Kumar Subramanyan, Senior Vice President, CPG Products, Blackstraw AI
Kumar Subramanyan is an experienced R&D leader in Consumer Packaged Goods (CPG), with extensive experience in Data Science and Digital Transformation, complemented with robust expertise in product innovation for the Personal Care and Beauty sector. Demonstrated proficiency in leveraging cutting-edge technologies that drive organizational growth and efficiency. With a proven track-record of spearheading ground-breaking solution design, leading cross-functional global teams, fostering collaborative partnerships and implementing organizational change at scale, he is widely recognized for strong commitment to teamwork, inclusivity, entrepreneurial leadership, and strategic vision.
Drawing on recent findings from MIT and Wharton, the author unpacks why the extremely high rates of failure as well as satisfaction among AI adopters are both true and what that contradiction reveals about how organizations are fundamentally misdeploying this technology
By Kumar Subramanyan, Senior Vice President, CPG Products, Blackstraw AI
Most companies are using artificial intelligence to optimize old processes. The real opportunity lies in reinvention. The artificial intelligence boom has created a curious paradox in corporate America. An MIT study from August 2025 delivered sobering news: over 95% of AI projects are not creating meaningful value. Yet just two months later, a Wharton report told a different story 82% of business leaders are actively using AI, with three-quarters reporting solid returns on investment. Both findings are accurate, and the disconnect between them reveals something crucial about how companies are deploying this technology.
Most organizations are using AI to automate routine tasks, speed up workflows, and trim costs by modest percentages. These improvements are real and measurable, which makes them easy to justify to boards and shareholders. But they are also leaving AI’s transformative potential largely untapped. Put simply, companies are using AI to build faster horse-drawn carriages when they should be inventing the automobile.
The Incremental Gains Trap
Customer service is one of the most popular areas where companies deploy AI today, and it illustrates the problem clearly. Chatbots handle basic questions, route complex issues to humans, and give agents faster access to information. Companies love them because the results are obvious: support costs drop 20%, and response times improve. This is optimization: making the current way of doing business 10-20% better. There is nothing wrong with that, except that it falls well short of transformation.
Real transformation looks completely different. In a truly reimagined customer service model, AI analyzes millions of interactions and spots problems before they happen, proactively reaching out with solutions before customers even notice something is wrong. For everyday issues, AI agents with full memory of a customer’s history resolve over 90% of cases entirely on their own, not just suggesting fixes but actually executing them across internal systems. Human agents, meanwhile, focus exclusively on genuinely complex situations, supported by AI co-pilots that feed real-time guidance. This is not about shaving percentage points off a cost line. It is about creating customer experiences that competitors simply cannot match.
What Transformation Actually Looks Like
The gap between optimization and transformation plays out across every industry. In manufacturing, companies now use AI to predict equipment failures hours or days in advance, allowing maintenance teams to fix problems before production lines go down a meaningful step forward compared to waiting for breakdowns. But transformative AI goes further still. Digital twins of equipment continuously simulate failure scenarios, and when the system detects an issue, it automatically schedules maintenance, optimizes resources across the facility, and generates specific repair instructions. Technicians receive real-time augmented reality guidance that troubleshoots complications on the fly and learns from every repair. The goal is not fewer breakdowns, it is zero unplanned downtime through an entirely different approach to maintenance.
Retail follows the same pattern. AI currently helps shoppers find products they will like, sometimes with virtual try-ons, a genuine improvement for conversion rates. But the transformative version of this is far more ambitious: what if AI did not just recommend existing products, but created entirely new ones? Imagine a fashion company using AI to analyze emerging trends alongside a customer’s individual shopping history, designing a unique garment for that person, and triggering a micro-production run. That is not optimization. That is a new paradigm for both supply chains and customer relationships.
Why Executives Choose the Safe Path
The preference for incremental approaches is entirely rational. Optimization projects offer clear metrics, predictable timelines, and minimal risk. Walking into a board meeting with “we improved efficiency by 15%” is a safe conversation; no one questions it. Transformation, by contrast, means investing in ideas that might not pan out, tolerating uncertainty, and asking people to rethink how they work fundamentally. For organizations focused on hitting quarterly targets, that is a difficult case to make.
Organizational structure compounds the problem. Most companies run their AI efforts in silos, with marketing optimizing targeting, operations streamlining logistics, and customer service automating responses. Each group records wins and moves on. But in doing so, they miss the bigger picture. Real transformation requires stepping back to look across functions and identify where reimagining entire workflows can create disproportionate value. More often than not, the organizational barriers turn out to be harder than the technical ones.
Three Shifts That Enable Breakthrough Innovation
Moving from optimization to genuine transformation requires three fundamental shifts in thinking.
The first is changing the questions you ask. Stop asking “how can AI improve this process?” and start asking “what becomes possible with AI that we couldn’t do before?” Instead of “how can AI speed up customer onboarding,” ask whether AI could eliminate the need for traditional onboarding. Instead of “how do we optimize our inventory,” ask whether AI could enable a completely different supply chain model. This kind of thinking requires cross-functional teams willing to reimagine processes from scratch, unconstrained by how things have always been done.
The second is being ruthlessly selective. Attempting to transform everything at once is a reliable recipe for failure. The goal is to identify the specific areas where reinvention aligns with your strategic priorities and creates real, lasting value for customers. Some processes are genuinely fine as they are; a 15% improvement is enough. Others represent opportunities to do something competitors literally cannot replicate. The discipline lies in knowing which is which.
The third is recognizing that AI alone will not cut it. Real transformation means integrating AI with other technologies, IoT sensors, augmented reality, advanced analytics, and whatever else the situation demands. It also means building partnerships across your value chain, because no single company can reinvent complex processes in isolation. These relationships require investment before you have fully figured out what you are building together. That is uncomfortable, but it is how breakthrough innovation actually happens.
The Survival Imperative
The gap between optimization and transformation is no longer just a competitive issue; it is increasingly a question of survival. Companies using AI to polish existing processes will find themselves outmaneuvered by competitors who use it to create entirely new business models. The decision to adopt AI has already been made across corporate America. What remains undecided is whether organizations will deploy it for incremental gains or fundamental reinvention.
As AI capabilities continue to accelerate, the window for this kind of bold transformation will not stay open indefinitely. The companies making decisive moves now will not simply operate more efficiently than their competitors; they will be playing an entirely different game. The question is whether your organization will be among them.
