Compliance, culture, and cost: Experts highlight data challenges, focus on solutions

Big data is now an essential component of business operations worldwide. But persisting data challenges, be it siloes, limited reporting capabilities, slow analytics, poor quality and missing data, or maintaining scalable cloud platforms, impact companies significantly.

Data challenges can occur at any level in an organisation, leading to penalties, poor decision making and performance, and the loss of consumer trust. Navigating data challenges necessitates exploring strategic solutions involving operational efficiency, data migration and integration, security and compliance, cross-functional collaboration, and building a data-driven culture.

Oracle recently brought together an esteemed group of experts for a round-table discussion titled ‘Navigating Data and AI Integration for Optimal Performance’ to discuss these challenges and explore how Oracle Cloud Infrastructure (OCI) services can assist companies in truly leveraging the power of data. 

Speakers at the round table included Malaya Rout, Director, Data Science, Exafluence; Pallavi Chandolia, Founder and CEO, DotSmart Technologies; Subramaniam Hariharan, Senior VP & Head of Engineering, Zuci Systems; Saifudeen Khan, Director, Zuci Systems; Thiyaga B, Director Engineering, CaratLane; Mrudhula S, Head, Agrya; Sujatha Iyer, Head of AI (Security), Zoho Corp; Deepak Rajmohan, CEO, GreenPod Labs; Bharat Venkat, Sr. Technical Director, Denodo APAC; and Devanand S, Head - IT, Go Colours. The discussion was moderated by Swetha Kannan, Deputy Editor, YourStory Media. 

Unlocking data management challenges 

Businesses know that quality data can help them make informed decisions, identify new opportunities, and stay ahead of the competition. However, they struggle with data management, integration, and navigating AI. Data siloes can cause fragmented processes and departments across an organisation, resulting in poor decision making. The lack of a proper data governance system can compromise data security, neglect data compliances, and put data quality at risk. The lack of robust data infrastructure means businesses are overwhelmed by volumes of data, and the discussion began with a deep dive into the problems plaguing data-driven organisations today. 

Panellists described the unique challenges they faced when harnessing data, focusing on data accuracy, the increasing volumes of data, and data diversification. Speakers also brought up the challenges of staying compliant with local and international regulations along with data audits and security measures. Challenges of working with artificial intelligence (AI) were also spotlighted.

The speakers spoke about technical issues such as hallucinations, bias, incorrect answers, and the burden of the increasing level of compute power required to run LLMs. They described how consumer adoption of AI products has increased exponentially, while enterprises remain wary of AI. Problems such as model accuracy, standardisation, and adaptability have stopped many organisations from actively embracing AI. 

Oracle’s solutions for data challenges were discussed, including the inbuilt security and compliance features. The organisation recently introduced ZPR - Zero Trust Packet Routing - a user-friendly solution that helps Oracle customers to create security policies in Natural Language and vocabulary, making it easier to control access to sensitive data through specific access pathways. 

Speakers also discussed Oracle’s approach to AI. Instead of businesses moving data towards AI, Oracle allows them to bring AI to their data by actively embedding models and vector capabilities within databases. This allows companies to reduce costs, for training, security, and compute.

A solution for increasing cloud costs

The discussion spotlighted an interesting shift in the cloud landscape - companies opting for on premises over cloud computing. Panellists noted that cloud consumers initially adopted the cloud for the services. However, the cost of cloud computing increased exponentially as companies began to scale and they recognised the number of services they were dependent on when they examined their architecture. This led to arbitrary costs that nearly doubled the size of bills; the costs for data transfer between nodes or data processing charges were also huge. Escalating costs led companies to set up GPU towers in their own premises, use local LLMs, and run inferences by themselves - all to reduce escalating costs.

On the other hand, panellists also acknowledged the advantages of working with a cloud provider, particularly on matters of compliance. An on-premise approach does lower costs, but can also increase the need for manpower to manage resources. Additionally, companies that build systems on premises also have to deal with the need for additional security controls while managing fiscal control, fiscal security, compliance, audits, reports, and more.

A third option – a multi-cloud strategy – also came up. Oracle Cloud Infrastructure (OCI) features Oracle databases running on Google, Microsoft Azure, and AWS. The multi-cloud approach allows companies to have more control over costs, with the same benefits of on-premise systems. 

Data as a culture

A data-driven culture is one where the workforce uses analytics and statistics to optimise processes and accomplish tasks. Team members and company leaders collect information to get insights into the impact of their decisions before implementing new policies or making significant changes in the workplace.

Panellists said that like Platform as a Service, Data as a Service and Software as Service, ‘Data as a Culture’ should also be a priority for businesses. A data-driven culture should be driven by the top leadership and data should be made available for employees to take meaningful action.

Speakers stressed the importance of insight-driven decisions over intuition-based decisions, and discussed the role of data scientists. Instead of building models, data scientists must take the time to understand and measure data. The discussion also covered the ‘inter-departmental digital divide’ in organisations. As companies grow, inter-departmental communication must be digitised.

The panellists said data needs to be digital-first, even for internal complaints, warrants, tickets or any other forms of communication. Establishing these kinds of rules would emphasise the importance of data across the organisation, they said. The discussion covered a wide range of subjects concerning data, cloud computing, AI and data migration.

Oracle shared a range of services to address data compliance, accuracy and security challenges. It also highlighted its approach to data management, promoting data lake services to reduce costs and improve efficiency.