Why multi-account reconciliation gets messy
A client may have an IRA at one custodian, a taxable account at another, and a trust account managed under a separate registration. Your firm may track those relationships in its CRM, portfolio system, planning software, and reporting tools. Each system holds part of the picture, and each custodian sends data on its own schedule and in its own format.
The work of reconciling those records often starts with downloading files, checking which feed is current, mapping account numbers, and comparing balances by hand. An employee may spot a difference, then open several systems to work out whether it came from a timing issue, a missing transaction, an account mapping error, or an actual discrepancy that needs attention.
That process can be manageable with a small book and a few accounts. It gets harder as the firm grows. More clients, custodians, registrations, and feed formats create more places for mismatches to hide. The risk isn’t limited to a wrong balance in a report. A stale or misidentified account can also affect review preparation, portfolio oversight, and the adviser’s understanding of a client’s position.
The practical goal is not to make every source agree at every moment. Custodian files may arrive at different times, and transactions can settle after they first appear. The goal is to identify which differences are expected, which need a person to investigate, and which are urgent enough to resolve before someone relies on the data.
What the manual process usually involves
Most firms have some version of a daily, weekly, or month-end routine. Someone receives custodian data, imports it into an internal system, and checks that accounts and holdings loaded correctly. They may then compare data across a spreadsheet, portfolio platform, CRM, or accounting records.
The recurring work tends to fall into a few categories:
- Account matching: Confirming that the custodian account belongs to the correct client, household, and registration in the firm’s records.
- Balance checks: Comparing cash, market value, and total account value across sources.
- Position checks: Looking for missing, duplicated, or differently identified holdings.
- Transaction checks: Comparing buys, sells, transfers, income, fees, and other activity.
- Exception follow-up: Finding the source of a mismatch, recording what was checked, and deciding whether to correct a record or wait for a later feed.
Each check sounds simple by itself. The time adds up because staff have to repeat it across accounts and keep track of what changed between files. A team member may need to remember that one custodian’s file arrived before a particular transaction settled, or that a certain account uses a different identifier in the planning system.
A spreadsheet can help organize the work, but it won’t automatically make the underlying process reliable. If people copy data into a template by hand, there’s still a risk of selecting an old file, pasting values into the wrong row, or overlooking a line that shifted after an export changed.
The first step is to document how reconciliation works today. A useful process map identifies the systems involved, who owns each step, how often feeds arrive, what counts as a material exception, and how a resolved issue is recorded. That information determines what an agent should check and when it should leave the decision to a person.
For a deeper look at the workflow, see how advisory firms automate custodian data reconciliation.
Where mismatches come from
A reconciliation process needs to distinguish between a data problem and a real difference in account activity. Common examples include:
- Feed timing: One system reflects a transaction that hasn’t appeared in another source yet.
- Account mapping: A custodian account number is linked to the wrong household or internal account record.
- Identifier differences: The same holding appears under different symbols, security identifiers, or descriptions.
- Duplicate or missing activity: A transaction appears twice, or is present in the custodian feed but missing internally.
- Corporate actions: A split, merger, or other security event causes positions or descriptions to change.
- Cash differences: Fees, interest, dividends, or pending transactions cause a balance to vary.
- File or import issues: A feed is incomplete, a column has shifted, or a file has been processed more than once.
These exceptions don’t all deserve the same response. A small balance difference that is explained by a pending transaction is different from an account assigned to the wrong client. Treating every exception as equally urgent wastes staff time. Treating every small difference as harmless can let a serious issue pass unnoticed.
A sensible process records the source, the comparison date, the amount or position involved, the type of mismatch, and the next action. It also keeps a history of previous exceptions. If a difference appears regularly and resolves after settlement, the team can recognize that pattern. If a new mismatch occurs in the same account, the agent can flag that it doesn’t match the usual pattern.
What an automated reconciliation agent does
An agent for this workflow doesn’t replace the custodian, portfolio system, or the people accountable for client records. It reads approved data, applies agreed comparison rules, organizes exceptions, and presents the evidence for staff review.
A well-scoped process can look like this:
- Collect approved inputs. The workflow retrieves custodian files or data through the firm’s approved connection. It also reads the relevant account, household, position, and transaction records from internal systems.
- Check feed quality. Before comparing figures, it checks that the file is for the expected date, contains the required fields, and has not already been processed. Missing columns or an unexpected format become an exception, not a silent import.
- Match records. It maps custodian accounts to internal records using approved identifiers. Where the match is uncertain, it flags the account instead of guessing.
- Compare values and activity. It checks balances, holdings, and transactions according to the firm’s rules. Those rules can include tolerances, timing windows, and the source of truth for each field.
- Classify and rank exceptions. It labels likely timing differences, mapping problems, missing records, and other mismatch types. It sorts them by defined criteria such as potential client impact, size, recurrence, and age.
- Show the evidence. Each exception includes the source records, comparison date, relevant values, and the reason it was flagged. Staff shouldn’t have to repeat the full search just to understand the alert.
- Route work and record resolution. The agent sends the exception to the right team member or queue. Staff record whether they corrected a mapping, updated a record, requested a feed review, or confirmed a timing difference.
- Recheck after action. Once a correction or later feed arrives, the workflow checks whether the exception has cleared and keeps the resolution history.
The agent should not silently change client or account records. A better starting point is read-only analysis and a review queue. Once the firm trusts the matching rules and exception handling, it can consider tightly limited actions, such as creating a task or adding a note for approval. Corrections to financial records should remain under a clear human approval process.
Prioritizing the exceptions that matter
A useful exception queue answers three questions quickly: What doesn’t match? How important is it? What should happen next?
Start by agreeing on categories with the people who actually work the reconciliation. For example, a high-priority category might include an uncertain client or account match, a significant unexplained balance variance, or activity missing from an internal record. A lower-priority category might include a known settlement timing difference that typically clears on the next feed.
Avoid setting one universal dollar threshold before reviewing the firm’s accounts and risk tolerances. A difference that is immaterial for one account may deserve attention in another context. The agent can use rules agreed by the firm, but a person should own those rules and review them when processes, systems, or reporting requirements change.
Each exception should have a status, an owner, and a due point for follow-up. A common failure is to generate a report full of useful observations without giving anyone responsibility for resolving them. The queue should make it clear whether an item is new, under review, awaiting a later feed, resolved, or escalated.
It also helps to review recurring exceptions as a group. If the same mismatch repeatedly comes from a known import issue or account mapping problem, the firm may be able to fix the cause instead of assigning staff to close the same type of alert every week.
A practical example
Imagine a household with three accounts across two custodians. The latest files show that one taxable account has a different cash balance in the custodian feed and the firm’s portfolio record. A recent dividend appears in the custodian file, but not in the internal transaction list.
The agent first verifies that the feed is current and that the account is matched to the right household. It checks the transaction date and the firm’s rules for when activity should appear in the internal system. If the timing fits an expected delay, it labels the difference as a likely timing exception and schedules a recheck after the next feed.
If the difference doesn’t fit the expected window, the agent raises the priority and shows the values, transaction details, and previous history. A staff member reviews the evidence, confirms whether the transaction was missed, and makes or approves any correction. The workflow then records the decision and checks whether the records match after the next update.
This process reduces the time spent finding and sorting the issue. It doesn’t make the judgment about a client’s financial records on its own.
Data controls and oversight
Custodian data is sensitive, so the workflow needs clear access boundaries. Give the agent access only to the systems and fields required for reconciliation. Keep credentials and permissions under the firm’s control, and avoid moving account data into tools that haven’t been approved for that use.
You’ll also want a record of which files or records were compared, which rules ran, what exceptions were generated, and who approved any resolution. The review history should be useful to the operations team and consistent with the firm’s retention and compliance practices.
Test changes against representative examples before they affect live work. Include known mismatches, expected timing differences, missing fields, duplicate transactions, and uncertain account matches. Staff should be able to tell why something was flagged and correct a bad match without fighting the workflow.
For a more detailed discussion of data boundaries, read about AI agents for advisory firms with real data guardrails. Governance also needs to cover who can change exception rules and how those changes are reviewed. This is addressed in AI agent governance for financial advisory firms.
What the work is worth
Reconciliation is often treated as overhead because it doesn’t directly produce a client deliverable. But repeated manual checking takes staff time away from client service, planning support, and other work the firm needs done.
For a financial advisory firm, reconciliation work can take staff time away from other priorities. That isn’t a guaranteed saving from automating reconciliation. It’s a useful prompt to examine how much time is tied up in manual operations, where avoidable errors create rework, and whether staff capacity is being spent on tasks that can be organized more efficiently.
A practical estimate should start with your own volume. Count the accounts and feeds in scope, how often they’re checked, how many exceptions arise, and how long the team spends finding and resolving each one. Include the time needed for review and documentation, not just the first comparison. Then compare the current workload with a realistic future process that still includes human review for consequential decisions.
Better reconciliation can also improve the information advisers use elsewhere. For example, a meeting brief can bring together information advisers need before a client meeting. That work is only as useful as the data it receives. You can read more about AI meeting prep for financial advisory firms.
If you want to map your current feed, matching, and exception process, Book a 60-min Omni Audit. In 60 minutes, we’ll work through three outputs: a process map, a prioritized opportunity list, and a practical next-step plan. No deck.
Where Omni Ops fits
An automated reconciliation workflow can be designed around the firm’s process. The work starts with the systems and data sources you already use, the rules your staff apply, and the approvals that must stay with a person.
A reconciliation workflow can be designed to gather records, identify mismatches, rank exceptions, and prepare a queue for the team. The design should specify which comparisons it can make, what evidence it must show, which cases need escalation, and which actions it cannot take. Start with a defined group of accounts or feeds, test the outcomes with staff, and expand only when the exception process is working as intended.
Reconciliation can also connect to broader operations without turning into one oversized workflow. Each agent has a separate job, permissions, and approval path. You can see how Omni Ops supports operational workflows.
The aim is a controlled handoff. The agent does the repetitive comparison and sorting. Your team reviews exceptions, handles judgment calls, and remains accountable for changes to client records.
Your next step
Before selecting software or building automation, write down the current process for one custodian and a representative set of accounts. Note the files involved, the internal records used, the checks staff perform, and what happens when a match fails. That gives you a practical starting point and exposes whether the main problem is feed reliability, account mapping, exception volume, or unclear ownership.
Then decide what success would look like in your firm. It may mean fewer manual checks, faster resolution of high-priority mismatches, a cleaner audit trail, or more confidence in the data used for client reviews. Choose measures you can track with the team rather than assuming an agent will remove all reconciliation work.
To assess the workflow and identify where automation is a fit, see Omni for financial advisory firms. You can also Book your 60-min Omni Audit. We’ll leave you with the three outputs, a process map, a prioritized opportunity list, and a practical next-step plan. No deck.
The best first version is usually narrow. It checks a defined set of data, explains why each exception exists, and gives staff a reliable way to resolve it. Once that works, the firm can expand the scope based on evidence rather than adding automation for its own sake.
Your guide is ready
Check your downloads folder. If it did not open automatically, use the button below.
Download the GuideYour guide is ready
Check your downloads folder. If it did not open automatically, use the button below.
Download the GuideTalk it through
Talk it through with Sam
30 minutes on what a Command Centre would look like for your business.
Book a call