Range of AI—Human assistant to human replacement
The combination of parallel processing power, massive data sets, advanced algorithms and machine learning capabilities are spawning varied versions of AI systems.
Today, AI capabilities vary from specific/narrow to super, all-encompassing AI.
Narrow, or specific AI, is an intelligent assistant that can aid humans in making complex decisions and enhance their cognitive powers by processing vast amounts of data. It can conceptualize and correlate data, recognize the patterns and deliver intelligent output.
For instance, soft AI can be used to detect frauds in various sectors such as banks.
A large sample of fraudulent transactions is fed into the AI system, which is trained to look for signs that separate fake transactions from genuine ones.
Another example of soft AI is the voice assistant that can understand voice inputs, analyse data about the users from a variety of sources (social media, smartwatches, etc.) to better understand their behaviour and deliver results tailored to users’ preferences.
Super, or strong AI, aims to make decisions on its own without any external support.
These machines can think, learn, decide and converse like humans. Hence, they have the ability to replace humans altogether.
However, super AI systems are yet to achieve breakthrough improvisation to fully comprehend human mind-maps and replicate human intelligence.
How is AI different from RPA and cognitive?
Though enterprises are increasingly understanding the benefits of AI, there still exists misperception around similar technologies—AI, robotic process automation (RPA) and cognitive.
AI is described as the decision-taking capability based on simulation of human intelligence processes by machines. These machines “can act" as human.
On the other hand, cognitive computing helps humans in fully or partially delivering judgement-based processes and assists in their decision-making. These systems deal with unstructured inputs, and “can think" as humans.
The third type, referred to as RPA, can automate rule-based tasks and “can do" what humans can. Such systems lack self-learning capability and are effectively dumb: they just perform exactly as programmed.
AI-use cases in business
As customers are becoming increasingly demanding, AI offers assistance on key requirements of evolving business:
• People-centric: The AI systems enable the enterprises to shift to a people-centric approach from being process-centric. The decisions are made based on unstructured real-time data rather than pre-defined processes. For instance, ride-sharing companies predict fleet demand based on factors such as weather forecasts, time of the day and historical customer behaviour.
• Ease of use: AI enhances customer experience with offered convenience and assistance. For example, enterprises are using “customer digital assistants" that can recognize customers by face and voice to have relevant conversations, and provide tailored choices to help them make purchasing decisions.
• Self-adaptive: AI has the capability to self-evolve, make connections between data, improve on past decisions and get smarter. For instance, machine learning-based intelligence enables an enterprise to improve sales performance by accurately predicting cross-selling and up-selling opportunities.
AI implementation strategy for enterprises
AI has the potential to disrupt the core of business processes. However, blind adoption of technology and hype-based purchase may not lead to the desired results. Enterprises can ride the wave of success with efficacious adoption of AI technology:
• Getting familiar with the concept: Rather than adopt the technology in haste, enterprises should first educate themselves on the basic concepts and capabilities of AI. The better a company understands what narrow/soft AI does, the more likely is its successful adoption.
• Identifying the problem to which AI is a solution: Enterprises should identify specific use cases in which AI could solve business problems and help them achieve specific project goals. They should further narrow down the possible AI implementations by assessing potential business and financial values.
• Bridging the talent gap: AI requires talent pool with a strong understanding of advanced programming, domain knowledge and business context. Enterprises should bring these skills together instead of waiting for one person to bring all the dimensions.
The importance of AI is well understood. However, its implementation remains limited.
It is imperative for firms to start applying AI for solving narrow-scope problems before expecting it to disrupt the core of the business.
AI can be employed for everything from managing targeted advertisements to optimizing logistics to tracking assets to understanding the customers’ social behaviour. The trick is to get started on the right note.
Milan Sheth is advisory partner and technology sector leader at EY India. The views expressed are personal.