It is truly amazing to see the evolution of Artificial Intelligence (AI) in our world today. Many companies are busy generating insights into the fabric of their business decision making. There is a keen realization of its applicability and impact, with the belief that Al will evolve to be a platform, one akin to mobility, that will be available ubiquitously. It will be just there – we will use it and we’ll have it everywhere. For e.g. no one says today we use Internet. It’s a given. And just like Internet disruption created companies like Facebook, Alibaba, Amazon, etc. AI will usher in new ones.
But there also a realization within the C-suite that this technology needs to be harnessed methodically and yet with agility. Literally EVERY C-Suite meeting that I partake in, I’m asked for a point of view on what they should be doing to be ready for AI and GenAI. Amidst all the hype and noise on this topic, if I were to tune into the signal, there are three areas that are important for every company to avail of their data/AI investments to create a true competitive advantage:
- Start with Value to prioritize your use cases
- No AI without IA – Information Architecture is key
- Talent and Data culture for Human-In-The-Loop ‘Hybrid Intelligence’
1. Value-Obsession: A match made in AI heaven
One thing is clear: while Gen AI 1.0 demonstrated ‘potential’ and piqued interest, Gen AI 2.0 is rapidly transitioning into actionable outcomes and tangible value. To build a data-driven organization and harness AI as a competitive advantage, companies must begin with a focus on Value. As one of my clients humorously remarked, “We have more AI pilots than Singapore Airlines, yet we’re still lacking enterprise impact”. Like any technological innovation, the success of AI hinges on anchoring its potential in business realities – understanding how this capability can create a competitive edge, where and when to invest, how to prioritize initiatives, and determining the right balance between urgency and thoughtful planning. Once this strategic foundation is established, companies can pursue MVP experiments to validate AI’s potential in specific use cases. This experimental phase represents where many organizations find themselves today. Once initial success is achieved, the next challenge is about scaling. Scaling requires careful investment in talent, infrastructure, and planning to ensure sustainable growth and enterprise-wide impact.
Leaders often grapple with the question of whether to wait or take immediate action. The answer lies in building a degree of readiness – establishing foundational capabilities in talent, infrastructure, and applications – to avoid being unprepared. Waiting too long risks irrelevance, while an overly aggressive approach can lead to misaligned efforts in the face of rapidly evolving technology (no one can beat the R&D budgets of FAANG). The optimal approach is to prepare a solid core and focus on low-risk use cases that demonstrate value. And the main value drivers for this capability remain same as in any technology strategy:
I. Revenue Growth: New Customers & Channels, New Products, Monetization Models
AI enables personalized customer experiences by predicting preferences, proactively addressing issues, and enhancing interactions. Solutions such as chatbots and recommendation engines boost customer engagement, fostering loyalty and driving revenue expansion across diverse opportunities.
II. Operational Efficiency: Workforce Augmentation, Asset Utilization, Process Efficiency
AI-driven automation streamlines routine processes, reducing costs while improving productivity. By optimizing workflows, addressing bottlenecks, and maximizing resource utilization, organizations can achieve greater operational excellence and scalability.
III. Risk Mitigation
AI-powered algorithms identify anomalies, assess potential risks, and prevent fraudulent activities. These advanced capabilities safeguard organizations against emerging threats, enhancing resilience and ensuring robust operational security.
As an example in retail banking, Generative AI has the potential to deliver huge value along all functions:
- Sales and Marketing – Automated email campaigns, personalized marketing, etc.
- Customer Engagement – Gamified financial education, virtual financial assistants
- Risk Management and Compliance – Fraud detection / financial crime investigation, digital onboarding for underwriting & KYC
- Technology – Code generation, code review, knowledge management and enterprise search
- Enterprise Functions like HR and Legal – contract generation and review, proposal creation, etc.
These AI use cases should be scored across viability and feasibility criteria to identify top use cases based on their alignment to strategic priorities:
- Value Viability: Is the AI solution aligned with company’s strategic priorities? Will the AI solution improve customer experience by improving engagement, information access, etc.? (Revenue Uplift, Cost Optimization, Risk Reduction, Customer Engagement/ Satisfaction)
- Complexity: Does the company have the right data, operations, and technology to implement the AI solution? (Data availability & readiness, technology maturity – skills & platforms, Compute requirements, etc.)
But the key is that companies don’t need chase the shiny objects only in AI/GenAI. I was recently working with a global retail bank – their new CMO and CEO agreed that their data layer was surely the first layer to build their digital foundation. We mapped their key customer journeys for the targeted personas to their value chain. The typical journey involved onboarding, credit check, approval, risk monitoring, etc. for their different products (deposits, loans, insurance, mortgages, wealth management, etc.). We pulled in key stakeholders from all parts of the value chain and built a shareholder value tree of drivers and use cases mapped to these drivers. We assigned some simple weights to help prioritize the same. One use case that popped up was the “Alerts/Notifications” to users when their balance was about to dip below a threshold. Cash flow understanding is a simple mechanism in AI – when a person typically gets her salary direct deposit, when the mortgage is paid out, when the car loan payment goes out, etc.. Customers always appreciate avoiding penalty fees. That’s when they begin to trust the “intentions” of their bank. Most customers trust banks in “execution” – when I go to the ATM, I’ll get cash and the accounts will be correctly balanced. But is the bank ‘really watching for my interests’?
2. No AI without IA
AI systems thrive on vast amounts of data. However, the mere presence of data is insufficient; it must be organized, accessible, and meaningful. This is where Information Architecture (IA) comes into play. By structuring data effectively, IA ensures that the models can access and process information seamlessly, leading to more accurate analytics and insights. The structured organization of data and information serves as the bedrock upon which robust analytics and AI systems need to be built. Without IA, AI initiatives risk inefficiency, inaccuracies, and potential failure. Companies are challenged with integrating their internal data from relational databases (RDBMS), external unstructured sources, and large language models (LLMs) into a cohesive information architecture: data integration and scalability, semantic alignment for context/ interpretation, and optimizing performance for latency issues. Companies are using many new frameworks for this. For e.g. Apache UIMA (Unstructured Information Management Architecture), an open-source framework that facilitates the analysis of unstructured information, enabling the integration of diverse data sources into a cohesive structure suitable for AI processing.
Start by standardizing structured data (e.g., tables, databases) into formats compatible with unstructured data processing. Use tools like natural language processing (NLP) to extract and encode insights from unstructured sources (e.g., text, images). Implement a centralized repository, such as a vector database, to store unified embeddings and metadata. Leverage LLMs fine-tuned for your use cases to interpret and query data holistically. Ensure your architecture supports APIs and middleware for easy data access. Continuously evaluate and refine to enhance data accessibility, relevance, and decision-making capabilities. Of course, this surge in AI utilization is placing unprecedented demands on data centers and network infrastructures – a necessity for scalable and efficient IA to manage and process the increasing data loads effectively.
3. Data Literacy and Culture: Human-in-the-Loop (HITL) for ‘Hybrid Intelligence’
In the race to adopt artificial intelligence, organizations often overlook two critical components: data literacy and a strong data-driven culture. Data literacy – the ability to read, work with, and communicate data – empowers employees to collaborate effectively with the AI systems. Without this skill, even the most advanced AI tools risk being underutilized. Equally important is fostering a data-driven culture. This involves instilling trust in data, encouraging cross-functional collaboration, and ensuring leadership champions the strategic use of data assets. Many boards are looking at the value of concepts like Reverse Mentoring from younger data literate professionals. Specially post-covid, a key trend is the rise of hybrid and remote work, which has prompted the adoption of collaborative tools such as Slack, Microsoft Teams, Asana, Workforce, etc. These platforms now feature enhanced AI capabilities, offering predictive task management, automated scheduling, and data-driven insights. These tools streamline repetitive tasks, automate data processing, and enable creative problem-solving, freeing up employees to focus on higher-value work. So then people ask the ‘adoption’ question – can this can run on its own or will a human loop always be needed? The answer will depend on what this is used for. One important human skill that will always be needed – deduction and inference – “asking the right questions” a.k.a. the Socratic Method. But with recent developments, AI is seeking to mimic the human brain. Artificial General Intelligence (AGI) is an area where AI is not constrained only by supervised learning but exploring new patterns altogether. This new inflection point is due to the convergence of 3 things: natural language processing, large amounts of training data with advanced transformer models, and huge compute power. This has started to create the beginnings of human-like cognition. Here “Intelligence” implies thinking, act, judging – an inorganic thing doing what humans do. And just like human cognition follows these steps: we sense, we decide, we create, we learn, and we remember the past (mostly that is :-)). One of my client’s Chief AI Officer mentioned “We are hiring folks with humanities background. These skills have to be at the intersection of business domain, data science, and human cognitive skills”.
The era of fully autonomous AI agents has not yet arrived; human judgment remains essential in many workflows. For instance, in financial services, fraud detection is a very important function. AI can flag unusual transactions for potential fraud. However, human analysts are needed to validate these alerts, distinguishing between genuine threats and false positives. This prevents unnecessary account freezes, improving the customer experience while maintaining robust security. Ultimately, ‘hybrid intelligence’ is about leveraging the strengths of both humans and machines. It’s not a competition but a partnership – one that maximizes the potential of technology while ensuring that decisions remain ethical, responsible, and context-aware. One of my clients put it in perspective well: “Think of AI / GenAI technology as an inorganic cognitive augment. Treat them as an intern that you can ask questions. But your experience is still key to ask the right questions and apply your cognitive judgement”.
Summary
In a nutshell, three key strategies for accelerating AI innovation: prioritize value-driven use cases, establish robust information architecture, and foster a culture of hybrid intelligence through data literacy.
Clearly AI requires a shift in leadership mindsets and behaviors – leadership needs to embrace this and prepare the workforce for the digital future by reskilling and upskilling talent. And another very important factor is that even after all this reskilling or training, the applications with embedded AI need to be made simple to use (employee experience for real customer experience).
 in our world today. Many companies are busy generating...){kind=link}