Companies have been investing heavily in artificial intelligence (AI) systems, but all that work is not paying dividends. The digital giants, including the cloud giants and others such as Apple, Facebook, and Netflix, are able to convert data science to business value, but other large enterprises can’t. The Fortune 100 aren’t there because of their old systems and ways. They want to do AI, but they don’t know how to get something out of it.
What are the problem areas? Why do implementations fail? Are companies new to AI, or do they use basic AI? Are there too high expectations from this niche technology? Insights from a new research from Infosys Knowledge Institute, the thought leadership and research arm of Infosys, throw some light on many of these questions.
Here’s a quick snapshot from the Data+AI Radar report: To begin with, companies are new to advanced AI. For instance, 85% of AI practitioners have not achieved top-tier capabilities that are closer to AI that can predict the future. Then, monitoring is the most successful AI task as it is also the most forgiving. Companies need advanced AI if they are to achieve the loftiest ambitions of AI and stand out from competitors. Amongst industry segments, financial services shows the most satisfaction with AI and telecom the least. Around 75% of companies want to operate AI at an enterprise scale.
Also Read: Lensa AI: How to use and everything to know about the viral app that turns your selfies into magic avatars | Tech Tips
A pertinent question arises. Is the global scenario wherein, the nimble, IT-savvy companies are getting the desired benefits from their AI deployments, but the bigger ones are unable to do so, getting replicated in India too? Sunil Senan, senior vice-president and business head, Data and Analytics, Infosys, explains: “There are many Indian companies leveraging AI successfully as these are IT-savvy companies and also from consumer-facing industry segments seeing the growing potential for AI-based personalisation. Food delivery companies are using AI to predict a lot of challenges in real-time, including last-mile delivery of food orders. Edtech companies are making their app more personalised, by providing real-time feedback on the work done by a student on a problem-set. Some of the tech-savvy hospitals in India have introduced an AI-powered personalised health risk assessment technology focused on preventive healthcare.”
According to Senan, inaccurate deployment of AI tools happens when the data from legacy systems is low in quality. Often, the approach to handling missing data generates a bias in the dataset. Data verification and AI infrastructure and computed resources are some of the top challenges companies face today followed by ‘Risk of bias’ in AI.
Most AI applications are also not able to deliver satisfactory business value because leadership or senior executives are not involved. Their involvement is necessary to create a bias towards the business value of the AI deployment, he says.
So what are the key parameters to keep in mind while executing AI deployments? Firms need to ensure they have data quality frameworks and validation processes which in turn ensure trustworthy data as input for AI use cases, says Senan. “A hub-and-spoke data management strategy where companies centralise platform and technology but give teams the flexibility to operate on their own is needed. More AI in the cloud also correlates with greater data sharing. Expanding deep learning and data sharing each has a positive influence on corporate profits.”
According to the Infosys report, companies can generate over $460 billion in incremental profit if they do three things: improve data practices, trust in advanced AI, and integrate AI with business operations. Needless to say, it’s worth walking the extra mile to transform AI dreams into reality.
EXPECTATIONS VS REALITY
81% deployed their first AI system in the past four years
85% of companies have not achieved advanced capabilities
63% of AI models are still driven by humans
Only 26% of practitioners are highly satisfied with their data and AI tools
