Financial institutions are on the verge of repeating the same mistake they made with telephone banking, internet banking and mobile banking: treating AI as another standalone service layer. In the first of our six-part series, we look at why the next phase of transformation will mean orchestrating end-to-end experiences, where AI works across the entire customer journey, carrying context across digital, conversational and human touchpoints.
When financial institutions introduced telephone banking, internet banking and mobile banking, each new channel improved access. However, very few institutions truly redesigned the end-to-end experience around the customer. Instead, they created parallel service models, fragmented ownership and disconnected journeys.
AI is now arriving in much the same way: layered in as another interface, another team, and another capability layer.
The pattern is easy to recognize. Smarter chatbots. Advanced Intelligent Virtual Agents (IVAs), powered by voice AI. More capable copilots. Increasing use of natural language and automation. Too often, these capabilities are being bolted onto already-broken journeys, rather than used to fundamentally redesign the customer experience.
What makes this transition different is the speed. AI is being adopted faster than any technology before. Generative AI has outpaced the uptake of both the PC and the internet, with tools like ChatGPT reaching mass adoption in months, not years.1
However, when AI is applied by financial institutions to fragmented customer journeys, the frictions increase. Customers moving from ‘app’ to ‘chat’ to ‘voice’ find they must repeat information to solve an issue. Service agents gain another tool, but not the customer context that would make it meaningful. Costs and frictions shift between channels rather than disappearing.
The fundamental issue is that the financial industry has a channel-era design problem. AI is just scaling it.
Capco and Raconteur research found that 62% of financial services leaders are planning major customer journey redesigns within 12 months, rising to 90% within five years.2 At the same time, the Bank of England and FCA’s 2024 survey found that 75% of firms were already using AI, with another 10% planning to do so within three years.3 The market is moving quickly, however, the faster deployment of a flawed design model simply scales the flaws.
This is why the next phase of transformation must be about designing AI Orchestrated Experiences, where AI works across the entire journey, carrying context across digital, conversational and human touchpoints. The value will come from designing unified, end-to-end journeys.
The customer experience is driven by their overall journey, not individual channels. However, most organizations still behave as though channels are the unit of design.
Digital teams optimize task completion in the app or on the website. Contact centers optimize handling time and containment rates, i.e. the ability to resolve issues inside their channel. Chat teams optimize deflection, i.e. diverting customers to self-service rather than human-assisted channels. Operations teams optimize the throughput of each process.
Each function improves its own metrics, but the customer pays the price of the gaps and frictions in between. These are becoming harder to ignore.
Capco’s recent work on conversational banking makes the point clearly: while financial services may have achieved multi-channel choice, omnichannel is the true end-state. Customers expect to start in one channel, continue in another and reach resolution without losing context. Isolated chatbot or IVA teams risk exacerbating fragmentation rather than reducing it.
This matters because AI deployed at the channel level inherits the channel’s limitations. A chatbot embedded on a website can only see what sits in front of it. A voice bot optimized for call deflection is incentivized to contain rather than resolve the customer issue. An agent copilot layered onto a customer service agent’s fragmented desktop can generate suggestions, but may not increase clarity.
Each tool may be more intelligent than its predecessor. But disconnected from journey context, it just becomes a faster way of reproducing friction.
That is why so many institutions are seeing a familiar pattern. More AI is being deployed but the overall customer experience is not getting better.
It would be easy to look at the last decade and conclude that conversational AI does not work in financial services. The story is more specific. Chat failed not because customers dislike a chat channel but because it was implemented in isolation from the journeys it was meant to support.
This is most visible in three journey types where fragmentation carries a high cost: servicing, fraud and claims.
Claims, in particular, are rarely one interaction. They are a series of moments: notification, data capture, status updates, clarification, exceptions, settlement and, sometimes, complaint. A disconnected bot may help with one of those moments, but only embedding AI into the wider journey can transform the experience.
Other evidence from the UK points in the same direction. Research by the Digital Regulation Cooperation Forum (DRCF) found that consumers see GenAI as one stage in a broader process rather than a complete solution, and still want the option to speak with specialists, especially for bigger financial decisions or where financial resilience is lower.4 That should change how firms design AI, not how fast they adopt it.
Value from AI will come from rethinking the experience end to end, not from deploying the most tools. At Capco, we call that AI Orchestrated Experiences, i.e. treating AI as the orchestration layer across the customer journey, not as another channel feature.
AI should understand intent, retain context, support decisioning, and guide the next best action. It should personalize and contextualize the experience in real time, without losing control, transparency or trust. AI should be embedded across products and journeys, rather than deployed through fragmented pilots.
This is also where UK market context matters. The FCA has been clear that existing frameworks apply to AI including Consumer Duty, accountability and governance. In practice, that raises the bar for customer journey design. If AI creates unreasonable barriers, breaks continuity or removes appropriate human support, the problem becomes a governance failure, not just a design one.
This is why the winners in the next phase will think differently. They will ask “where can AI reduce friction without removing trust?” rather than “where can we deploy AI?” They will design servicing journeys that prevent failure rather than merely absorbing demand. They will design fraud journeys that orchestrate urgency, reassurance and escalation. They will design claims journeys that connect moments in the overall customer experience, rather than digitizing them in isolation.
And they will recognize something that many institutions still have not fully accepted: human support becomes more important in an AI-enabled future, especially in the moments that matter most. Capco and Raconteur’s research found that 99% of leaders expect human advisers to remain integral to customer journeys, yet fewer than one in ten feel very confident in their ability to build long-term trust through AI, and only 5% feel very confident about aligning with regulatory expectations. The challenge is how to orchestrate both human professionals and AI properly.
Adoption is happening fast. Design is lagging. That is a mistake, and it is one that many institutions – having spent the last decade untangling telephone, internet and mobile fragmentation – cannot afford to repeat.
Technology will keep changing. However, design choices made now will define the customer experience for the next decade – opening up opportunities for those that recognize and solve the underlying problem earlier this time.
To discuss how Capco can help your organization design and implement AI Orchestrated Experiences, speak to our experts.
1 https://www.nber.org/papers/w32966
3 https://www.bankofengland.co.uk/report/2024/artificial-intelligence-in-uk-financial-services-2024
