Not many banks have a direct partnership with OpenAI. Instead, they prefer to go through Microsoft Azure, AWS, or Google Cloud, which offers a mix of regulatory caution, commercial pragmatism, and technology strategy.

The primary reason being banking is highly regulated, especially around data sovereignty, explainability, and compliance. A direct deal with OpenAI often translates to custom integration work and increased responsibility for governance, which is not preferred by banks in general.

Despite this, UK-headquartered bank NatWest partnered directly with OpenAI earlier this year. But why?

Banks that go direct often want early or exclusive model access along with more control over data handling and model fine tuning.

Gunjan Kumar, MD, global head of digital platforms and India head for Wealth CDIO, NatWest Group, told AIM: “The partnership will give NatWest early access to OpenAI’s latest models and help accelerate the rollout of high-value, transformative AI use cases.”

This includes customer-facing pilots that could offer personalised financial coaching, empathetic real-time voice support, and new ways for customers to manage their finances.

The focus is on applying GenAI in regulated and high-context areas such as fraud prevention, risk modelling, and customer service, ensuring that technology augments rather than replaces human engagement.

“Being the first UK bank to sign with OpenAI gives us a direct line to the latest AI advancements, helping us respond to customer needs faster and more effectively,” Kumar mentioned.

Scaling AI Across the Bank

Before partnering with OpenAI, NatWest had already implemented a range of AI-driven initiatives across its operations. Its Cora+ chatbot handled over 11 million customer interactions in 2024, increasing satisfaction by 150% and cutting down to half the number of cases needing human intervention.

In wealth management, Kumar shared that AI-powered call summarisation tools saved advisers around 15 minutes per call by reducing manual note-taking. Fraud detection systems supported approximately 7,000 colleagues in identifying and preventing scams earlier, offering greater protection to vulnerable customers. 

Developer productivity tools such as GitLab Duo generated code suggestions, performed security checks, and created test cases. AI also sped up complaints handling, reducing resolution times by an average of 20 minutes per case. These efforts were underpinned by NatWest’s Kepler ML platform, which enables data discovery, model training, and deployment at scale.

With over 17,500 employees across Gurugram, Chennai, and Bengaluru, NatWest India plays a critical role in driving these innovations. The India team has been instrumental in developing Cora+ and call summarisation capabilities. 

Roadmap for the Next 12–18 Months

NatWest plans to scale the priority AI use cases it identified in 2023, broaden access to GenAI tools for more engineers, and provide prompt engineering training to enhance adoption, said Kumar. 

He added that the bank is also working to simplify its data architecture to accelerate AI deployment, while reinforcing responsible AI practices through a dedicated code of conduct and expanded ethics training. 

Building on its existing partnerships with Accenture and AWS, NatWest is now collaborating directly with OpenAI to further strengthen its AI capabilities and deliver more personalised banking experiences.

Despite increased automation, the NatWest Group emphasises that AI is designed to augment, not replace human interaction — especially in sensitive areas like financial advisory, according to Kumar. The bank sees AI as a way to enable self-service, provide faster resolutions, and give employees more time for high-value work, all while keeping human oversight central to customer trust.

Turn of Events

OpenAI has reportedly given banks a major breakthrough by releasing GPT-OSS, its 20B and 120B parameter open-weight reasoning models that can run locally on a MacBook Pro, any cloud provider, or even on-premise. For years, large banks have been slow to adopt AI — not due to a lack of interest, but because strict regulations and cybersecurity concerns made SaaS AI tools risky. 

JPMorgan’s CISO Patrick Opet has publicly warned that SaaS can “quietly enable cyber attackers,” weaken the global economic system, and has already forced the bank to isolate compromised providers after supply chain incidents. Every AI query to OpenAI previously meant sending customer financial data to a third party, a non-starter for regulated institutions. 

Now, with GPT-OSS, banks can deploy powerful AI without data ever leaving their infrastructure, eliminating vendor risk, privacy concerns, per-token costs, and rate limits. 

Meanwhile, banks like G-SIB, as reported, have slashed the time needed to compile environmental, social, and governance (ESG) reports from eight months to just one or two days by deploying OpenAI’s advanced models. The shift translates into nearly 39,000 hours of manual work saved and millions in cost efficiency.

Beyond ESG, analysts are already seeing a boost in productivity while using AI to speed up data extraction, content summarisation, and compliance-heavy workflows.

The post NatWest takes its AI engagement further with OpenAI Deal appeared first on Analytics India Magazine.