
Automated bank reconciliation in Business Central: Copilot
Automated bank reconciliation in Business Central: see what Copilot adds to automatch, where it errs, and how to keep a person approving every posting.
Home » Automated bank reconciliation in Business Central: Copilot
Yes. Automated bank reconciliation in Business Central runs in two passes: Match Automatically applies text rules, then Copilot proposes matches for the remaining lines, and an accountant reviews and posts.
That split matters to a controller. It shrinks the pile of unmatched lines without taking sign-off away from your team, which is what month end close automation should look like in practice.
This guide covers what Copilot adds, how it differs from automatch, how to switch it on, where it errs, and a year-end routine with one approval step. We checked every feature claim against Microsoft Learn on 16 November 2026. The Learn page was last updated on 2 October 2026 and still labels the feature preview, so re-check it before you go live.
Who approves what Copilot matches in your close?
Talk through your bank reconciliation routine with our team and see where an approval step fits before the year-end close.
Book a 30-min call →Book a 30-min call if you would rather walk through your own bank accounts with us.
Quick Answer Automated bank reconciliation in Business Central means Match Automatically handles text-based matches first, then Copilot proposes more matches from dates, amounts and descriptions and suggests G/L accounts for leftover lines. Copilot runs on Business Central online only, and a person reviews every proposal and approves every posting.
Copilot does not replace the reconciliation you already run. It works on the lines that automatch leaves behind, which is where your team’s hours go.
We use one simple model to keep the order straight, and we reuse its stage names through the rest of this post.
The QServices Framework Approval-First Reconciliation = Import → Automatch → Copilot → Review → Post
The order matters because each stage hands a smaller, harder pile to the next, and nothing reaches Post without a person at Review.
Microsoft Learn calls the feature bank account reconciliation assist (see Reconcile bank accounts with Copilot). It does two jobs.
In short, Copilot supplements automatch in Business Central. Microsoft’s own wording is that it is intended to supplement the automatch operation.
Here is how Copilot suggests bank reconciliation matches in Business Central: you run it on a reconciliation, it inspects the open statement lines and bank ledger entries, and it returns proposals you can keep or delete. The headline example from Microsoft is a customer who pays several invoices in one lump sum. Copilot reconciles that single statement line with the multiple ledger entries.
Nothing it proposes is saved until you select Keep it and close the window. That one design choice is why the Review stage in our model is not optional.
Key Insight Per Microsoft Learn, nothing Copilot proposes is saved to the database until you select Keep it. The review step is built into the feature, so skipping it means skipping part of the feature itself.
First, two terms. Bank reconciliation proves that the cash on your books agrees with the bank statement. Ledger matching is the line-by-line pairing that gets you there.
The match shapes work like this in Business Central terms:
Manual matching handles one-to-many. Learn adds a limit worth remembering: a many-to-one match with differences cannot be posted, even if the differences net to zero.
Automatch is text-based and rule-driven. Copilot adds a second pass that weighs dates, amounts and descriptions together, then proposes matches for a person to confirm. Automatch runs first through the Match Automatically button, and Copilot works on what is left.
The difference between automatch and Copilot in Business Central starts with inputs. Automatch compares text on a statement line with text on bank ledger entries and applies a Transaction Date Tolerance (Days) setting. You can inspect the basis for any match with Match Details.
Learn says these algorithms can sometimes leave many transactions unmatched, and its FAQ calls the rules inflexible. Partners report the same pattern when a document number is missing from the bank description. If your bank strips invoice numbers, expect a long unmatched list from rules alone.
Here is the automatch vs Copilot comparison, with manual work as the baseline.
| Manual | Match Automatically | Copilot-assisted | |
|---|---|---|---|
| Matching logic | Person reads each line | Text compare plus date tolerance | Dates, amounts and descriptions together |
| Fuzzy descriptions | Handled by judgment | Weak when text differs | Better, but can be wrong |
| Grouped payments | One-to-many by hand | Mostly one-to-one | Can match one line to several entries |
| Review needed | Every line | Spot checks via Match Details | Every proposal before Keep it |
| Audit trail | Posted entries only | Inspectable match basis | Saved only after Keep it; your sign-off is the evidence |
| Setup effort | None | Import format and tolerance | Highest: online only, admin activation, page permissions, extra review step |
The last row is the honest downside. Copilot adds a review step and setup work that rules alone do not need. That is a good trade only when your unmatched pile is big enough to justify it.
Setup is a configuration task, not a project. Most of the work is permissions and one pilot account.
Follow these steps to turn on Copilot bank reconciliation in Business Central online. This is the Copilot bank reconciliation setup in order:
Admins can also deactivate the capability or grant and deny access per user, which gives you a clean off switch. See Enable AI features in Business Central for the admin side.
No, plan on online only. Copilot runs on Business Central online only, and Learn lists online use as a prerequisite. We could not confirm how the feature behaves in any other deployment, so do not build an on-premises plan around it.
This is the flow we run for automated bank reconciliation in Business Central:
Capacity is bounded. Learn gives roughly 1,000 open bank ledger entries in the statement’s date window for 2025 wave 1 and earlier, and roughly 3,000 for 2025 wave 2 and later, varying with description length.
Microsoft publishes no defaults for these, so treat the list below as our starting points, not product settings.
A false match costs more than a missed one. An invoice wrongly shown as paid hides in the books until the customer calls, and it feeds exactly the mess described in how to clear unapplied cash.
Where Copilot gets bank statement matches wrong is fairly predictable. Learn’s FAQ says suggestions might sometimes be incorrect or incomplete, and that users must review them before Keep it.
Two limits are documented. Copilot works best when G/L account names, ledger descriptions and bank descriptions share a language, and mixed languages often give fewer matches. Microsoft also says it is not intended for fraud detection or to check whether customers paid on time.
Microsoft tested with synthetic data and published no accuracy percentage. We found no independent benchmark either, so measure your own match rate.
This is an illustrative scenario for a US operating account in USD, not a client result. The four invoices behind the ACH deposit are $6,200.00, $5,150.00, $4,300.00 and $2,800.00.
| Line | Match Automatically | Copilot proposal | Human action | Result |
|---|---|---|---|---|
| ACH batch deposit $18,450.00 | No, no single entry | One line to the 4 invoices | Confirm the four add up exactly | Matched 1:N |
| Wire fee $35.00 | No ledger entry | Suggests the Bank Charges account | Check the account, accept | Posted to Bank Charges |
| Check cleared 6 days late, $3,980.00 | No, outside 3-day tolerance | Proposes by amount and description | Confirm check number | Matched |
| Short-pay $12.50 (invoice $2,400.00, received $2,387.50) | No, amounts differ | May propose the invoice | Decide: write off or chase | Posted with a reason |
| Duplicate vendor payment $6,200.00 | First debit matches | Nothing to match | Investigate, ask for a refund | Stays open until credit arrives |
| Other 20 lines | 17 matched | 3 proposed | Review the 3 | All matched |
The duplicate payment is the line that proves the point. No matching tool should make it disappear, and Copilot correctly leaves it alone. For the short-pay, our guide to tracking each short pay shows the reason codes to attach.
| Line type | Who decides |
|---|---|
| Exact match | Rules, with a spot check |
| Partial payment | Person |
| Grouped deposit | Copilot proposes, person confirms |
| Bank fee | Copilot suggests account, person accepts |
| FX difference | Person |
| Duplicate | Person |
Four failures show up again and again in month-end bank reconciliation.
| Need | Match Automatically | Copilot | Custom agent |
|---|---|---|---|
| Exact text matches | Yes | Not needed | Not needed |
| Fuzzy or grouped lines | Weak | Yes, with review | Yes, with review |
| Rules across many banks | Per reconciliation | Not a documented feature | Yes |
| Intercompany matching | No | Not a documented feature | Yes |
| Exception routing and approvals | No | No | Yes |
| Build and upkeep effort | Lowest | Low | Highest |
A custom agent is worth it only when exceptions pile up in places Copilot does not cover. Otherwise it is extra upkeep.
Is auto-post switched on in your tenant?
If matched lines can post without review, a short call can help you check the setting and design an approval step that fits.
Book a 30-min call →No. Leave auto-post off. Learn documents a Post if fully applied option, but our policy is that a person approves every bank reconciliation posting.
Not by itself. Learn says that turning on Post if fully applied posts the reconciliation ‘when all lines (100 percent) are matched and you select Keep it.’ So even with the switch on, someone still selects Keep it.
The wording does not spell out whether Copilot-proposed matches count toward the 100 percent check. We treat that as undocumented, so test it in a sandbox. Even if you allow it, Business Central reconciliation auto-post skips the Review stage, which is the one stage that catches a wrong match.
Key Insight A reconciliation can reach 100 percent matched and still be wrong. A fully matched statement proves every line found a partner, not that each partner was the right one.
This is our policy, not a rule from Microsoft or a standards body. The reasoning is plain: the person who prepares should not be the only person who approves, and a tool that suggests and posts removes both people.
Undo exists on the Bank Account Statement List, and you correct erroneous ledger entries by reverse posting. Both work, but they cost a reopened period at the worst moment. Approval first is cheaper. For the wider principle, see why human-in-the-loop is non-negotiable.
Real-World Scenario A typical case we see: a controller runs six bank accounts with about 2,400 statement lines a month and a 5-day close target. In the naive setup, Post if fully applied is on for every account, and a clean-looking account posts with one grouped deposit matched to the wrong customer’s invoices. The error surfaces in the aging report two weeks later. In the Approval-First model, the Review stage catches it: the preparer keeps proposals, the approver reads the test report, and the grouped deposit gets a second look before anything posts. This is an illustration, not a client result.
Standards & References
- Reconcile bank accounts with Copilot: what the feature does, Keep it, Post if fully applied
- Enable AI features in Business Central: admin activation, online prerequisite, data movement
A year-end bank reconciliation checklist in Business Central differs from the monthly one because year-end close is a separate process, and errors now land in audited numbers. Pair this with our month end close process checklist.
The approval comes once, after the review work, so it stays quick and meaningful. For the monthly rhythm around it, see our accounts receivable reconciliation routine.
Baseline four numbers before the pilot: minutes per statement line, auto-match percentage, exceptions count and days to close. Do not borrow vendor percentages. Here is an illustrative month of 400 lines, not a client result:
| Rules only | Rules plus Copilot | |
|---|---|---|
| Statement lines | 400 | 400 |
| Matched by rule | 288 | 288 |
| Proposed by Copilot | 0 | 64 (55 kept, 9 deleted) |
| Manual lines | 112 | 57 |
| Time at 2.5 min per manual line, 0.75 per proposal | about 280 min | about 190 min |
The saving is real but modest, because the review time does not disappear.
We start at the Review stage, because that is where the control lives, and then build backwards through Automatch and Copilot. As a Business Central partner, we run the same team for build and upkeep.
Implementation Checklist
- Business Central online confirmed for the target environment
- Bank account reconciliation capability active on the Copilot & agent capabilities page
- Pages 7250 and 7252 permissions granted to preparers only
- Allow data movement setting reviewed against the Azure OpenAI geography
- Post if fully applied left off
- Preparer and approver are different people for every bank account
- Baseline minutes per line, auto-match percentage, exceptions and days to close recorded
- Parallel run completed on one account with test report filed
Pilot one operating account for one close with a parallel run, and measure auto-match percentage, false-match rate, minutes per statement line and days to close before you add more accounts for automated bank reconciliation. Book a 30-min call and we will set up the pilot with you.

Written by Rohit Dabra
Co-Founder and CTO, QServices IT Solutions Pvt Ltd
Rohit Dabra is the Co-Founder and Chief Technology Officer at QServices, a software development company focused on building practical digital solutions for businesses. At QServices, Rohit works closely with startups and growing businesses to design and develop web platforms, mobile applications, and scalable cloud systems. He is particularly interested in automation and artificial intelligence, building systems that automate routine tasks for teams and organizations.
Yes, in two passes. Match Automatically applies text rules, then Copilot proposes matches for leftover lines. This automated bank reconciliation Business Central setup needs Business Central online, because Copilot runs there only. A person still reviews proposals and approves the posting.
Only when you let it. Business Central reconciliation auto-post means the Post if fully applied option, which posts when all lines match and you select Keep it. Nothing is saved before Keep it. We leave the option off so a person approves every posting.
An administrator controls it. They can deactivate the capability or deny access per user, and the Allow data movement toggle applies when geographies differ. Microsoft says suggestions can be wrong, so review every one. Run a privacy review before go-live.
Setup itself is configuration: activate the capability, grant permissions on pages 7250 and 7252, and check data movement. In our experience the real timeline is one parallel-run close on a pilot account. Automatch vs Copilot results are visible after that single cycle.
Pick one operating account, record minutes per line, auto-match percentage, exceptions and days to close, then run Match Automatically and Copilot alongside your manual process for one close. Compare results, split preparer and approver, and expand only after that.

Automated bank reconciliation in Business Central: see what Copilot adds to automatch, where it errs, and how to keep a person approving every posting.

What is unapplied cash? It’s a credit held as a liability until matched. See the journal entries and when to apply, refund or write it off in your ERP.

An AI service desk agent triages tickets and drafts first answers from your knowledge base. See the hand-over rule that keeps a person in charge of changes.

Integration monitoring for ERP, CRM, bank and store syncs: what to watch, how to alert, and how an agent drafts the incident note before anyone looks.

Proposal automation that drafts first versions from CRM notes and past wins, while your reps write the parts that win deals and approve every send.

AI in accounts receivable can read remittances and flag late payers today. See what works, what’s hype, and where your team still approves every action.
Next step
Ready to put approval first in year-end?
Bring your current reconciliation steps and we’ll map where Copilot helps and where a person signs off on every posting.
Book a 30-min call →
Automated bank reconciliation in Business Central: see what Copilot adds to automatch, where it errs, and how to keep a person approving every posting.

What is unapplied cash? It’s a credit held as a liability until matched. See the journal entries and when to apply, refund or write it off in your ERP.

An AI service desk agent triages tickets and drafts first answers from your knowledge base. See the hand-over rule that keeps a person in charge of changes.

Integration monitoring for ERP, CRM, bank and store syncs: what to watch, how to alert, and how an agent drafts the incident note before anyone looks.

Proposal automation that drafts first versions from CRM notes and past wins, while your reps write the parts that win deals and approve every send.

AI in accounts receivable can read remittances and flag late payers today. See what works, what’s hype, and where your team still approves every action.
Every quote comes from a published review on GoodFirms, Clutch or our own site. Click a source to read the full review.
Trustpilot5.0★★★★★See reviewsGoodFirms5.0★★★★★45 verified reviews
G25.0★★★★★See reviewsClutch5.0★★★★★Quality 5.0 · Schedule 4.8
Google4.7★★★★★Mohali office“QServices was highly professional, knowledgeable, and approachable. We continue to work with them on other projects.”
“Quality assurance is often overlooked, but these guys take that very seriously. … their team of experienced, dedicated developers and highly skilled engineers quickly dissected and decomposed every problem and presented solutions and options that directly met my needs”
Erik BrownBusiness owner and software engineerClient review“The supplier delivered the application on time and was responsive to our needs. He was also flexible, which allowed us to make changes on the fly.”
Zoltan BodnarCFO, Yusen LogisticsGoodFirms“Thank you Sahil and team for wonderful collaboration right from day 0. … Highly recommend the developers.”
Ranbeer MakinClientClient review“They are always looking for new ways to improve their products and services and to deliver value to their customers.”
“I have had an excellent experience working with QServices. We have a highly collaborative relationship and I value them not just as contractors but as partners.”
Alex VanderElsClientClient review“Sahil undertakes every project with a high degree of professionalism. His communication style is unmatched and he is always available to resolve issues or just discuss the project.”
Philmore J.ClientClient review“Responsive, patient and professional. Consistent weekly attention to our project, and new ideas we hadn't asked for.”
Michael WilliamsonPresident and CEO, United Way of Southeast LouisianaGoodFirms“I have been working with them on a project for a year now … They have been able to solve problems that come up with bugs or development and able to communicate well with me.”
Steven LewisClientClient review“We had been working on the app for two years. QServices built it in two months.”
Brian WallerPresident, Technology Association of IowaGoodFirms“All projects were delivered on time and were fully completed upon delivery.”
“They had excellent quality. The hypercare they provided after the project went live is exceptional.”

Founder and CEO

Chief Sales Officer
Thanks, we’ve got your details. A named member of our team will reply within one business day.