AI has taught me that the journey from hype to reality is less about magic and more about discipline, data, and persistence. What once promised instant transformation now demands careful groundwork and collective learning.
Chief Data Officer of RPG Group
When I look back at the past 15 months of living with AI at RPG Group, I can’t help but reflect on the gap between promise and reality. Fifteen months ago, we were told I would write novels in the morning and solve world hunger by afternoon. Today, what I mostly see is automation of emails, creation of meeting minutes, or the generation of summaries. The journey hasn’t been the joyride that was promised; at times, it has felt more like a horror show.
From Hype to Hard Truths
In those early days, AI was painted as a superhuman force. What I discovered instead is that adopting AI is a messy, complex, and far from plug-and-play process. Information overload, shifting technologies, and constant new releases mean that just as you begin to understand one tool, another arrives. While there is undeniable progress, most enterprises, including ours, continue to struggle with integration, data readiness, and scaling challenges.
Why “Me Too” AI Doesn’t Work
One of my strongest learnings has been the futility of simply chasing trends. Too often, I’ve seen existing solutions relabelled “AI” to gain traction. But fundamental transformation doesn’t come from changing the name. It comes from addressing foundational gaps in data availability, security, and ecosystem maturity. Without that groundwork, projects risk becoming pilots that never deliver meaningful business value.
Data: AI’s Hungry Engine
What has become crystal clear to me is this: AI is not starved of algorithms; it is starved of data. I often describe AI as a starving child with a heavy diet. It constantly demands clean, contextual, and timely inputs. Unless we can ensure data governance, privacy safeguards, and domain expertise, AI will fail to deliver its full potential. And while we discuss automation, we must also remember that people are at the center of it all. AI can’t run without people who know how to run AI.
Counting the True Cost
The other reality check I’ve faced is around cost. AI is expensive, not only in terms of cloud and API usage but also in terms of talent. Specialists are scarce, and their price is rising. And then there is the environmental dimension, which few talk about. A single generative AI query can consume as much carbon as a month of driving. For organizations like ours, deeply committed to ESG principles, this is a risk we can’t ignore.
My Playbook: Awareness, Discipline, and Scale
Despite these pitfalls, I remain optimistic. But optimism must be grounded in discipline. The approach that works for us begins with raising awareness among cross teams and business users about what AI can and cannot do. Not every workflow automation problem is an AI problem. Not every integration issue requires machine learning.
I frame AI in three layers: everyday AI, business operations AI, and game-changing AI. Each needs to be evaluated for feasibility, impact, and scalability. And one principle guides me always: just because you can doesn’t mean you should. AI, I often say, is like shaadi ka laddoo. Whether you eat it or not, you’ll regret it. But you can’t ignore it either.
A Journey, Not a Destination
For me, AI adoption has never been about a single project. It’s a continuous loop of experimenting, learning, and scaling. It’s about empowering cross-functional teams, validating use cases, and ensuring that proofs of concept are designed with scale in mind.
As I reflect on this journey, I’m reminded of the words of poet Majrooh Sultanpuri: “Main akela hi chala tha, jaanib-e-manzil magar, log saath aate gaye aur karwan banta gaya.” I began this journey alone, but along the way, others joined, and together we built a movement.
That is what AI feels like to me today: not a magic wand, not an instant solution, but a shared journey of discipline, persistence, and possibility.