Sunil Senan on harnessing AI and Data for enterprise transformation

In this exclusive interview, Sunil Senan, Senior Vice President and Business Head of Data Analytics and AI at Infosys shares his perspectives on how organizations can harness the power of data and AI to drive business transformation. From modernizing enterprise cores and optimizing data for AI applications to overcoming legacy challenges and capitalizing on AI-driven innovation, Sunil discusses the strategic importance of cloud, data governance, and developing digital skills to thrive in the AI-first era. His responses highlight the key role of responsible AI, data readiness, and a robust digital foundation in shaping the future of enterprises.

Excerpts:

DQ: How can organizations effectively modernize their enterprise core with data and AI to stay competitive in the digital age?

Sunil: Data and AI can revolutionize core business functions. By leveraging these technologies, businesses can “responsibly” leverage Data & AI for accelerated growth, unlock efficiencies at scale, and build new ecosystems. Organizations can modernize their core functions by streamlining operations, improving decision-making, enhancing customer experiences, and gaining a competitive edge. Human resources can optimize talent acquisition and retention, finance can detect fraud and forecast performance, sales and marketing can personalize campaigns and optimize lead generation, operations can optimize supply chain and quality control, IT can strengthen cybersecurity and automate tasks, and provide exceptional customer services.

However, navigating the path to modernizing the enterprise core to realize value-at-scale is not linear. It demands a holistic approach by developing a comprehensive data and AI strategy, establishing a strong & responsible data foundation, building a modular infrastructure that enable adaptable & scalable AI product development, and fostering a collaborative workforce.

Also, by addressing data organization, fingerprinting, access control, and security measures, organizations can create a solid foundation for AI initiatives, accelerating development, improving model accuracy, and ensuring data privacy and reliability. This enables enterprises to harness the full potential of Data and AI to drive innovation, improve decision-making, and enhance customer experiences.

DQ: What role does data play in driving AI initiatives, and how can companies ensure their data is optimized for AI applications?

Sunil: Data is the fuel that powers any AI initiative. Without high-quality, relevant data, AI systems would be severely limited in their capabilities. Data is used for training AI models, feature engineering, model evaluation, and real-time decision-making. It plays a crucial role in explaining how AI models arrive at their conclusions, enhancing transparency and trust. Data is used to evaluate the model's performance and identify areas for improvement. Its quality, diversity, and ethical handling are essential for building effective, reliable, and beneficial AI solutions.

Infosys research highlights “Being Data Ready” as a prime challenge in getting “Enterprises AI Ready”. Given the volume and variety of data, it is often very challenging for enterprises to assess the value of the data & possibilities their data and AI can drive for their business. The biggest challenges enterprises face today include an unclear business strategy for AI, an unorganized data landscape & missing data governance, a lack of trust in AI control, limited AI skills and proclivity, enterprise culture and people, talent and skill shortage, etc.

To optimize data for AI and to enable autonomous operations, organizations must address key considerations such as contextual metadata, data integration, and robust governance. By providing AI systems with rich metadata such as industry context, sensitivity, entitlements, provenance, and freshness, organizations can enhance the effectiveness of AI models and decision-making processes. Additionally, integrating diverse data sources and establishing effective governance frameworks are essential for ensuring trust, ethics, quality, and compliance. In the era of generative AI, effectively connecting, harvesting, and correlating information from all data sources is fundamental for enterprises seeking to drive substantial business value at scale.

DQ: How can organizations overcome challenges like legacy technology and outdated processes while implementing AI-driven data strategies?

Sunil: Legacy technology and outdated processes can hinder the successful implementation of AI-driven data strategies. Heterogenous infrastructure, technology debt, outdated hardware, complex business processes with significant human interventions, lack of end-to-end business process visibility, and insufficient data from these processes are some of the common obstacles. Moreover, manual processes, a lack of data-driven culture, data quality issues, and limited data access can further impede the effective utilization of data and AI.

As enterprises identify high-value AI use-cases, to overcome challenges and harness the full potential of AI, organizations should adopt an AI-first transformation approach built on a cloud foundation. This approach establishes a resilient and secure infrastructure, streamlines business processes, and enhances data visibility, enabling effective scaling of AI initiatives. To ensure success, enterprises must prioritize data governance, privacy, and security, fostering a data-driven culture, and leveraging data fingerprinting techniques. A data-driven culture empowers employees to make informed decisions based on data insights, while data fingerprinting techniques enable organizations to untangle complex data ecosystems and optimize data management. These measures are crucial for overcoming legacy challenges and empowering organizations to harness the full potential of AI, driving innovation and achieving competitive advantages.

DQ: Can you elaborate on the convergence of cloud, data, and AI, and how this trifecta delivers cognitive solutions that drive business transformation?

Sunil: AI-first transformation powered by generative AI, advanced analytics, and a cloud backbone is revolutionizing businesses in unprecedented ways. Cloud platforms provide the scalable infrastructure for AI initiatives, while data serves as the fuel that powers these intelligent solutions. The synergy between cloud and AI enables organizations to harness the vast amounts of data stored in the cloud, unlocking valuable insights and driving innovation.

In the enterprise context, AI’s massive calls on computing, data gravity, storage, security, and networking for delivering real-time insights, need the cloud not just to pivot, rapidly experiment, and pilot transformative programs but to scale AI enterprise-wide as well. In addition, the complexity of gen AI mandates implementation through connected enterprise applications and scalable enterprise cloud platforms, connected both within and outside the enterprise. Disconnected pilots and initiatives run by individual development teams can prove counterproductive. Enterprises will need to deploy ‘FinOps for GenAI/AI’ strategies to maximize value and drive cost efficiencies for GenAI/AI initiatives, as Traditional FinOps is not equipped to handle Data & AI workloads. It needs to be seamlessly incorporated and integrated into the AI value management process.

To maximize ROI from cloud, data, and AI, companies must focus on high-value business cases and build a robust cloud foundation optimized for AI workloads. By preparing their data for AI and scaling innovations effectively, organizations can unlock the full potential of these technologies. Common pitfalls that hinder cloud value include unrealized use cases, cloud sprawl, and stalled adoption.

By overcoming these challenges, organizations can leverage data and AI to:

• Create disruptive business models: Develop innovative products and services.
• Transcend industry boundaries: Expand into new markets and seize new opportunities.
• Accelerate transformation: Drive digital transformation and achieve competitive advantages.
• Reduce costs and Improve productivity: Streamline operations and unlock efficiencies at Scale

The convergence of cloud, data, and AI services offers unprecedented opportunities for enterprises to accelerate their transformation and embrace an AI-first future.

DQ: What best practices should organizations follow to ensure the quality and governance of their data when deploying AI models?

Sunil: As organizations evolve from being data-driven to embracing an AI-first strategy, data continues to be the essential fuel for their AI engines, making its quality & governance more crucial than ever. Governing data in the AI world comes with its own set of challenges such as lack of trusted sourcing, anomalies & bias in data, the presence of large unlabeled structured and unstructured data along with aggravated privacy & security concerns.

To manage these challenges organizations should follow below mentioned best practices -

• Data & AI COE – Enterprises should establish a dedicated Data & AI Governance CoE with stakeholders from legal, privacy, business, AI model & data owners to govern, publish policies, and guidelines & manage the AI operations
• Data Fingerprinting – Organizations should fingerprint their extensive data footprint through the creation of data/model catalogs, data lineage, classification, data dictionary, etc. This approach will enhance overall governance by ensuring that data and model aspects are documented, tracked, traced, and evaluated, fostering trust in the AI systems.
• Data Quality Monitoring – Organizations must establish controls over Data & AI, including profiling, data remediation, bias management, training data labeling, and data annotations, to ensure that high-quality training, validation & test data is used by the models.
• Data Privacy & Compliance Controls – Subsequently, controls such as personal data identification, personal data protection, consent management, and compliance assessments with respect to various regulations (GDPR, EU AI Act, CPRA, etc.) including risk assessments etc. should be implemented to ensure privacy and compliance in governance.
• Data Security Controls – Create a robust security layer over the AI systems to protect data from attacks such as data poisoning, inferencing, prompt injections, etc. through context-based access, security guardrails, etc. in addition to the standard security controls
• Data Discovery & Consumption – Consumption of Data & AI through a centralized marketplace for easy discovery & responsible distribution across the organization

Infosys has been committed to driving AI-first for business growth in a responsible manner and is industry's first organization to be certified in ISO / IEC 42001:2023 (Artificial Intelligence Management System).

DQ: How can AI be leveraged to convert unstructured information into structured data, and what are the key benefits of doing so?

Sunil: With a large portion of data generated in unstructured formats like text, audio, and video, it can be challenging to extract insights. NLP, Computer Vision, and Deep Learning techniques can be used on unstructured data to bring structure. NLP techniques can process text documents, emails, social media posts etc. It can analyze the structure and semantics of sentences, extract entities and relationships, and derive insights from unstructured text data. Both supervised and unsupervised techniques can be used to bring this structure to an unstructured text. Computer Vision techniques can extract valuable insights from unstructured data sources like videos of security footage, inventory in warehouses, product image classification, and medical images to detect diseases.  With AI-powered unstructured data integration, organizations can save time, reduce costs, and make informed decisions. The organizations can uncover the enormous business value of unstructured data.

DQ: What are the most common challenges organizations face when trying to integrate AI into their data-driven decision-making processes?

Sunil: Data Analytics and AI’s fearsome elements are the potential risks associated with it. User Generated Data, Digital, Device, Unstructured data, missing governance, unclear goals, and inaccuracy & lack of trust in data-driven insights are just the tip of the iceberg. "Machine ambiguity" and data hallucinations demand ethical guardrails and robust observability. Enterprises that implement data protection, security, and privacy to merely satisfy the law demands or customer’s expectations, need to rewrite the plot. In the age of generative AI, connecting, harvesting and correlating information from all the data with privacy and security at the core is one of the critical foundational needs for enterprises to get their data ready for AI. Applying the responsible by design principles across the life cycle of the data and AI is a key competitive advantage that will eliminate distrust and risks from both cyberattacks and regulatory scrutiny.

DQ: How can businesses use AI to continuously learn from data and automate business problem resolution for improved efficiency and innovation?

Sunil: As we know there are several ways businesses can use AI to continuously learn from data or use insights from data for their decisions. Most examples are function specific and we think that is the manner in which this will progress. However, for those functions, there are instances where AI platforms or experimentation platforms are being implemented to make AI usage easier for users and builders (users use AI while builders build applications with AI).

For example, for a major electronics retailer, we deployed AI models to analyze curated consumer voice and text data, generating aspect-oriented sentiment analysis. This automated approach improved business insights, facilitated customer problem resolution and increased the accuracy of detecting nuanced emotions.

In another example, we successfully automated data retrieval leveraging an LLM-powered knowledge bot for a large home improvement retailer, accessing over a million filter combinations from thousands of databases. This resulted in increasing monthly average usage and enhancing answer accuracy.

Nonetheless, Automation applied in responding to these insights generated by AI is not being productized by most firms due to certain risks involved but over time we expect businesses to use systems that make many decisions autonomous with certain thresholds, over which they get routed for business user review.

DQ: With AI models constantly evolving, what steps should organizations take to ensure they remain agile and can quickly adapt to new advancements in AI?

Sunil: According to the AI Index report published by Stanford University, a total of 149 foundation models were released in 2023, while AI-related regulations in the US grew by 56.3% during the same period. The rapid proliferation of AI models and increasing regulations underscore the need for organizations to prioritize data readiness, responsible AI, and AI education while focusing on business value generation.

Learning from the digital revolution, the winning formula is not a template or architecture. It is a culture or a philosophy. Being nimble, modular, designing for scale but bits at a time, and above all being value driven. Not an AI for AI’s sake approach, but rather AI as a tool to solve a problem. Tools will evolve but solutions cannot wait. Leveraging what is available to the best extent possible, and then not rushing to upgrade just because the next generation is available, rather focusing on extracting most business value from the solution already created. A COE that stays abreast of developments and provides guidance to other teams on what are the best options available. Reviewing past solutions and re-visiting the options only at reasonable intervals such as a year to let the solution start delivering business value and prove that the use case has adoption before investing further. At the review and upgrade stage, upgrade only if the cost-benefit makes sense. Similar to an automobile R&D division in their approach to upgrading past models and launching fresh models.

DQ: How can organizations ensure they have the necessary digital skills to capitalize on AI advancements and build a data-driven enterprise?

Sunil: Infosys recently released a report on the future of technology skills, which found that the areas with the largest need for training are advanced statistical analysis (8% gap), machine/deep learning (6%), and cloud computing (6%). We’ve found success with tech boot camps and online certification programs, as alternatives to traditional academic degrees, as successful tactics for upskilling. There has been a major shift in organizational recruiting and training strategies as businesses advocate for a more dynamic and continuous learning model to bridge the widening skills gap.

To keep up with the shifting landscape, businesses are moving away from hiring or training workers based on expertise in a single technology. Instead, they are seeking talent proficiency across multiple disciplines. Also, companies should start looking for intangible human soft skills like intellectual curiosity, adaptability, learnability, critical thinking skills like objectivity & logical reasoning, empathy, emotional intelligence, etc., that can enhance AI collaboration.

 Building diverse talent pipelines and offering competitive opportunities to advance hard technical skills and soft communication skills are effective strategies. Valuing hands-on skills over specific degrees and being open to candidates from boot camps are meaningful strategies for businesses. Motivating employees to gain proficiency across technologies will also be imperative. Valuing adaptable policies, strategic technology investments, and workforce development allows leaders to drive organizational success. Also, companies are increasingly aware of the value of investing in “Data readiness” which will help foster a culture of responsible experimentation as part of upskilling.

Hands-on training in AI, ML, advanced analytics, and cloud technologies will be critical to ongoing success, and this highlights the need to balance skills required for basic infrastructure, such as the cloud, with those needed for emerging technologies such as AI. Companies that invest effectively in developing their workforce’s technical skills will boost innovation and harness digital transformation.