AWS & Cognizant setting the North Star for AI in financial services – from compliance-first innovation to scalable transformation across the industry.
The Consumer Duty provides a set of rules set out by the financial regulator, the Financial Conduct Authority (FCA). It is designed to introduce a new standard of protection, mandating financial businesses in the UK to “act to deliver good outcomes for customers”. As such it imposes rules on four related areas: products and services, price and value, consumer understanding, and consumer support.
In its own right, the FCA Consumer Duty makes regulatory demands on financial services firms, but it also acts as a proxy for the compliance-intensive environment in which these organisations operate. Regulation is a particular challenge at a time when organisations are looking to scale artificial intelligence (AI) projects. Agentic and generative AI (GenAI) offers financial services organisations a unique opportunity to reinvent both operational processes and customer experiences – but it requires, too, that those firms understand the interrelationship between innovation and regulation.
A key question persists: how can organisations accelerate GenAI efforts with the level of governance required in the age of Consumer Duty? And it’s the question that was at the heart of an AWS and Cognizant event – appropriately called Scaling AI in the age of Consumer Duty – which took place in mid-November 2025.
The AI use cases moving from pilot into production
We are witnessing two shifts in the adoption of GenAI in financial services, offered one voice at the event. First, if 2024 was the year of the proof of concept, this has been a year of full deployment. Second, if banks, insurers and others started by looking at back office processes as candidates for GenAI adoption, increasingly firms are looking at customer-facing applications. “We started internal but now we are looking outwards.”
By way of example, insurance firms are seeking to apply AI to claims and underwriting workflows. There remains, too, a focus on research where automation is saving time, bringing efficiencies to the decision-making process. Lastly, there are early signs that GenAI has a role to play in payments processes, too.
Picking up on one of the examples above, another attendee boiled down a key area of deployment to one word: documents. “Documents are everywhere,” he said. “They are ‘heavy’ so wherever they are used – for an internal process or as part of a customer journey – there is value to be had from automation.”
The call center is another arena ripe for AI transformation. Firms are using GenAI to augment the skills of human operators, building analytical platforms to help better understand where customer frustrations lie, and the most effective paths to resolution.
AI catalyzes cloud and data strategy
The two most significant enablers of AI are, arguably, cloud and data. In turn, the eagerness to adopt GenAI – as well as build agentic solutions – has forced organisations to revisit both their cloud and data strategies.
On cloud strategy, AWS suggests that 20 percent of customers are modernizing their infrastructure so that they can leverage AI. A larger percentage of customers said they are actively engaged in reviewing their strategic approach in an effort to generate business value by moving to cloud.
On data strategy, one attendee said he was finally getting a hearing thanks to AI. He has been promoting the importance of quality data for 15-20 years but too many paid lip service to the idea. Now adherence to data lineage practices – where organisations properly track how data is generated, transformed and used throughout its entire lifecycle – is being taken seriously. Elsewhere, another voice expressed the view that AI was democratizing access to data. By making it available to relevant business functions, organisations acknowledge that individual parts of the business know best how to leverage data.
In order to avoid data cleansing becoming an endless task, it’s better to treat it as a piecemeal task. As one attended advised, assess the data required for your priority AI workloads and devote time and resource to “getting it into shape”. Over time you will begin to raise the overall quality of your entire data estate.
AI is not a “sticking plaster” for business as usual
This is an important moment for organisations willing to embrace the AI opportunity. The inference here is that many may not. “Most businesses have done digitization in the past; few businesses have done true process digital transformation,” argued one attendee. He noted that many firms “lift and shift” existing processes and applications when moving from analogue to digital. He offered creating a PDF facsimile of a paper-based form by way of example.
Digitalization, by contrast, demands that businesses rethink and optimize those processes. GenAI provides the latest opportunity to do so. Organisations that fail to use this moment to transform – not simply replicate – the way they operate are likely to fall behind. In the words of a one attendee, that means avoiding using AI as a “sticking plaster” for existing processes. Rather, it’s a chance to reconsider how those processes work and how they add value through, for example, generating greater efficiency or creating new business models.
Getting the innovation / compliance balance right
Pragmatism was a word that came up several times in the course of the evening. This was especially the case when discussing the best approach to innovation in a compliance-heavy market sector.
Decision making must match an organization’s risk appetite and, suggested one attendee, there is always room for maneouvre. While the old adage, “don't ask for permission, ask for forgiveness”, doesn’t quite apply here, a willingness to work at the edges of compliance is certainly an option for some. The question this attendee asks himself is: “What’s the minimum I have to do in order to gain value?”
Others expressed greater caution. One voice around the table said it is worth thinking about regulatory duty through “three planes”. The first is an understanding that “AI is a general purpose technology”, and as such “it should be decentralized”. The second plane, invoking the spirit of Consumer Duty, dictates that the organisation must do what’s right for the customer. The third plane – perhaps most chilling for those looking to take risk – is an understanding that financial services firms operate under senior manager and certification regime (SMCR), a UK regulation that holds individuals more accountable for their conduct and competence.
Elsewhere, some lamented the ever-changing regulatory landscape, recalling how his organisation had asked him to tack closely to the EU AI Act one week, only to reverse that decision a week later.
Consumer Duty: it’s good business
Despite some frustration about regulatory overreach in financial services (see above), a number of voices around the table argued that guidance such as the FCA Consumer Duty serves a useful purpose. It focuses mind, or in the words of one business analyst: “It’s good business”. Why? Because if you deliver more personalized products and services, better value, improved consumer understanding, and superior consumer support – per the demands of the regulation – you are more likely to prosper.
In further support of modern regulation, it was argued that the nature of standards setting and enforcement has changed for the good over recent decades. “The regulator has started to keep an open mind to the possibilities of technology. If I was to dial back 20 years, the thinking was from a perspective of safety and policy, not so much, ‘how can we participate and make innovations safe for all?’
“It is a marked shift in the regulatory landscape because they're actively participating in the ecosystem. They are helping to drive things like sandboxes, to engage partners, and to understand what is else is possible.”
