We’re now living in the AI era, with leading organisations having transformed their approach to innovation. Already, 8% of financial services organisations expect to be AI-native by the end of the year, and 26% plan to achieve this within the next two years. What these leaders build over the next 12 months could set the industry’s competitive order for the rest of the decade.
Our research with more than 1,000 financial services and fintech leaders reveals that a small group of early movers is no longer experimenting with AI, but building the operational benchmarks that others will be forced to meet. Beyond simply adopting new tools, they are redesigning how decisions are made, how services operate and how value is created. By the end of next year, those doing this work will have something their competitors cannot buy later: institutional knowledge.
The question facing leaders is no longer whether to become AI native, but whether they are building capability now while benchmarks are still being defined, or preparing to play catch-up once they become the baseline.
Moving beyond agile to innovate
For two decades, agile has provided a trusted playbook for digital change. It brought disciplines of iteration, collaboration and incremental delivery that transformed how financial services built products. But with the rapid pace of AI evolution, the tempo has changed.
In our research, three-quarters of leaders say traditional agile processes are already creating bottlenecks that limit their ability to scale innovation quickly. This challenge is reflected by leaders we work with who describe a recurring pattern: every week, they feel set back to level zero as new models, features and tools emerge. Progress is constantly being disrupted by having to start again, creating:
- Perpetual fear of being a laggard – concern that a competitor has just adopted a capability you have not even evaluated
- Strategic fatigue – too many options, but not enough clear signals about which will endure
- Execution paralysis – teams waiting for the next release before committing to this one
- Risk asymmetry – if the upside is shared across the organisation but the downside is personal, leaders may default to safety
- Fear of looking behind the curve – no one wants to admit they do not fully understand how to use AI effectively, so decisions drift rather than surface difficult but necessary questions
- Attachment to proven metrics – initiatives that cannot be expressed in cost savings can struggle to win support, even when they build capabilities competitors will later rely on
Most financial services leaders (94%) agree that AI-native ways of working will be essential within the next two to three years. However, in practice, that belief collides with the much messier reality of personal risk. For senior decision-makers in established institutions, the stakes are high. If an AI initiative fails, it can be seen as a misjudgement, a governance failure or even a regulatory exposure.
Risk profile also differs by institution. AI-native start-ups with limited legacy exposure treat experimentation as necessary, while incumbent banks with large balance sheets and stringent regulators treat experimentation as a potential liability. Both positions are rational in context. The challenge is that competitive advantage does not wait for psychological comfort. During an inflection point, playing safe by waiting for certainty can become the riskiest choice available.
This challenge is compounded by the shift many in the field are observing, where progress is driven as much by breakthroughs in architectures and reasoning as it is by scaling.
Moving first into outstanding experiences
For long-established institutions, competitive advantage has historically been built on scale, brand recognition and product breadth. But as they turn to younger customers, those foundations are weakening.
New generations are less attached to brands and more attached to seamless experiences. They trust devices, recommendations and digital journeys more than legacy logos and will follow the service that feels most relevant and contextual in the moment.
Agentic AI is accelerating this shift because it is built around behaviour rather than static segmentation. It can:
- Observe patterns in spending, saving and travel
- Infer preferences for risk, convenience and value
- Tailor prompts, products and timing to how someone actually lives, not who they resemble on paper
The same behavioural understanding that helps a bank suggest a timely holiday deposit or short-term savings plan can help a lender spot when past credit issues no longer reflect a customer’s current discipline.
As agentic AI matures, this could become even more important, with a clearer separation between the financial utility layer, and the experience layer that customers actually see. The utility layer will still look like a bank, insurer or payment provider. It will hold deposits, manage risk, handle compliance and connect into clearing and settlement. But for many customers, the everyday relationship will be with an AI agent that sits above this layer.
In this landscape, competitive battleground moves from owning the customer’s account to understanding the customer’s context. The institution that best reads behaviour and acts on it helpfully wins, even if it is not the one that issued the first card or opened the first account.
This creates a sense of urgency in being the first mover – while there may be some risk around adopting new technologies, it may make the difference in capturing a new generation of consumers.
Preparing for change in 2026
Over the coming year, AI-native leaders will begin to pull ahead. Rather than waiting for perfect clarity or betting everything on a single flagship project, leaders can act in more deliberate ways.
Here are four decisions that will shape competitive position:
1. Define your risk posture explicitly
Are you prepared to be an early adopter in selected domains, or will you only move once patterns are proven elsewhere? Both can work, but they imply different expectations for market share, margins and pace of change. Make the trade-offs explicit rather than drifting into caution by default.2. Choose where you will compete in the value chain
Will you focus on being the most trusted utility in the background, powering others’ experiences, or do you intend to own the primary AI relationship with customers? Trying to do both without clarity risks under-investing in the capabilities either position demands.
- 3. Build behavioural understanding as a first-class capability
Traditional data programmes often stopped at reporting and dashboards. Competing in an AI-native market means treating behavioural data as an operational asset: continuously collected, interpreted and acted on. This includes not just financial actions but also how customers respond to prompts and nudges. - 4. Design human roles for an AI-native world
As AI agents take on more decision-making and execution, human roles will shift from manually processing work to supervising systems, setting constraints curating contextual tangible outcomes while dealing with exceptions. Investing in these skills now – product thinking, judgement, communication and oversight – will be as important as investing in the technology itself.
Partnering to lead
By 2028, most financial services organisations expect to be AI native in some form. The institutions using the next 18–24 months to build AI-native operating models, behavioural capabilities and trust frameworks will enter that future with years of learning already embedded. With progress coming from rapid breakthroughs rather than predictable scaling curves, waiting for clarity is riskier than acting early. Leaders who experiment now will compound learning long before the market stabilises.
The AI era is not waiting for anyone. The question is whether your organisation will help define the benchmarks others must follow, or spend the years after 2026 adapting to a competitive landscape shaped without you.
Discover how your organisation compares with competitors by exploring the full report and our research with 1,000 leaders.