By Martyn Whistler, Associate Director, Knowledge, Ernst & Young LLP.
Very often these days, we hear people expressing fear about how artificial intelligence (AI) will take over the world and rule humanity – or at least business life. Most of these conversations validate my own fears: that our own understanding of AI – both soft AI (which aids us in tasks such as sending texts, booking hotels or restaurant reservations and ordering taxis) and hard AI (which makes decisions on its own and mimics or even replaces human intelligence) – is inaccurate or incomplete.
We might not have realized it, but AI has become almost omnipresent now. It is part of virtual personal assistants, voice recognition software, call centers, weather forecasting, power grids and even large industrial processes. If you look at the conversations around digital in recent years, they have been mostly about connecting devices, networks and data to create a digital environment from where data can be collected and manipulated. By introducing AI into these connected environments, I believe it is possible to achieve results that have previously been considered impossible.
The obvious benefits and the not-so-obvious challenges
The more obvious benefits are efficiency and effectiveness. Organizations become more efficient by empowering bots to answer customer queries online or by using software to resolve customer queries before they get to a human handler. Efficiency gains can be significant. In power grids, for example, the traditional approach was to have engineers monitor grid functioning. They could push capacity but their goal was to keep the lights on. With AI, however, this monitoring can be more automated and the power grids can potentially be optimized to 100% of their capacity. More importantly, the power grids can function more intelligently and help enable management to make decisions on where in the network to scale up or down, or on how to optimize networks to levels unattainable by solely human control.
AI can also help improve effectiveness by, for example, helping a doctor with a diagnosis. The AI-powered diagnosis can potentially be far more accurate than a human diagnosis, and the medical staff can focus on attending to the more human and subjective aspects of health care – so everybody wins.
There are obstacles to implementing AI too. Cost is certainly one. So are legacy and work forces with institutional knowledge that can act as roadblocks to addressing the low-hanging fruits. Although AI has evolved over the years, it still has significant untapped potential and scope for innovations. I wouldn’t say companies are yet to acknowledge the need to bring AI into their organizations, but that many are only starting to figure out where and how AI could fit into their organizations.
How can CIOs get the ball rolling?
Start experimentally: Adopting an enterprise-wide AI strategy in one go will not be easy, so CIOs can start small – probably by identifying challenges AI can help address and then expanding slowly to the rest of the organization. But, to get there, organizations should start by experimenting and learning from the challenges and to help stay ahead of peers when AI becomes more common.
Talk to C-suite peers to identify opportunities: There are sector-specific nuances that organizations will have to work out depending on the domain of their organization’s operation. However, regardless of the sector, AI could be implemented to help improve efficiency in almost all areas of an organization, including marketing, finance and logistics. So having conversations with peers such as the CFO, the CMO and supply chain leads are vital.
Spot the AI that your company can benefit from: There are many high-level research-oriented discussions happening around AI. While CIOs can benefit by following these discussions, their focus should also be on what can be implemented in their organizations today. The number of organizations offering AI solutions can itself seem overwhelming. Identifying a solution that works for your organization can be a challenging, and unavoidable, activity.
When AI controls decisions, who controls your company?
Before CIOs even try to answer this question, it is important to say that we are very far away from reaching the stage where AI can operate without human intervention. But when AI becomes common, organizations will have to learn to manage risk and make sure they have the right balance between the human and technological elements. AI, in other words, is a supplement to human decision-making, not a replacement. It can automate numerous transactions in sectors such as financial services. But can it replace salespeople in a retail store or the nurses and doctors in a hospital? Probably not. But it could certainly make their jobs easier and help them make good decisions; it could also free up their time to work on more value-added tasks and customer/patient focused activities. AI should not be seen with fear, but as an opportunity.
EY would love to hear from you. What are your experiences with AI? Share your ideas and comments, and let’s start a lively discussion!