A recent survey has shown that only 21% of organizations have a clear Artificial Intelligence (AI) strategy, with C level support and applied across the enterprise. The road from proof of concepts to larger scale programs can be difficult if the best practices are not in place, ranging from a methodology framework to a needed diversity within the AI transformation team.
AI is already changing the business environment and is bringing a variety of challenges to industries. From healthcare to finance to professional services, many industries are preparing to face those challenges. So far, a majority of companies are working on AI with a pilot approach, which they then implement within certain areas of their business without a clear strategy aligning the different programs. This results in fragmented applications and prevents companies from benefiting from the full potential of such technologies.
Pilots and experiments are necessary and fundamental for any AI project or initiative to get started. Nonetheless, a specific way to measure its efficiency and its results needs to be implemented across the whole organization and not only from a technological, but also from a business, perspective. Performance analysis and measurements against goals can help highlight concrete added value brought by new AI powered solutions. Setting up goals and measurements are one step, but how do you make sure that the topics on which a company is working on are the right ones?
In order to gain the maximum advantage of an AI implementation, the AI system needs a lot of learning data to provide results on one specific task or action. It needs to be built, configured and trained with a very specific purpose. This technology is not the right one for general and undefined tasks for random application across random use cases. This is why the scenarios need to be very specific and very well defined. The main areas where executives expect results from AI projects are mainly “improve or develop new product/services”, “achieve cost efficiencies” and “accelerate decision-making”. This can be tackled by combining a business approach and looking at different problems and asking “how can we solve it?” with a technological approach by knowing the potential of a specific technology asking “how can we improve our processes by using AI?”. In a nutshell, having the ability to narrow down to an area where improvement is needed and identifying even a very specific task where AI could help an organization, is key.
Many companies still struggle to go from a successful Proof of Concept (PoC) to a larger scale enterprise program. The difficulty to find the right people with the right skills to drive this kind of transformation program is still present, even if the trend is declining. Other challenges seen on the market are for example the integration of AI in existing processes or activities. It is given that without a clear strategy in terms of adoption, a training plan to empower every person within the organization to be part of a continuous learning atmosphere, it is quite difficult to leverage the necessary value from AI.
The specific applications of AI with clearly defined goals, as a road to success, will also have an impact on the employees’ view on AI. Changing the perception of AI of a job or reducer and replacer to the benefits of AI allowing an augmented workforce to focus on value added tasks that bring competitive advantage to the company while AI technology is focusing on the automation of repetitive tasks.
Associate Partner, Digital & Analytics, EY Luxembourg
Manager, Digital & Analytics, EY Luxembourg