Middleware Modernisation: The Bank Team’s Role in Making Transformation Work
Who owns the outcome of middleware transformation? A practical guide to bank stakeholders, decision authority, acceptance evidence and service ownership—from planning to retirement.
Perspectives · Banking · Architecture · Transformation
The implementation partner can deliver a platform. The bank’s stakeholders must define what it should achieve, establish how it should behave, and take ownership of the service it becomes.
In a middleware transformation meeting, the dashboard shows green. The gateway is running, interfaces are deployed and the integration partner has completed its planned tasks.
Then someone asks: Who approves a retry when the original payment outcome is uncertain? Who owns a customer complaint while two vendors investigate? Who confirms that the old disaster-recovery route can no longer execute the same instruction?
Those questions expose the bank’s real role in transformation. Its participation cannot stop at requirements workshops and a final sign-off.
Our previous article on middleware modernisation covered the plan, architecture and delivery approach. This companion looks at the people inside the bank who make those choices effective.
Start by separating delivery from accountability
A system integrator (SI) may implement services. A multi-service integrator (MSI) may coordinate vendors, dependencies, testing and readiness. Product vendors may supply the platforms. The bank must retain enough capability to challenge proposals, understand dependencies and make informed decisions.
India’s RBI IT governance Directions address Board oversight, management responsibilities, risk controls, continuity and assurance. Its IT outsourcing Directions emphasise that outsourcing must not diminish the institution’s obligations to customers or impede supervision.[1] [2] These provide context for accountable bank participation; the operating model below is an editorial recommendation, not a prescribed organisation chart. Banks should map it to their applicable requirements and approved delegation of authority.
For each material decision, identify one accountable decision owner, the responsible delivery teams, required control reviews, evidence and escalation deadline. A single accountable owner does not remove mandatory approvals or segregation of duties.
Who does what inside the bank?
1. Board and Board-level committees: set direction and oversight
The Board and relevant committees should understand why the bank is investing, which critical services are affected, and how the programme fits its strategy and risk appetite. Their role is oversight of material exposure, capability and progress at the appropriate level.
Ask for reporting on customer impact, concentration risk, unresolved control weaknesses, investment outcomes and recovery readiness. Detailed endpoint approvals belong with delegated management and technical authorities. Escalation should make material issues visible without turning the Board into the project’s daily delivery team.
2. Executive sponsor, CIO and COO: connect investment to a viable service
The executive sponsor resolves priorities across departments and secures business participation. The CIO owns technology delivery within the bank’s governance, including architecture, lifecycle and sustainable support. The COO connects the change to processing operations, staffing, cut-offs and customer-service continuity.
Together, they should establish funded outcomes, realistic migration waves and decision deadlines. They must resolve a familiar conflict: a launch date is fixed, but a dependency owner has not allocated people. A supplier escalation cannot replace an internal resource decision.
3. Business, product and journey owners: define correct behaviour
Payment, lending, deposits, treasury and digital-channel owners must define what success means for their journeys. Specify eligibility, limits, validations, customer messages, cut-offs, exception treatment and the distinction between accepted, pending, failed and completed.
These owners approve business requirements and acceptance criteria, assess customer harm and own business acceptance under delegated authority. A business analyst can document the rules, but an accountable bank owner must resolve ambiguity. “Technical team to decide” is an unsafe answer to a question about a duplicate debit or an expired customer instruction.
4. Operations, branches and customer support: make exceptions workable
Operations teams identify the work that happens outside the happy path: rejected files, pending transfers, manual repairs, reconciliation breaks and end-of-day dependencies. Branch and contact-centre representatives explain what frontline teams need to tell customers and which enquiries they can safely resolve.
They should help design exception queues, maker-checker controls, case ownership and escalation procedures. Test whether an operator can trace one instruction without unrestricted database access. Agree who contacts an affected customer, what information they can rely on, and how the case remains owned across shifts.
5. Enterprise and solution architecture: preserve coherent boundaries
Architects establish capability boundaries, integration patterns, information authority, shared standards and the permitted exceptions. They should challenge unnecessary synchronous dependencies, duplicated business logic and designs that make every change depend on a central middleware team.
Maintain architecture decisions and explicit trade-offs. A reference framework such as BIAN can support capability discussions; it does not settle the bank’s deployment choices. Confirm how the design works across primary and recovery environments and how it will evolve without breaking consumers.
6. Application owners and the middleware team: own both sides of the contract
Core banking, channel, payment, CRM and partner-interface owners must confirm the real behaviour of their systems: response meanings, capacity, maintenance windows, error codes and supported recovery operations. The bank’s middleware owner maintains integration standards, the interface catalogue, platform lifecycle and service ownership.
Together, they establish versioned contracts, stable business identifiers, timeout and retry rules, duplicate controls and consumer migration plans. Assign an end-to-end journey owner as well as component owners. Otherwise, every component can meet its local target while the customer journey fails between them.
7. Data owners, reconciliation teams and reporting: establish what can be trusted
Data owners approve meaning, mappings, validation rules, lineage, freshness expectations and access to banking data. Reconciliation teams define comparisons between instructions, source outcomes and downstream records, including who investigates differences and how exceptions close.
An ODS is a query-only replicated data store, populated by replication engines to support reads and reduce source-system read load. It cannot execute, retry, reverse or update a banking transaction. Its owners must publish freshness and quality expectations; absence from a lagging replica cannot establish that a source posting failed.
Reporting teams validate downstream totals, timing and regulatory or management-report dependencies. A middleware payload change may appear technically compatible while altering a business interpretation or missing a reporting cut-off.
8. Infrastructure, network, database, cloud and IAM teams: prove the foundations
These teams provide capacity, connectivity, routing, name resolution, storage, platform configuration, backup and recovery. Identity and access management (IAM) and certificate-management owners establish service identities, privileges, trust stores, renewal responsibilities and emergency-access controls.
Evidence should include realistic load and failover behaviour, dependency capacity, certificate rotation and access revocation. A firewall rule that works in production says little about an untested recovery route. Record who owns each certificate and credential, including those used by jobs and third-party connections.
9. CISO and security engineering: build protection into the journey
The CISO function establishes security requirements, challenges threat assumptions and oversees security risk treatment within its mandate. Security engineering translates those requirements into controls across gateways, services, brokers, logs and source connections. NIST’s zero trust guidance cautions against granting implicit trust solely because an asset or user sits inside a network.[3]
Require evidence for authentication, business authorisation, least privilege, encryption, secret handling, logging and vulnerability remediation. The security operations centre (SOC) needs actionable detections and a response path. A scanner report is one input; it does not establish that a stolen valid identity cannot misuse an authorised API.
10. Risk, compliance, legal and privacy: challenge exposure and define obligations
Risk teams independently challenge scenarios, materiality, compensating controls and residual exposure. Compliance identifies applicable obligations and reporting expectations. Legal reviews contracts and enforceable responsibilities. Privacy owners review purpose, personal-data exposure, retention and third-party processing.
Each function should record its assessment within its remit. Residual risk acceptance must follow the bank’s authorised governance, with escalation where limits are exceeded. A steering committee cannot make a mandatory obligation optional. Give exceptions an owner, treatment, review date and expiry; do not bury them inside meeting minutes.
11. QA, business testers and release management: prove and control the change
Quality assurance coordinates traceability from requirements and risks to meaningful tests. Application teams supply component evidence; business testers establish whether the complete journey behaves correctly. Include uncertain outcomes, duplicate delivery, incorrect entitlements, peak loads, source slowness and failover with work in flight.
Release and change management control the execution window, dependencies, approvals and recovery steps. Their decision process should consume business, technical, security and operational evidence. A change ticket marked approved is not a substitute for missing acceptance or unresolved material risk.
12. IT operations, service management and continuity: accept the running service
IT operations and service owners need monitoring, incident ownership, escalation contacts, configuration records and usable runbooks. Business continuity and disaster-recovery owners connect recovery procedures to critical banking operations and validate that people, partners and infrastructure can recover together. Basel’s operational resilience principles provide a broader reference for this focus on continuity through disruption.[4]
Accept the service through demonstrated tasks: trace a transaction, diagnose a backlog, rotate a credential, recover a dependency and reconcile pending work. Define hypercare shifts, decision cover, handovers and exit criteria before launch. Permanent teams need practice and access before project specialists leave.
13. PMO, procurement, finance and people teams: sustain the delivery capability
The programme management office (PMO) maintains an integrated plan, dependency register and decision log. It should report missing evidence and overdue decisions alongside completion percentages. Procurement and vendor management establish supplier obligations, escalation routes, support, knowledge transfer and exit provisions with legal and service owners.
Finance tracks transition costs, dual running, consumption and savings that depend on actual retirement. People and change teams arrange training, backfill and protected subject-matter-expert time. Transformation repeatedly stalls when bank experts are expected to support daily operations and full-time delivery simultaneously.
14. Internal Audit: provide independent assurance
Internal Audit assesses governance and control effectiveness through its independent, risk-based mandate. It may identify weaknesses during the programme and review whether management addressed them.
It should not become the owner of controls, the author of the design or the routine go-live approver. Those management decisions would compromise the separation the bank needs. Audit findings must have accountable management owners, actions and follow-up evidence.
Turn the stakeholder list into decision ownership
A RACI matrix—Responsible, Accountable, Consulted and Informed—is useful when it identifies decisions rather than simply listing departments against “middleware implementation.” Name the actual role-holder and deputy for each decision.
| Decision | Accountable bank role* | Required inputs and evidence |
|---|---|---|
| Approve journey behaviour and business acceptance | Designated business or product owner | Operations, application owners and testers; state rules, exception cases and end-to-end results. |
| Approve technical design or an architecture exception | Designated architecture authority | Application, data, security and platform reviews; decision record and exception conditions. |
| Accept a residual risk within delegated limits | Authorised risk-acceptance authority | Risk and control assessments, impact, treatment owner and expiry; required escalation. |
| Authorise the migration wave | Named release or cutover decision authority | Required acceptance and control approvals, readiness evidence, stop criteria and recovery plan. |
| Accept steady-state support | Designated service owner | Operations and continuity evidence, trained teams, access, roster and unresolved-issue ownership. |
| Authorise legacy retirement | Designated legacy service owner | Consumer confirmation, recovery-path checks, retention review and commercial closure plan. |
*Illustrative assignments. Map them to the bank’s approved structure, delegation and mandatory control approvals. The MSI coordinates evidence and actions; it does not inherit these bank decisions.
Engage the right people at each stage
Discovery: business and operations teams expose real journeys and exceptions; application and data owners verify dependencies; procurement and finance establish commitments and costs. Architecture, security and risk teams identify constraints before a product decision becomes difficult to change.
Design and build: owners approve contracts and controls in manageable increments. Bank experts attend decision workshops with authority to resolve questions. Record assumptions, owners and dates rather than allowing the implementation team to guess.
Testing and cutover: business, source-system, security, infrastructure and operations teams supply evidence against pre-agreed criteria. The cutover authority checks that required decisions are complete, knows who can stop the wave and confirms cover for in-flight transactions.
Hypercare and retirement: the bank owns customer-impact decisions and operational priorities. The MSI can coordinate investigation across vendors. Service owners accept support, track recurring issues and verify removal of old routes, jobs and credentials before expected savings are declared.
A practical example: every team is green, but the journey is not ready
A fictional bank modernises its loan-disbursement integration. Middleware and core teams pass their interface tests. Infrastructure demonstrates failover. Business testers confirm the successful journey. During a recovery exercise, however, an old retry worker attempts to resubmit a disbursement whose core response was lost.
The core application owner establishes the authoritative posting outcome. The middleware owner traces the original instruction and prevents competing execution routes. Operations and the product owner define treatment of the pending case and any customer communication. QA reproduces the failure and verifies the fix. Risk reviews the remaining exposure; the authorised bank decision-maker determines whether the wave can proceed.
The MSI coordinates the investigation and evidence across suppliers. It does not decide that a duplicate-posting risk is acceptable for the bank. Internal Audit may later assess the control and governance, but it does not take over the release decision.
This is why stakeholder participation must begin before the final readiness meeting. The missing control sits between ownership boundaries, and no single supplier’s completion report can resolve it.
Our perspective
A strong bank team does not need to perform every implementation task. It needs to know what it is buying, challenge how it will work, supply timely decisions and own the service after the programme closes.
The most useful question in a transformation meeting is often simple: “Which bank stakeholder owns this outcome, what evidence will they use, and when will they decide?” Clear answers help the delivery team move faster and give the bank a service it can govern.
Continue with Learning
Who can accept the remaining risk? Separate supplier recommendations, bank risk decisions, release authority and independent assurance in a migration readiness exercise.
Sources and further reading
The stakeholder assignments, decision matrix and fictional example are editorial recommendations. The references provide governance, outsourcing, security and resilience context; they do not prescribe every role assignment in this article.
- RBI — Information Technology Governance, Risk, Controls and Assurance Practices Directions, 2023. Check applicability and the current regulatory position for the institution.
- RBI — Outsourcing of Information Technology Services Directions, 2023. Reference for retained institutional obligations in outsourcing arrangements; read with applicable subsequent directions.
- NIST SP 800-207 — Zero Trust Architecture.
- Basel Committee — Principles for operational resilience.
Continue with Unscripted Perspectives
- Middleware Modernisation and Transformation: From Plan to Operating Model
- Secure Banking API Development Part 4 — Testing and release evidence
Put this perspective into practice.
Explore the concepts, make a decision and test what changes when the situation changes.
Topics: API Architecture, Banking Architecture, Inside the Banking Backbone, Operational Resilience, Perspective