Financial services firms racing to adopt AI may be overlooking the foundational work required to make those investments pay off, according to Syed Ali, CTO of Global Financial Services at Ensono.
While many banks, insurers, and asset managers are experimenting with generative AI and agentic tools, Ali said the organizations seeing the strongest results are not necessarily the ones chasing the flashiest use cases.
That reality is creating a growing opportunity for MSPs, consultancies, and SIs. As CIOs face mounting pressure to demonstrate AI returns, many organizations need help addressing fragmented data environments, modernizing operating models, and connecting AI initiatives to measurable business outcomes.
AI success starts with data and process discipline
Ali said many financial services organizations are still struggling with fragmented data environments, inconsistent definitions, and unresolved technical debt. In one example, he described a bank with 17 different sources of truth for defining a customer.
“If you can’t define what your customer profile is, how can you expect AI to give you measurable results?” Ali said.
For organizations that have been more successful, the work often starts with what Ali called “the plumbing”: classifying data, improving metadata, consolidating silos, and ensuring that AI systems operate on reliable information.
Financial services faces higher stakes for AI mistakes
Ali said the financial services sector has a longer history with AI than many industries, with earlier applications in data mining, classification, regression, fraud detection, and risk analysis.
But generative AI and agentic systems introduce new questions because they create new content, recommendations, and decision pathways. Traditional deployments have historically focused on analyzing past information.
That shift raises the stakes for governance, especially in areas where AI-informed decisions directly affect consumers.
“If an AI system incorrectly refuses a mortgage to a household, that’s a catastrophic event for that family,” Ali said. “You’re dealing with decisions that could change the course of life for average citizens.”
For MSPs, consultancies, systems integrators, and other service providers supporting regulated clients, the message is clear: AI strategy cannot stop at deployment.
Financial services firms will need partners that understand governance, auditability, explainability, and exception handling as deeply as they understand automation.
Many CIOs are not seeing expected AI returns
Despite high levels of investment, Ali said many CIOs remain unconvinced that AI is delivering the returns they expected.
He said part of the problem is that some organizations are funding AI through budgets historically tied to robotic process automation, business process outsourcing, or broader process automation initiatives.
If those earlier efforts failed to address broken business processes, AI may simply amplify the same inefficiencies.
“When you don’t fix the underlying issue of your business process, and you plaster AI on top of it, chances are you are not going to realize what AI has to offer,” Ali said.
That gap between AI investment and measurable return creates both a challenge and an opportunity for service providers. Many clients are not simply looking for another AI tool; they need help identifying broken processes, cleaning up data environments, and connecting AI projects to business outcomes.
Providers that can diagnose those foundational issues before layering AI on top may become more valuable strategic partners as CIOs face pressure to show ROI.
AI adoption requires operating model changes
With all that said, organizations focusing on AI adoption only through a technical lens are still likely to miss the broader picture around change management.
For Ali, the most important AI conversations with executives are often less about specific technologies and more about how organizations are structured to use them.
“It’s no longer about management of resources. It’s now about who owns the outcome,” Ali said. “The person who owns the outcome — inclusive of people, process, technology, and AI — is going to adopt it faster and be more successful.”
AI will reshape entry-level roles
Ali said AI adoption will reshape parts of the workforce, particularly entry-level roles that involve repetitive administrative work, data entry, and highly standardized processes.
“I’m not going to shy away from the fact that, yes, there are certain roles that are going to disappear,” Ali said.
At the same time, he expects demand to grow for workers who can oversee AI-enabled processes, evaluate exceptions, and apply business judgment when automated systems encounter edge cases. Organizations will also need new expertise around AI governance and decision support.
The challenge for employers, he said, is determining how workers will gain experience and develop specialized skills if traditional entry-level training paths become less common.
Instead of learning primarily through repetition, future employees may spend more time understanding business processes, managing exceptions, and working alongside AI systems that handle routine tasks.