Advances in artificial intelligence (AI) over the past couple of years have made the technology an important part of business transformation. Whether in search of efficiencies, savings or simply more advanced ways to deal with increasing quantities of data, businesses from all sectors have begun to seriously consider how they can implement AI to achieve an advantage.
However, AI is not the panacea many would have you believe; as with all technical investments, the success of an AI installation is highly dependent on the way it is selected, implemented and integrated with the rest of the business. AI does deliver incredible ROI for some businesses, but for others, it can result in wasted time and financial losses.
This is because there are tangibly ‘right’ and ‘wrong’ ways to implement AI. Approaching it as a vanity project, or without a clear objective, will inevitably lead to problems with its adoption and performance. So how can businesses ensure that their AI implementations provide return on their investment?
Necessity or vanity?
This first point should be obvious, but implementing any technology for the sake of it will lead to failure, one way or another. The seemingly miraculous results that some organizations have seen since adopting AI have not been achieved by magic! Businesses that are turning to ambitious AI transformation projects for the optics are inevitably throwing money down the drain. The first step to achieving ROI with your AI project is making sure it’s aligned to your business needs, and is applied to solve a specific set of problems. Viewing AI as a miracle tech fix is a recipe for failure.
Speculative technology implementations are rarely likely to find the problems that need solving efficiently. Instead, all transformation projects of this kind should begin with a thorough appraisal process to evaluate where AI will be most effective.
If you’re in, you’re in
Once you’ve identified one or more potential use cases, decisive action is important. According to a report [link] from Boston Consulting Group (BCG) and MIT, in which companies were ranked according to the degree of AI integration into their operations, their commitment to AI and the depth of their understanding, the companies that fully commit to implementing AI are the ones that show the greatest return on their investment. Those that are ‘dabbling’ or just ‘testing the waters’ are the ones that struggle to see any significant returns.
To build or not to build?
Once they’ve committed to understanding the impact AI could have on their use cases, business leaders should ask themselves three key questions to decide whether to build an AI solution in house. Firstly, is this for a business capability where I am already more advanced than any competitor or supplier, and will it provide me an enduring competitive advantage? Secondly, do I have access to more data that is better quality and more representative of the real world than competitors or suppliers can access? Thirdly, do I have the access to talent, capital and organizational buy-in required to make a success of an ultra-high risk technology project? If the answer to any of these questions is no, do not build it in-house!
The most serious barrier to any in-house AI project is access to data. An AI needs a lot of good quality data that accurately represents the environment it’s operating in. This enables it to continuously learn, optimize, remodel and apply its ‘knowledge’ to your own business. However, with a few exceptions, it’s impossible for in-house solutions to access the enormous, industry-level data that specialist solutions will have, particularly when they involve hundreds or thousands of third-party, or even competitor, datasets.
Another difficulty for in-house AI projects comes in linking the technical expertise with deep domain-specialist knowledge. Finding and hiring talent will be crucial to the project’s success, but in an extremely competitive (and not to mention, expensive) field, your business will automatically be competing for talent with tech giants that can offer phenomenal employment packages, working environments and diverse problems to solve.
As the report from BCG and MIT notes, there is an important difference between the roles played in an organization by AI ‘production’, or the development of an in-house AI capability, and AI ‘consumption’, which includes the buying-in of available solutions. While elements of both in-house ‘production’ and ‘consumption’ exist alongside each other in the most successful AI implementations, the report finds that buying-in solutions is where the real gains are made, particularly when it comes to scalability and transformational capacity.
Not ‘just another IT project’
Perhaps surprisingly, the BCG and MIT research reveals that returns on AI investment are significantly higher when the projects are the responsibility of a C-level exec that isn’t the CIO. IT teams will be familiar with the challenges of becoming siloed, and the importance of integrating transformation projects across the business, and the need to appropriately message and manage any major change. This is especially true for AI, which shouldn’t become ‘just another IT project’.
Ensuring that your CFO or COO, for example, is involved in, if not in charge of, the selection and adoption process of your AI implementation immediately ensures wider buy-in for the project across the company. It is crucial that technological transformations like this fit within wider business objectives and, most importantly, complement cultural transformation as well. Inevitably, the CIO and IT team will be central to these technology projects, but maintaining organizational collaboration is crucial for generating good returns.
This organization-wide collaboration includes training and awareness efforts, making sure that all departments of the business are included in the process and educated in the objectives and benefits of the project. As part of this training, it is important to ensure that the interface between the technology and the human personnel is productive and allows for iteration in order to best fit the needs of the business. A major part of achieving this cross-organizational cooperation comes from building a culture that embraces change and is able to continuously adapt and improve. Benefits for your company’s staff should always be at the heart of these strategies, ensuring you get whole-team buy-in and don’t create larger-scale operational problems with your technological solution.
AI can rapidly generate impressive ROI (we often see this achieved in less than four weeks), grow profits and dramatically cut costs when it’s implemented strategically. But, in order to fulfil its potential, it must have proper buy-in at an organizational level, and not simply be siloed with IT. Clearly-defined and practicable objectives are essential. With clear use cases, the right personnel and decisive leadership, returns from AI can be transformational and quickly realized.
Philip Ashton, Co-founder and CEO, 7bridges