Scaling the AI ladder gets easier with trusted AI backed by the right data architecture

Soma Tah

Amid the accelerated pace of digital transformation during the pandemic, many organizations in India turned to artificial intelligence (AI) to address the exposed flaws in business continuity and become resilient. From absorbing spikes in customer-service volumes to recalibrating broken or uncertain supply chains and maintain operational flexibility, AI came in handy for businesses all the time. Over half (53%) of Indian IT professionals said their companies have accelerated the rollout of AI due to the COVID-19 pandemic, revealed the Global AI Adoption Index 2021 by IBM. 

India is uniquely positioned to succeed in the AI economy, says Viswanath Ramaswamy, Vice President, Technology, IBM Technology Sales, India/South Asia. AI is now becoming part of business functions such as hiring, supply chains and customer service. While industries such as BFSI, Retail, and CPG have been at the forefront of AI adoption, other sectors like Telecom, Manufacturing are also not far behind. But the reality is that the struggle does not end after the adoption of AI. As AI moves on from experimentation to adoption fast, it makes businesses revisit their AI adoption strategies. Scaling the AI investments and maximize the returns from them also gets challenging. 

Ramaswamy tells us how IBM is handholding Indian customers in their AI maturity journey and help them unlock the full potential of AI for their businesses with trusted AI.

Excerpts:

As the first-mover advantage wanes gradually, how are businesses revisiting their AI adoption strategies nowadays? 

As business leaders navigate a succession of pandemic-fuelled transformation imperatives, AI continues to enable organizations to address urgent and immediate business priorities—quickly and at scale. As per IBM IBV CEO Study, in India, 45% of CEOs expect AI/ML to deliver results in 2-3 years. Organizations are integrating AI in their digital transformation journey to enhance client value, bring in high-impact use cases, and to bring in a differentiated approach vis-à-vis their competitors. An interesting observation is that Indian organizations are investing in capabilities that are necessary preconditions of AI: such as strengthening data and process management, and human capital leading to speeding the progress to more advanced adoptions and time to value. 

For organizations, it’s clear that an enterprise’s path toward greater maturity is through the AI journey – and this can begin with tangible steps. For example, deploying an AI-enabled virtual agent or re-optimizing demand forecasting with an intelligent recommendation engine can quickly achieve positive returns.

For example, ICICI Prudential Life Insurance has introduced a virtual assistant powered by IBM Watson to converse with policyholders addressing their queries and providing personalized account-specific information. Using IBM Watson, ICICI Prudential Life has also created an email automation solution that classifies the user requirement and maps an appropriate response accordingly.  

Similarly, an AI-enabled mobile application powered Emerald Jewellery's B2B business in India amidst COVID-19. The collaboration with IBM has enabled Emerald Jewellery to redesign the buying process for their 200+ dealers, from 500,000+ jewellery design selection to placing a purchase order, using a zero-touch engagement model.

Why do businesses invest in AI solutions- for improving efficiency or for creating a competitive differentiation? 

I would say it is both. As AI moves beyond experimentation and towards adoption, it is now widely adopted as a key lever of competitive advantage and profitability across businesses and industries. AI improves a company’s cost base—augmenting human capability to motivate greater efficiencies and helps enhance or protect top-line revenue, experience and engagement. 

Many use cases emerged in the last couple of years. For example, IBM is helping Parle with an intelligent AI-based supply chain. Using IBM Watson Studio, we are helping them design a critical KPIs-based performance management framework to measure and monitor the health business processes and their outcomes. Intelligent workflows help Parle drive its agenda on the appropriate product mix, faster planning and scheduling, and optimized supply chain costs.    

Similarly, Bestseller, which owns Vero Moda and other key fashion brands, uses AI to predict the right merchandise for the consumer at the right time. With IBM Watson mining deeply into big data, the retailer can determine the right assortment plan for each store, predict the next best product to incorporate into its mix, and improve the efficiency of its supply chain. They are working with IBM Watson to predict the next big trend and the most relevant styles, colors, and size ratios. Higher relevancy means a sharper, better-selling assortment, helping them meet consumer expectations while becoming more efficient and competitive in the market. 

With its value-expanding impact on both sides of the financial equation, the essentiality of AI is clear, both in and before the current crisis.

As the discussion continues to heat up with AI ethics and AI governance issues, are businesses ready to rely on AI-led outcomes? 

For AI to help our work and improve our lives, it must respect our data and the insights about us, and it must be transparent and explainable. Today, organizations recognize the importance of a holistic approach to building quality and governed data platforms to build on an AI strategy. 

The Global AI Adoption Index 2021 found that nearly all (95%) IT professionals in India believe that it is critical or very important to their businesses that they can trust the AI’s output is fair, safe and reliable. Further, 3 in 4 (78%) Indian IT professionals think it is very important to their companies that they can build and run their AI projects wherever the data resides.

We believe trust is essential to AI adoption. It allows organizations to understand and explain recommendations and outcomes and manage AI-led decisions in their business (which can be critical for regulatory requirements, among other things)—while maintaining the full ownership and protection of data and insights.

Drawing on our expertise of more than 40,000 IBM Watson client engagements globally, we are helping build trustworthy AI services on hybrid clouds for a wide range of industries. IBM’s scaled data science methods and governance frameworks help organizations to cultivate responsible AI stewardship based on their unique values and regulatory requirements. We have launched tools like AI Fairness 360 and AI Open Scale to help organizations monitor their AI operations and mitigate bias.

What challenges do businesses face as they look to scale AI and maximize the returns from their AI investments?

As per the Global AI Adoption Index 2021, Indian IT professionals are most likely to see limited expertise or knowledge (52%) and increasing data complexity and data silos (50%) as barriers to AI scaling and adoption. While recent advances in the technology are making AI more accessible than ever, companies are facing a challenge in realizing the full potential of AI: unifying complex, siloed data across decentralized IT infrastructures, ensuring security and compliance of the data and technology, building trust in AI systems and outcomes, and developing a clear strategy with the appropriate skills and cross-functional teams to scale and support AI to meet business needs. But with the right approach, architecture, and capabilities, businesses can overcome these challenges to achieve sustainable solutions that span the organization and deliver more value from AI investments faster.

IBM is committed to lowering the barriers to entry and making AI more accessible to businesses by offering an architecture (AI Ladder) based on client insights. The four key areas clients must focus to implement AI capabilities successfully are: Collect Data, Organize Data, Analyze Data and then Infuse AI. 

IBM Watson provides businesses with a range of AI software and services, such as tools to help build AI i.e. APIs for developers and tools for data scientists to build and deploy custom models as, AI applications targeted at self-service, risk and compliance, and financial and IT operations that can run anywhere on the Cloud Pak for Data. IBM's Data Science Elite Team tackles the skills-shortage in data science and AI, by assisting clients in their AI and digital transformations.

How is IBM honing its AI capabilities in terms of NLP, Deep learning, computer visions, etc. ? What are some of the newly added capabilities in Watson?

We are witnessing an uptake in Natural Language Processing (NLP) technologies, Automation and Trust to help lower costs and get trusted information into the hands of their customers, partners and employees in a distributed work environment.  

NLP: Chat and Voice-based bots are growing tremendously. AI that understands the language of business spanning human language, documents, contextuality is essential. But, unlike numbers and images, language varies from country to country. Additionally, most large enterprises are global and must operate in numerous languages. As a result, enterprise NLP solutions must work in many languages and ideally without the need to undergo training for every language individually. IBM Watson Discovery now includes support for 10 new languages, including Hindi.  Some of our products: IBM Watson Natural Language Understanding, IBM Watson Assistant, IBM Watson Text-to-Speech/Speech-to-Text, IBM Watson Discovery

Automation: Automating data collection and sorting work is critical to facilitating the deployment of AI. Automation and lifecycle management tools, in combination with chip and system-level advances, are essential to scale AI. This can be implemented across business processes and IT. IBM Cloud Pak for Watson AIOps uses AI to simplify IT operations management and accelerate and automate problem resolution in complex modern IT environments. Bharati Airtel has been working with IBM to integrate advanced end-to-end automation and plans to embed AI capabilities in the future as a core part of its network transformation. IBM and Airtel have co-developed a “single-click” automated hybrid cloud design and deployment capability and “light touch” operations. 

Trust: Principles for Trust and Transparency that we follow to build and strengthen trust in technology. They make clear that:

-The purpose of AI is to augment - not replace - human expertise, judgement, and decision-making;

-Data and insights generated from data belong to their creator, not their IT partner;

-Powerful new technologies like AI must be transparent, explainable, and free of harmful and inappropriate bias for society to trust them.

Our engagement on these principles not only shapes our business practices, it also drives our participation in the broader societal dialogue around trust in technology.

What are some of the next-gen hardware capabilities that IBM is working on to unlock the full potential of AI for businesses? 

AI demands substantial computing resources and storage space. Hence, setting up any application of AI needs a clear infrastructure agenda. As businesses turning their attention to training their AI models to draw faster inferences with AI applications, they require robust and scalable IT Infrastructure to feed in the AI data for faster throughput and better accuracy. For AI infrastructure, businesses need to rethink the ways they are managing their data also. 

To help businesses in this transition, IBM has introduced a concept called "AI Infrastructure," which consists of a single platform for data pipelines, services, and AI. Essentially it is an end-to-end server, storage, and software platform designed for businesses to stand up their version of the AI revolution. Fundamentally, the AI infrastructure can scale the storage as the volume of data grows. The next-gen IBM Power processors with an embedded math accelerator will help businesses derive much faster AI inference at the data source. The same processor chip gets used for both traditional enterprise workloads and modern-day AI applications.