over time as and when the need arises for new capabilities. M&A brings not just new staff and customers, but another jumble of applications. Nobody plans to have a set of often disconnected applications, it just happens organically. Two divisions, for example, independently looking to improve customer engagement may select different CRM systems, or internal recruiters select an on-boarding system that does not integrate with the main HR management platform. Easy SaaS subscriptions on the company credit card and departmental applications make it all worse. Portfolio management is, as a result, part an art and part a juggling act, ensuring that scarce resources are applied for the greatest business benefit. Adding AI to the mix changes these calculations. Some decisions are easy, for example there is no point in rushing to apply AI to an application that is used once a year by one person, but after that the decisions are more nuanced. But vital. And at Data Kinetic we’re deploying our Future Practice approach to ensuring that yesterday’s best practices are not holding back, or worse. misdirecting your application enhancement program.
Making your organization AI-ready is a big job, so don’t make it even bigger by starting in the wrong place. Sure, proof-of-concept projects help build some tech skills, but they won’t magically scale and make your whole business ready for AI. The six pillars identified here are the starting point for change to ensure that your AI enablement will not go wrong the same way that more than half of digital transformations did. Building a strategy that prepares the workplace for AI is essential to success. Hopefully these points will point you in the right direction, but we’re here to help with our uniquely flexible subscription advisory and AI technology services. Contact us and see how we can prepare your organization for successfully adopting Applied AI.