Organisations will not stand out of the crowd simply by deploying expansive artificial intelligence (AI) infrastructures. Instead, it will boil down to their ability to navigate the complexities and use AI agents to their advantage.
Significant investments are being poured into data centres and retrofitting these sites with massive compute power to meet growing AI demand.
These efforts eventually are likely to converge and differences between the AI models, to power what enterprises need for their AI workflows, will be marginal, said Alvaro Garrido, Standard Chartered’s COO of technology and operations and CIO of information security and data.
While compute power still will be essential, ultimately, what matters is how well companies orchestrate the models and navigate AI to enhance their customer interactions, Garrido said in a video chat with FutureCIO.
Based in Singapore, the executive was appointed to his current role in May 2025 and has been with the bank since May 2022, when he was group CISO.
Garrido noted that Standard Chartered itself would likely deploy a combination of on-premises and cloud services, rather than run a full-scale data centre, tapping services from cloud software vendors and suppliers.
“What will change fundamentally, though, is that we will get to a point where we’re no longer thinking about retrofitting or overlaying AI into processes,” he said.
Rather, he envisages a future where AI agents will be leveraged to differentiate services and how the company operates.
“Will we be able to do things we can’t even conceive now?” he posed.
This underscores the need to “scale up” its employees, who will be part of this evolving ecosystem, so they have the expertise and access to data required to support the transition, he said.
And like other banks and financial services institutes, talent is a challenge for Standard Chartered. he noted.
“Data now, as it is connected to the AI ecosystem, is becoming an increasingly complicated tool for talent,” Garrido said. “For us, the solution isn’t to simply hire outside, which we’re doing, but at the same time, [we’re also] cross-training and reskilling our workforce and trying to identify skills that are reusable and can be retained.”
These encompass the ability to manage the different phases as AI agents continue to advance, which introduces layers of complexities, he said.
Focus on agentic control framework
Garrido added that the necessary policies and processes must be established, such as whether an agent should be the primary decision-maker and if another agent should be set up to monitor and control another agent to ensure compliance.
He stressed that agentic objectives should be defined by humans, with controls decided by humans, while agents could be tapped to support the execution of all processes.
“I don’t see a case of agents operating on their own, but I do think we will see things we don’t know we can do,” he said. “We need to find the sweet spot between the two.”
Standard Chartered currently is reviewing its processes and running agentic AI pilots.
“We’re not preoccupied or [overly] concerned about the platform or model or technology of choice,” Garrido said, noting that this infrastructure served mainly to connect AI agents and data models as well as support technical capabilities, such as local inferencing. “It is the extraction layer to make it easy for business users to tap [agentic services].”
“For us, the key point is to provide a control framework around the agent, so we always know how a decision is made and that the decision is made within our regulatory framework, and we have traceability,” he said.
“There are a lot of tools out there and new models coming up all the time…we need to make sure we’re in control,” he added.
He noted that Standard Chartered is working to define what an agent should be and to create agentic workflows based on each individual user’s role.
These efforts include establishing chain flows between agents, he said.
In this aspect, operating in a highly regulated sector lends itself well since it means the bank already has the necessary lexicon to guide how it manages data, including tagging its provenance and restrictions.
It provides the foundation for the agentic layer, ensuring data continues to operate and propagate those requirements, Garrido said.
He added that AI also can be applied to manage data classification, as well as monitor and track the behaviour of AI agents.
Ultimately, the bank’s agentic initiatives should fortify user interface and not diminish the quality of its service delivery, he said.
Navigating crowded AI market
Asked how the vendor community could ease his job, he pointed to the crowded AI marketplace.
“We’re in that space right now where it’s hard to navigate what’s useful for your situation and what’s a good tool to incorporate to your tech stack,” Garrido said.
“We’re trying to understand how to use all the tech and tools in a good way,” he said. “It’s a little intense sometimes to distinguish what’s valuable.”
Standard Chartered in March 2025 launched its GenAI (generative AI) tool SC GPT, across 41 global markets, which it then said offered more automation and scale on how risks were measured, recorded, and reported.
It added that its employees could use the GPT tool to develop customised solutions for “market-specific needs”, including content generation for digital marketing and translation services.
The bank in April 2026 also extended its partnership with Singapore’s Agency for Science, Technology and Research (A*Star) Institute of High Performance Computing (IHPC) to drive AI research for the financial sector.
The three-year collaboration includes a SG$15 million investment to establish an AI for Banking Innovation Lab, which houses data scientists and researchers from both organisations.
The teams have been tasked to explore innovation across various areas, including fraud detection and natural language processing, with a project to develop a natural language interface for querying structured databases already underway.
The lab’s initiatives will tap advanced AI and machine learning to enhance operations and client experience. Successful projects may be spun off and facilitate next-generation financial services, the two partners said.
The latest agreement builds on previous collaborative work between Standard Chartered’s fintech investment and ventures arm, SC Ventures, and A*Star IHPC. Their initiatives then had looked at the use of GenAI in regulatory compliance and research in AI fairness.
In May this year, Standard Chartered CEO Bill Winters faced criticism for saying AI was replacing “lower value human capital”, as the bank announced plans to automate some roles over the next four years. He later sought to clarify his remarks, elaborating that employees would be offered “every opportunity to reposition” and reskill.
In a separate announcement that same month, the bank revealed it had hired Shebani Baweja as its group chief data officer responsible for shaping the company’s data strategy, including how data should be governed and used across the organisation.
Also based in Singapore, Baweja reports to Garrido.
