Enterprise AI is at a crucial juncture, as a significant portion remains ineffective at scale. Within the financial services sector—encompassing wealth management, insurance, and lending—organizations have made substantial investments in AI pilots, proofs of concept, and large language models. However, only a select few can demonstrate consistent, measurable improvements in revenue, customer experience, or operational efficiency.
Kamal Kishore, Chief AI and Technology Officer at Centricity WealthTech, emphasizes that the problem does not lie with the technology itself. “AI success is 20% model and 80% system design. Most organizations are still optimizing the 20%, while the real gap lies in how AI is operationalized,” Kishore stated. This critique challenges the prevailing industry focus on models and benchmarks, pointing out that the true limitation resides in the systems, processes, and operational frameworks that dictate the value AI can deliver.
The demand for technology leadership within financial services has shifted significantly. What was once about meeting delivery timelines and cost controls is now directly tied to business outcomes. Kishore remarks, “Earlier, technology was about delivering projects and managing costs. Today, boards expect technology leaders to influence revenue, growth, productivity, and speed of decision-making.” This shift is particularly evident in customer-facing functions like onboarding, advisory, and partner distribution, where organizations are increasingly pressured to provide quicker, more personalized, and seamless experiences.
At Centricity WealthTech, AI is not viewed as an experimental layer but as a key driver of performance. “AI should not exist as a capability. It should show up in business performance and measurable outcomes,” Kishore asserted. This perspective necessitates embedding AI into core workflows, allowing it to directly impact conversion rates, turnaround times, and productivity. The challenge is not just deploying use cases but effectively integrating them into daily operations. “The key difference is whether AI is treated as experiments or as a platform capability. Unless it becomes part of core workflows, it will never scale,” Kishore explained.
Organizations that have successfully scaled AI are building it as a shared platform and measuring success through business metrics rather than use case volume. Conversely, others are hindered by disconnected pilots that “exist to demonstrate possibility, not to deliver outcomes,” Kishore noted.
The industry’s fixation on models—be they GPT, Gemini, or others—may risk neglecting broader execution challenges. “A great model in a bad system delivers zero business value. Without the right architecture, even the best AI will fail to scale,” he emphasized. Within financial services, AI must function effectively in complex environments characterized by legacy systems, regulatory pressures, and high data sensitivity, making integration the real hurdle.
While many organizations focus on model investment, they often underinvest in associated ecosystems, including data quality, workflow integration, governance, and user adoption. Kishore points out, “Almost 90% of initiatives stall after one pilot. The system around the model is not designed for real business usage.” Data quality, which is central to this challenge, directly impacts AI effectiveness. “Data has always been an enabler. But without strong governance and quality controls, it becomes the biggest constraint,” he added.
With the rise of generative AI, context has emerged as a focal point. Structured data, domain expertise, and regulatory compliance are essential for producing reliable results. “What matters is how effectively you provide business context to the models. Context determines the quality of outcomes,” Kishore explained. Furthermore, governance must transition from an obstructive to an enabling role. “Governance has to be architectural, not heavy. It should enable scale, not slow it down,” he noted.
By embedding governance into system design, organizations can meet compliance requirements while scaling AI effectively. “Once the architecture is in place, it enables scaling without restricting experimentation. That is where governance becomes a growth enabler,” he stated.
AI signifies more than just a technological upgrade; it represents an operational model transformation encompassing people, processes, platforms, and partnerships. “AI is not just a technology shift. It is an operating model shift involving people, processes, platforms, and partnerships,” Kishore explained. This change necessitates a reevaluation of workflows, roles, and decision-making processes across the enterprise, potentially reshaping advisory models, underwriting processes, and customer journeys.
Initially, organizations may focus on the number of features developed, but over time, attention must shift towards the business outcomes those features facilitate. More advanced organizations are utilizing AI to enable real-time decision-making, thereby gaining a competitive edge. AI is also transforming workforce dynamics, allowing organizations to redeploy talent to higher-value tasks like innovation and customer engagement, resulting not merely in efficiency gains but also in a structural shift in work execution.
As financial services firms transition from experimentation to scale, the distinguishing factor won’t be superior models, which are increasingly accessible. The real advantage will lie in designing systems capable of operationalizing AI at scale, integrating it into workflows, and aligning it with overarching business outcomes. “The ROI comes when AI is embedded into how the business runs, not how experiments end,” Kishore stated. In an industry characterized by trust, speed, and customer experience, effective system design will ultimately differentiate organizations that successfully scale AI from those mired in pilot projects.
(With inputs from Diksha Negi)
Published On May 9, 2026 at 09:15 AM IST
