Bank Merger Data Migration: How to Combine Two Institutions' Data on One Core
Bank merger data migration moves the acquired bank's customers, accounts and history onto one core. How to match customers, map products and convert.

Bank merger data migration is the conversion of the acquired institution's customer, account, loan and transaction data from its core system onto the surviving institution's core, followed by re-pointing every system and file that exchanged data with the old core.
Unlike a core replacement inside one bank, a merger conversion has to combine two institutions' data: match the customers who bank at both, map two product sets and two charts of accounts onto one, resolve account numbers that collide, and tell customers what changes before it changes.
The conversion itself is usually run by the surviving institution's core provider, on a conversion weekend agreed with that provider: some institutions convert at the legal close, many some months after it. The work that decides whether that weekend goes well is the matching, mapping and reconciling before it.
Where engini fits: engini does not convert core data; the core provider's conversion team does. engini connects to both institutions' systems, including databases and file drops behind each firewall through its on-prem agent, so your team can automate the repeated work around the conversion: comparing customer lists, tracking mapping exceptions in a table, reconciling every mock conversion with the same queries, and rebuilding the feeds that must point at the surviving core. Start with the on-prem agent guide or the engini docs index.
Key Takeaways
- A merger conversion combines two institutions' data, so it adds four problems a single-bank migration does not have: shared customers, two product sets, two general ledgers and colliding identifiers.
- Customers with deposits at both banks matter for deposit insurance: the FDIC insures the acquired deposits separately from the customer's existing deposits at the acquiring bank for an initial six months.
- Under Regulation DD, acquired accounts need no new account disclosures, but a mapping that worsens a term needs notice at least 30 calendar days before the change takes effect.
- An account a bank acquires through a merger is not an "account" under the Customer Identification Program rule, but the acquired institution's identification and due diligence data still has to move in usable form.
- Reconcile at every mock conversion with the same queries, by institution and in total, and keep an old-to-new cross-reference for every identifier that changes.
What is different about data migration in a bank merger?
Data migration in a bank merger is different because two sets of records describe overlapping customers with different products, codes and account numbers, and all of them have to land in one system that already holds live data. In a core replacement, one institution moves its own data to an empty core. In a merger, the acquired institution's data is loaded into a core that is already running the acquirer's business.
| Core replacement (one institution) | Merger conversion (two institutions) | |
|---|---|---|
| Target core | New, empty | Already live with the acquirer's customers |
| Customers | One set of records | Two sets, with customers who appear in both |
| Products | Re-created on the new core | Acquired products mapped to the acquirer's existing products |
| General ledger | One chart of accounts, remapped | Two charts of accounts merged into one |
| Account numbers | Can usually be kept | May collide with numbers already in use |
| Customer communication | Mostly about new systems | Also about changed terms, numbers, cards and statements |
| Timing | Set by the project | Set around the legal close: at closing or some months after it |
The conversion steps themselves, mock conversions, cutover weekend and post-conversion checks, are the same as in any core migration. For that sequence and the checks used at each stage, see core banking migration: cutover, parallel run and validation.
How is a bank merger conversion planned?
A merger conversion is planned backward from the conversion date, with data work starting as soon as the parties can share data and the customer notices timed against fixed regulatory deadlines. A typical sequence:
- Inventory both institutions' data. List every system that holds customer or account data at the acquired institution: the core, loan origination, card processing, online banking, the general ledger, the data warehouse, and any internal databases or file transfers.
- Agree on mapping rules. Map each acquired product, fee schedule, account type, branch and general ledger account to its target on the surviving core. Every code needs a target and an owner.
- Match customers across the two institutions. Decide which records describe the same person or business, and which only look alike.
- Resolve identifiers. Decide which account numbers can be kept and which must change, and build the cross-reference table.
- Rehearse the conversion. Run mock conversions on copies of production data and reconcile each one.
- Notify customers. Send the notices that the mapping requires, such as changes in terms, new account numbers, new cards and new online banking access, on dates that meet their deadlines.
- Convert and reconcile. On the conversion weekend, run the final conversion and the same reconciliation used in every rehearsal, then decide go or no-go.
- Run the post-conversion checks. Check the first statements, payment files and reports, and keep the acquired core available for reference as long as your plan requires.
Tip: Set the reconciliation queries after the first mock conversion and change them only on purpose. When the final numbers come from the same queries as every rehearsal, a difference on conversion weekend is easy to trace.
How do you match customers who bank at both institutions?
You match shared customers by comparing identifying fields across both institutions, usually the taxpayer identification number first, then name, date of birth and address, and by sending every uncertain match to a person rather than merging it automatically. A wrong merge combines two people's accounts; a missed match leaves one customer with two profiles, two statements and two sets of notices.
| Match result | What it means | What to do |
|---|---|---|
| Exact match | Same taxpayer ID, and name and date of birth agree | Link the acquired accounts to the existing customer |
| Probable match | Most fields agree, one differs (a married name, a moved address) | Review by a person, then link or keep separate |
| Conflict | Same taxpayer ID, but name or date of birth disagree | Investigate before conversion; it may be a data error at either institution |
| No match | The customer exists only at the acquired institution | Create a new customer on the surviving core |
| Business customers | Same business at both, possibly under different legal names | Match on taxpayer ID and legal name, and check signers and owners |
The match also matters for deposit insurance. When one insured institution's deposits are acquired by another, the FDIC insures the acquired deposits separately from the depositor's existing accounts at the acquiring institution for an initial period of six months, to give depositors time to restructure their accounts. The grace period does not apply to someone who held funds at only one of the two institutions, and acquired time deposits follow their own rules (FDIC: Merger of IDIs). The shared-customer list from your match is the list of depositors this rule affects.
How do you map two product sets and two general ledgers?
Map each acquired product to the surviving institution's closest product, then check every term that changes, because a mapping that worsens a customer's terms triggers a notice. Products rarely line up one to one: interest calculation methods, fee schedules, minimum balances, overdraft rules and statement cycles all differ between institutions.
Note: Under Regulation DD, an institution that acquires accounts through a merger does not have to give new account disclosures, but the advance notice rules apply if terms are changed (official comment 4(a)(1)-2). Notice of a change that may reduce the annual percentage yield or adversely affect the consumer must be mailed or delivered at least 30 calendar days before the change takes effect (12 CFR 1030.5). Build that deadline into the conversion plan: the product mapping has to be final before the notices go out.
The general ledger needs the same treatment. Each general ledger account of the acquired institution is mapped to an account in the surviving chart of accounts, and the trial balance is reconciled after each mock conversion, by institution, so that a mapping error shows up as a difference in a named account rather than in the combined total.
| Mapping area | What to compare | What goes wrong |
|---|---|---|
| Deposit products | Rate, interest method, fees, minimum balance, statement cycle | A worse term with no notice; interest calculated differently on the first cycle |
| Loan products | Rate type, payment schedule, escrow, late fees | Payment amounts that change at conversion |
| Fee schedules | Every fee code and amount | Fees charged that the customer never agreed to |
| General ledger | Each acquired GL account to its target | Balances posted to the wrong account, found at month end |
| Branches and cost centers | Acquired branches to surviving branch codes | Reports and regulatory filings by branch that do not add up |
What happens to account numbers, cards and payments?
Account numbers change when an acquired number is already in use on the surviving core or does not fit its format, and every change has to be carried into the systems and payment flows that use the old number. Keep a cross-reference table from each old identifier to its new one, and make every downstream system use it.
The identifiers that usually need a cross-reference:
- Account numbers, for deposit and loan accounts that are renumbered.
- Customer numbers, when shared customers are linked to an existing profile.
- Card numbers, when cards are reissued on the surviving institution's program.
- Online banking user IDs, when customers move to the surviving institution's platform.
- Payment references, for direct deposits, recurring debits and bill payments that name the old account.
History must stay joinable across the conversion date: statements, disputes and tax forms after conversion often need transactions from before it, and the cross-reference table joins them. For more on moving legacy banking data into new systems, see AI data migration in banking.
What compliance data has to move with the accounts?
The acquired institution's identification, due diligence and monitoring records have to move with the accounts in a form your compliance team can use, even though acquired accounts are not new accounts under the Customer Identification Program rule. That rule's definition of "account" excludes "an account that the bank acquires through an acquisition, merger, purchase of assets, or assumption of liabilities" (31 CFR 1020.100). Whether to refresh any of that information is a decision for your BSA team; the migration's job is to carry it over complete.
In practice, plan where each of these lands on the surviving core or in its connected systems:
- Identification data and documents collected at account opening.
- Customer risk ratings and due diligence notes.
- Open alerts, case files and their history.
- Holds, restrictions and legal orders on accounts.
- Beneficiaries, authorized signers and powers of attorney.
Note: Your core provider is a third party. The interagency guidance from the OCC, the Federal Reserve and the FDIC on third-party relationships applies to the provider that runs your conversion (OCC Bulletin 2023-17).
Where do AI agents help in a merger conversion, and where do they not?
AI agents help with the repeated comparison and triage work around a merger conversion, and they should not be the system that converts or loads core data. The conversion belongs to the core provider's tools, which are built and tested for it.
| Task | Good fit for an AI agent | Why |
|---|---|---|
| Group reconciliation differences by likely cause | Yes | Many differences share a cause; grouping them saves review time |
| Propose probable customer matches for review | Yes, with a person deciding | The agent reads both records; a person decides the link |
| Draft the exception note for each difference | Yes | The note follows the same pattern every time |
| Check mapping tables for codes with no target | Yes, or a plain query | A rule finds most of them; an agent can explain the rest |
| Convert and load core data | No | That is the core provider's conversion |
| Decide go or no-go | No | Named people decide, against written criteria |
Give an agent read access where reading is enough. An agent that only compares the two institutions' databases does not need permission to write to either.
How engini helps with bank merger data migration
engini handles the cross-system work of a merger conversion: reaching both institutions' systems, running the same comparisons at every rehearsal, tracking exceptions, and rebuilding the feeds around the surviving core. It does not map or load core data.
- Reach both institutions' systems. The on-prem agent runs on a server inside each network and opens every connection outbound, so neither institution opens inbound ports. It reaches on-premises databases (SQL Server, Oracle, IBM DB2, MySQL, PostgreSQL and ODBC sources), file drops in local folders and network shares, and internal applications. IBM AS/400 systems, the classic core in many older financial institutions, are among the systems reached this way.
- Run your comparison queries. On SQL databases, workflows run your own SQL statements and existing stored procedures (MS SQL, IBM DB2), so every mock conversion is reconciled with the same queries, by institution and in total.
- Track matches and exceptions in a table. Probable customer matches, unmapped codes and reconciliation differences can each be written to an engini Table, with an owner and a status. A table automation starts a workflow when a record is created or updated, for example to alert the product owner.
- Use an AI agent step where judgment helps. An AI agent step with read access to both sides can group differences by likely cause, compare two customer records and draft the exception note. Your team decides what is linked, fixed and signed off.
- Rebuild the feeds. Transaction files, clearing reports and payment instructions can be loaded into the surviving core or the local financial system by workflows that check the data first. Every integration runs both ways, so results can flow back to the systems that sent them.
- See every run. Run history shows each execution and can be exported, and notification rules email you when a workflow fails. The audit log can be exported and streamed to your SIEM.
- Control access. engini acts as the account you connect, so a read-only database account keeps a comparison workflow read-only. Roles including custom roles, single sign-on, and separate development, staging and production environments are part of the platform. engini holds SOC 2 Type II and ISO 27001; details are on the security page. See how engini supports financial services.
- Code or no code. Operations teams build in Workflow Studio; developers call the same tools from the Python or TypeScript SDK, the REST API or the CLI.
When you do not need engini: for the conversion itself, use the core provider's conversion tools and team. If the provider's conversion reports already cover your reconciliation and customer matching, and neither institution has on-premises systems to reconnect, you may not need another platform. For a one-time bulk load of history into a data warehouse, a dedicated ETL tool is the usual choice. For connecting AI agents to legacy systems more broadly, see MCP for finance systems and legacy cores.
Agent gateway
Ship agents into systems you can’t afford to break
Connect AI agents to your ERP, CRM, financial and legacy systems through MCP, SDK, API or canvas, with secure auth, tool permissions and an audit trail built in.
Frequently Asked Questions
What is bank merger data migration?
Bank merger data migration is the conversion of an acquired institution's customer, account, loan and transaction data onto the surviving institution's core. It includes matching shared customers, mapping products and general ledgers, and re-pointing every connected system.
How long after a bank merger does the core conversion happen?
There is no fixed rule; the conversion date is agreed with the core provider, at the legal close or some months after it. Leave time for mapping, mock conversions and the customer notices that must go out before the date.
What happens to customer data during a core system migration?
Customer data is extracted from the acquired core, mapped to the surviving core's formats, matched against existing customers and loaded on conversion weekend. The acquired institution's records are kept for reference as the plan requires.
Are deposits insured separately after a bank merger?
Yes, for an initial six months. The FDIC insures deposits acquired in a merger separately from the depositor's existing accounts at the acquiring bank for that period, with separate rules for time deposits.
Do acquired accounts need a new Customer Identification Program check?
No, the CIP rule's definition of account excludes accounts a bank acquires through a merger. Your BSA team still decides what due diligence information to carry over or refresh.
Can AI agents run a bank merger data migration?
No, the conversion itself should be run with the core provider's conversion tools. AI agents help around it by grouping reconciliation differences, proposing customer matches for review and drafting exception notes.
Sources
- FDIC: Merger of IDIs (Financial Institution Employee's Guide to Deposit Insurance)
- CFPB: 12 CFR 1030.5, Subsequent disclosures (Regulation DD)
- CFPB: Official interpretation of 12 CFR 1030.4, comment 4(a)(1)-2, Acquired accounts
- eCFR: 31 CFR 1020.100, Definitions (Customer Identification Program for banks)
- OCC Bulletin 2023-17: Third-Party Relationships: Interagency Guidance on Risk Management
- engini docs: On Prem Agent
- engini docs: MS SQL and IBM DB2
- engini docs: Table Automations and AI Agent
- engini docs: Workflow Run History and Notifications