Lead, Partner or Participate – Orchestration Choices for AI Banks
Why control of the customer journey may become the next strategic fault line in banking
Imagine a large UK bank, one that has already spent several years strengthening its mortgage origination capabilities. Its models are accurate. Decisioning is fast. Customer experience scores have improved dramatically.
Then a technology platform, already sitting in the customer journey through open banking data aggregation, launches its own mortgage comparison and application product. Overnight, the bank's origination capability becomes a back-end service inside someone else's journey. The economics, the relationship and the data feedback loop now begin to move away from the bank.
In this scenario, the vulnerability was not the quality of the bank’s technology. It was instead the position from which that technology reached the customer. The bank had built a better capability without securing the layer that controlled how that capability was surfaced, selected and experienced.
This is the competitive battle that is now taking shape in banking. Many institutions still frame AI advantage primarily in terms of internal capability: better models, faster processes, lower costs and greater personalisation. Those capabilities matter. But in an agentic financial ecosystem, value also depends on who controls the interface, the workflow logic and the feedback loop through which customer decisions are made.
In the agentic era, balance sheet strength will remain essential. But durable advantage will increasingly compound around institutions that control the orchestration layer around the customer.
While other articles in this series have focused on the foundations of the AI-enabled bank – from architecture and operating model to workforce and governance – this one digs into what happens when autonomy moves beyond the institution and into the ecosystem surrounding it.
What does orchestration control really mean?
Banking has always had intermediaries. Brokers, aggregators, comparison platforms and open banking apps have all inserted themselves between institutions and their customers. What changes in the agentic era is the depth and persistence of that intermediation.
Intermediaries typically engage at specific moments, be that a sale, a product comparison, an application or a switching decision. An agent that helps a consumer manage their financial life on an ongoing basis could be involved across a wide range of interactions and decision points potentially touching payments, savings, credit, insurance and investments. While the regulatory outlook for such deep integration is still a matter of debate1, what is clear is that the entity controlling such an agent would have great scope to shape the customer's financial relationships more deeply than any comparison site or switching platform ever could.
Orchestration control has three compounding sources of advantage:
- The interface through which the customer experiences a financial service
- The workflow logic that determines which products and providers are surfaced
- The data feedback loop that improves decisions over time.
Each can create durable advantage. Yet each can also be owned by an institution that manufactures no financial products at all.
|
|
|
|---|---|
|
Where this is happening already |
Apple's expansion into financial services is not primarily a product strategy. It can also be viewed as an orchestration play, giving Apple a privileged position across the customer interface, identity layer and payment moment. Technology platforms do not need to replicate every banking product to gain influence. Instead, they can control the customer layer while regulated institutions provide the balance sheet, account, card or payment capability that sits underneath. |
Four ecosystem positions raising different questions
Most strategic discussions concerning ecosystem participation treat the choice as binary: either build your own platform or join someone else’s. In practice, however, an agentic financial ecosystem could create several distinct positions, each with different economics, governance requirements and competitive dynamics.
The pressing issue is that many institutions may drift into one or more of these positions before making an explicit choice about the roles they want to play, having carefully considered the opportunities and challenges each role presents. Some firms may combine roles deliberately. Others may move between them over time.
The purpose of the framework is therefore not to force a single choice, but to clarify where the institution intends to compete, where influence sits, where accountability remains and where economics accrue.
Table 1: Four agentic ecosystem positions
Source: Deloitte
The critical observation is that the Orchestrator position is where platform economics are most likely to compound over time in the institution's favour. Other positions can still be valuable, particularly where the institution has scale, regulatory trust, specialist capability or balance sheet strength. But they face greater risk of margin pressure as orchestrators accumulate data, relationships and leverage.
Yet not every institution should pursue the Orchestrator position, where governance complexity, investment requirements and execution risks are substantial. For many institutions, a specialist capability role or regulated product and balance sheet role may be a better fit. The important point is that this should be a deliberate choice, rather than an accidental outcome.
|
|
|
|---|---|
|
Strategic diagnostic |
Map your three most important customer journeys. For each, identify who controls the interface, who owns the data feedback loop and where the margin accrues. Where the answer is a third-party, ask whether that position reflects a deliberate strategy or an accidental dependency. |
The ecosystem competitor hiding in plain sight
While banks are typically benchmarked against one another and, to a lesser extent, established FinTechs, in the agentic era the more consequential threat may sit outside both categories.
The big global tech giants each possess components that could make them powerful financial services orchestrators in their own rights: customer interfaces, identity layers, operating systems, cloud infrastructure, payment touchpoints, data scale and AI capabilities. Their strategic advantage does not require them to become full-service banks, and they need only influence the layer through which customers access banking.
In contrast, the global digital commerce giants illustrate this pattern from a different starting point. Through marketplace relationships, payments, seller financing and cloud infrastructure, they can place financial services inside existing commercial journeys rather than asking customers to begin with a bank. The regulated entity may still provide the underlying financial product, but the customer context, data trail and decision moment increasingly sit elsewhere.
Elsewhere, players majoring in search, maps, payments and cloud create yet another form of orchestration dynamic, their core businesses already shaping discovery, identity, location context, device access and payments behaviour. In financial services, that creates the potential for banks to feature less as the owner of the relationship and more as the regulated provider inside another’s customer journey.
A critical strategic question for banks is whether their ecosystem position deepens the customer relationship, or whether their banking infrastructure merely provides rails for someone else’s journey?
Data network effects, where they compound and where they are constrained
Agentic systems learn from interaction. An agent that manages more customer journeys, across more financial products and life events, builds a richer model of financial behaviour. That can improve recommendations, risk assessments and personalisation. That feedback loop also creates durable advantage for whoever controls the data, provided they can use it lawfully and responsibly.
This qualification matters, since regulators constrain how personal data can be used in automated decision-making and profiling, particularly where decisions have legal or similarly significant effects. They also require institutions to provide positive customer outcomes, and firms must be able to show that data-led journeys support them, avoiding foreseeable harms.
The implication is that data advantage in banking is built through disciplined stewardship, not data accumulation at scale. Institutions with higher-quality, better-governed and more ethically deployed data will be better placed to build more effective agentic systems than those with larger but less well-governed data estates. The competitive advantage is real, but it accrues to quality, governance and permissioned use.
That distinction also determines which partnerships create durable advantage and which create regulatory exposure. Data-sharing arrangements that enrich a third-party's model without clear governance concerning how that data is used, for what purpose and under which constraints create commercial, conduct and data protection risks.
The ecosystem dependencies that are harder to control
Third-party dependency management is not new. Banks already operate within established expectations for oversight of outsourcing, supply chains and operational resilience. What changes in an agentic ecosystem is where dependency can begin to affect the customer journey. Some dependencies will still sit inside services the bank has chosen to outsource. Others, though, may sit within orchestration layers the bank does not control, but through which its customers discover, compare, select and use financial products.
That distinction matters. If a customer uses a whole of market financial agent or platform to manage their finances, the bank may become one product provider inside another organisations journey. The bank may not control how its product is presented, what assumptions the agent makes, or how the expectations of the customer are shaped. Even so, if the product does not perform as the customer expected, the bank may still face reputational consequences, service queries or complaints handling costs, even where responsibility for the recommendation or orchestration sits elsewhere.
For leadership, the question is therefore practical: which customer journeys depend on external platforms, data sources, infrastructure, and how does that impact their line of sight to the financial behaviours and expectations of their customers?
Of course, such questions also have regulatory and supervisory dimensions, which this article does not seek to address. Our focus here is rather on the strategic leadership questions surrounding how much of the customer journey the bank understands, influences and protects as more activity moves through shared ecosystem infrastructure.
Table 2: Examples of ecosystem dependencies that leadership teams should understand
Source: Deloitte
Taken together, these examples show why ecosystem position matters. A bank may understand its own systems and suppliers, yet still have limited visibility over how shared infrastructure, external platforms, model providers and data intermediaries shape customer journeys beyond its organisational boundaries.
The strategic issue is whether leadership can see where these dependencies sit, how they affect customer experience and where the bank’s influence is strongest or weakest.
Cross-institutional accountability
Open banking and API connectivity have already created workflows that cross institutional boundaries. Agentic AI [EH4.1]deepens that interdependence considerably. An autonomous agent may initiate a payment, verify an identity, adjust a credit exposure and source liquidity across several institutions in seconds, all in response to a single customer instruction.
When that workflow produces an error, a discriminatory outcome or a financial loss, accountability may be difficult to trace. The problem may have emerged from the interaction between components, from the orchestration logic that connected them, or from a data handoff between systems that each met their own standards but produced a combined output that served the customer poorly.
As more customer journeys are shaped by agents, platforms, data providers and regulated institutions working together, the practical challenge becomes one of visibility and handoff management. A bank may own one part of the journey, a platform may shape the interface, a data provider may influence eligibility or personalisation, and another party may provide a decisioning or workflow tool.
This does not mean responsibility disappears or becomes impossible to locate. The issue is that leadership teams may need a clearer view of how the journey works in practice: what the bank owns, what partners provide, where customer-impacting decisions are made, how issues are escalated and how the customer experience is protected when something goes wrong.
|
|
|
|---|---|
|
Governance requirement |
For every agentic workflow that crosses an institutional boundary, banks need to map accountability explicitly. That includes who is accountable for the part of the workflow inside the regulated firm, how exceptions are escalated across organisational boundaries and who has the power to intervene, pause or override. Where the workflow affects a bank’s decisions, customers or risk profile, the bank should be clear which outcomes remain within its own accountability, which sit with partners and how those responsibilities are governed, evidences and escalated. If ownership of those outcomes, escalation paths and intervention rights is unclear, the workflow is operating with a governance gap. |
Building ecosystem roles deliberately: three moves that compound advantage
Ecosystem strategy in the agentic era requires deliberate sequencing. Banks need to decide which role, or combination of roles, they intend to play, then make partnership, data and governance choices that support that strategic direction.
Three moves matter most.
The questions leaders should be asking
The AI-enabled bank will still need strong products, efficient operations, robust models and resilient balance sheets. But those capabilities will create less strategic value for the institutions involved if they sit behind some else’s customer journey. The next phase of competition will be shaped by ecosystem, turning on who controls the interface, who governs the workflow, who owns the feedback loop and who is accountable when autonomous decisions cross institutional boundaries.
The institutions that embrace competition openly, choose their ecosystem role, or combination of roles, deliberately and govern their ecosystem relationships with the same rigour they apply to their own systems will be far better placed to compound advantage over time. Meanwhile, those that drift may discover that market-leading capability can still become invisible infrastructure.
In the agentic era, intelligence only creates advantage when itreaches the customer through ecosystem relationships the bankunderstands, chooses and governs deliberately.
References
1. Our UK AI Hub provides space for Deloitte’s topic experts to publish their perspectives on the major transformation questions raised by AI adoption in banking and financial services: how institutions modernise their technology estates, redesign operating models, reshape workforces, strengthen governance and scale AI safely and effectively. Regulation is central to that agenda and is a substantial and rapidly evolving topic in its own right. The regulatory implications of AI depend heavily on use case, institution, type of customer or client affected, the jurisdictions involved and the ways in which technology is designed, deployed and controlled. For that reason, this series of article focuses on the transformation themes at hand. For detailed information on regulatory considerations related to AI, we direct readers to our European Centre for Regulatory Strategy (ECRS) for the latest regulatory developments in AI. https://www.deloitte.com/uk/en/blogs/ecrs.html
2. For further details, please see: The UK’s Critical Third Parties regime is finalised (Deloitte, 19 Nov 2024).
3. For further details, please see: Stronger Together: building Third Party Resilience in the Financial Sector (Deloitte, 01 April 2026)
