New Delhi, June 23 -- Companies across various industries, such as finance, healthcare, e-commerce, and manufacturing, are rapidly embracing Artificial Intelligence (AI). While AI excels at automating repetitive tasks, analysing large data volumes, and predicting outcomes, there is a caveat. The lack of quality data often acts as a barrier to effective AI implementation. Therefore, experts believe, before hastily integrating AI into business processes, it is crucial to ensure the correctness of the data.
Data quality is defined as the degree to which data meets a company's expectations of accuracy, validity, completeness, and consistency. Quality data is key to making accurate, informed decisions and is especially key to the successful implementation of AI-led tools and technologies.
Debojyoti Dutta, Chief AI Officer of US-based technology provider, Nutanix, explained that AI systems require substantial amounts of data. "They utilise what is known as 'training data', as these datasets serve as instructional material for the AI model.
"If the data is insufficient, incomplete, or inaccurate, the resulting AI will produce subpar or incorrect results. Hence, the effectiveness of AI models in your business is fundamentally dependent on having suitable data that is well-organised, clean, and thorough," he said.
Why is good data so important for AI?
While it is known that good data is key for AI to boost business and innovation, a survey by market research firm Gartner published in March demonstrates that over 63% of organisations either do not have or are unsure if they have the right data management practices for AI. The survey found that organisations that fail to realise the vast differences between AI-ready data requirements and traditional data management will be at risk of failing in their AI efforts.
Experts like Gopal Patwardhan, CEO of GTT Data Solutions, believe that clean data is key to AI innovation because AI itself isn't innovative; it's a trained model that relies on data to set objectives and improve performance. For instance, AI analyses historical data to predict customer behaviour or enhance processes, requiring precise and relevant data for credible results.
Prakash KS, Head of GenAI CoC at Siemens Healthineers Global Development Center, emphasises that AI output accuracy hinges on data quality. According to him, flawed data leads to unsatisfactory outcomes, regardless of AI sophistication. "Biased, incomplete, or inaccurate data will produce skewed results. Conversely, clean data enables AI to provide valuable insights that improve business strategies, operations, marketing, and customer service."
Moreover, AI cannot fix bad data-it can amplify its flaws. As Dutta noted, contrary to the perception of AI as a magical solution, it can worsen data problems by relying on inaccuracies and biases, disrupting operations. Well-organised data empowers AI to reliably expedite business processes and boost productivity.
The need for using clean and real-time data for building various AI systems will continue to increase with the emergence of more advanced AI techniques. "You cannot have generative AI without clean, real-time data flowing through your systems with the right permission, guardrails, and governance," Ashok Srivastava, Senior Vice President and Chief Data Officer at Intuit, said in an interview with TechCircle.
Ensuring quality data in AI projects
Companies should begin by identifying and correcting errors, eliminating duplication, and verifying data accuracy. Properly investing in data quality early in the process can help support business strategy and finances. Trying to fix errant data at later stages can be costly. Correcting it at the beginning can prevent important mistakes from occurring in the first place.
"We should begin an enterprise's data landscape, then build a clean, structured data foundation that the entire organisation can utilise, believes Patwardhan. He added, rather than focusing on isolated point solutions, we create a strong database and layer AI tools on top to build the applications and interfaces that clients need today.
It is important to normalise data from various systems so the AI can work with it, said award-winning data scientist and AI researcher Akshata Upadhye. "There's a need to bring data from different places (e.g., sales transactions, HR documents or customer info) into one place to allow AI to have a more complete look at all needed features for better suggestions." She also advised on enforcing data governance rules for clear ownership and responsibilities regarding security, privacy, compliance, etc.
Besides, businesses should address biases in data, as it prevents AI from perpetuating unfair outcomes. Keeping data updated is also important for AI systems to generate relevant insights.
Ultimately, clean, accurate, and organised data enables AI to boost business, drive innovation, and improve RoI and long-term viability, believe experts.
Published by HT Digital Content Services with permission from TechCircle.