Automating Remittance Matching in 2027
Automating remittance matching means letting software read the explanation that comes with every customer paymentdetermine which invoices it settles, and apply the cash without a person retyping anythingng. AI can now read a remittance advice it has never seen before, agents can log in to a customer's AP portal or email a customer to ask what a payment was for, and the payment rails themselves are starting to carry far richer data. For most finance teams, the question is no longer whether remittance matching can be automated, but how much of it, how quickly, and with which platform.
It's important because remittance matching is where cash application succeeds or stalls. A payment that cannot be matched sits as unapplied cash, the invoice stays open, the customer's credit limit stays tied up, and sooner or later a collector chases someone who has already paid. Multiply that across thousands of payments a month (for some of the businesses we have seen) and the effect shows up in DSO, in the month-end close and in customer relationships.
Viewpoint Analysis is a Technology Matchmaker: we help finance leaders find and select the right technology fast. This piece is part of Our Viewpoint, a series in which we set out how we would approach a significant technology decision if we were sitting in the buyer's chair. It is written for finance, AR and shared services leaders who want to understand what has changed in remittance matching, which vendors are leading the way, and how to plan a project that delivers.
What This Guide Covers
• What remittance matching is, and how it differs from cash application
• Why matching is still hard, even for well-run AR teams
• How the technology has developed, from bank references to AI agents
• The six changes reshaping remittance matching in 2027
• The vendors at the forefront, grouped by approach
• How we would plan a remittance matching project, the terms worth knowing, and the most common mistakes
What Is Remittance Matching?
A remittance advice is the information a customer sends to explain a payment: which invoices it covers, any credit notes taken, and the reason for any amount held back. It can arrive inside the payment itself, in a bank file, as a PDF or spreadsheet attached to an email, in the body of an email, through EDI, on a lockbox image, or sitting in the customer's own AP portal waiting to be downloaded.
Remittance matching is the work of linking that explanation to the payment it describes, and then to the open items on your ledger. It is one step inside the wider process of cash application, which also covers receiving the bank data, posting the result to the ERP and handling the exceptions. Put simply, cash application is the whole job, and remittance matching is the part that decides whether the job can be done automatically. Our Cash Application Software Options 2026 report looks at the wider process and the specialist vendors in more depth.
Why Remittance Matching Is Still So Hard
The root of the problem is that the money and the explanation usually travel separately. The bank tells you that money has arrived. The customer tells you, somewhere else and often later, what it was for. Everything difficult about remittance matching comes from bringing those two pieces of information back together.
• Payment references are short. In the UK, Faster Payments and Bacs references are limited to 18 characters, so a payment covering a dozen invoices cannot describe itself. The detail has to come from somewhere else.
• One payment, many invoices. Large customers often pay a whole statement or a week of invoices in a single transfer, sometimes across several of your legal entities.
• Amounts rarely line up. Short payments, deductions, early payment discounts, bank charges, currency differences and overpayments all mean the payment total does not match any obvious combination of invoices.
• Remittances arrive in every format. No two customers send the same layout, and the same customer may change format when it changes its own systems.
• The payer is not always the customer. Payments arrive from parent companies, shared service centres, factoring firms or payment providers whose names do not match anything on your customer list.
• Some remittances never arrive at all. Others sit in an AP portal that someone on your team has to log in to, one customer at a time.
Well-run AR teams cope with all of this, but usually by throwing experienced people at it. Their knowledge of which customer pays how is often the most valuable asset in the department, and very little of it is written down.
The Real Cost of Unmatched Cash
Unmatched cash is not lost cash. The money is in the bank. But until it is applied, the business behaves as though it is not, and that has consequences well beyond the AR team.
Customers stay on credit hold for invoices they have paid. Collectors spend time chasing balances that are already settled, which is one of the quickest ways to irritate a good customer. DSO and aged debt reports overstate what is really owed, so the cash forecast is built on numbers everyone quietly adjusts. Month-end close waits for unapplied cash to be cleared or explained, and auditors ask why items have sat in suspense for weeks. None of these costs shows up on a single line of the P&L, which is exactly why they are easy to underestimate when building the case for automation.
A Question Worth Asking Your Team |
Before you speak to any vendor, ask how much cash was sitting unapplied at the last three month ends, how long the oldest items had been there, and how many hours a week go into matching by hand. Those three numbers will tell you more about the size of the opportunity than any benchmark, and they become the baseline for your business case. Our guide on How to Build a Business Case for AR Automation shows how to turn them into a financial case.  |
How Remittance Matching Technology Has Developed
Automation in this area is not new, but each generation has solved a different part of the problem. Knowing which generation a platform belongs to, or which mix of them it uses, makes vendor conversations much easier to follow.
Approach | How it works | Where it struggles |
Manual matching | People read bank statements and remittances and apply cash by hand in the ERP | Volume, speed, holidays and staff turnover |
Rules and templates ("auto cash") | The ERP or a bolt-on matches on exact references and amounts, with templates built for each common remittance layout | Anything that does not fit a rule, and every new layout needs setting up |
Machine learning | Models learn from past matches and propose likely matches with a confidence score | Needs history to learn from, and still depends on getting the remittance data in |
AI agents | Language and vision models read any remittance first time, and agents fetch missing data, request remittances and resolve exceptions | Newer, so governance, audit trails and accuracy on your own data need proving |
Most platforms on the market in 2027 combine several of these layers. A common pattern is exact matching first, configurable rules second, and an AI model or agent handling what is left, with a person reviewing anything below an agreed confidence level.
Six Changes Reshaping Remittance Matching in 2027
1. Software Can Now Read Any Remittance, First Time
For years, the main obstacle to automation was getting data out of remittance advices. Each new layout meant a new template, and a customer who redesigned its remittance broke the process overnight. Large language and vision models have largely removed that barrier. Modern software can extract data from complex remittance advices even when it sees a layout for the first time. For buyers, this changes the question from "which formats do you support?" to "how accurately do you read our formats, and how do you show your working?"
2. Agents That Go and Get the Missing Data
Reading a remittance is only useful if you have it. The newest platforms send agents to collect remittance data from where it lives: logging in to customer AP portals, triaging the shared remittance mailbox, and, when nothing can be found, writing to the customer to ask what a payment covers. When the software cannot resolve a match, it can send the customer an automated remittance request rather than leaving the cash unapplied. This is a real shift, because it moves the work of chasing remittances, one of the most tedious jobs in AR, from people to software.
3. Matching Even When There Is No Remittance
Some payments will always arrive with nothing useful attached. AI models that learn each payer's habits, such as which invoices it tends to pay together, how long it usually takes and which deductions it normally takes, can now propose matches from the payment amount and open-item history alone.
4. Richer Data From the Payment Rails
The move to ISO 20022 messaging is changing what a payment can carry. Under the cross-border standard, up to 9,000 characters of structured remittance information can travel with a payment, enough for invoice numbers, amounts and dates. Swift's cross-border rules require more of each payment's data to be structured from November 2026, and domestic systems in the UK, Europe and the US have been moving to the same standard. The catch is that the benefit only arrives when the payer fills in those fields and your bank passes them through in its statement files, such as camt.053. Ask your banks what they can deliver today, and check that your matching platform can use structured remittance data rather than just store it.
5. E-Invoicing Puts a Clean Reference on Every Invoice
E-invoicing mandates are spreading quickly. France began its phased B2B mandate in September 2026, the EU's VAT in the Digital Age package brings structured e-invoicing to intra-EU B2B trade from July 2030, and the UK government has confirmed that e-invoicing will be mandatory for VAT invoices from April 2029, with Peppol chosen as the exchange network. When your invoice lands directly in your customer's AP system as structured data, the invoice number they pay against is exactly the one you issued. That will not fix remittance matching on its own, but over the next few years it should cut the number of payments that quote a wrong or partial reference. Finance teams planning a remittance matching project in 2027 should ask vendors how their roadmap connects the two.
6. New Ways of Working With AI, and New Ways of Paying for It
Two further changes are worth knowing about. First, AR platforms are opening their data to general AI assistants. In July 2026 one vendor launched an MCP server giving AI assistants such as Claude and Microsoft Copilot read-only access to its invoicing, payments, cash application and analytics data, so a controller can ask questions about unapplied cash in plain English. Second, pricing is changing. another vendor introduced outcome-based pricing in February 2026, with no implementation or subscription fee until go-live and fees then linked to measured results. Whichever model you choose, be precise about how success is measured, because a match rate can be defined in more than one way.
Vendors at the Forefront of Remittance Matching
The vendors below are listed in no particular order and grouped by the kind of approach they take. Every platform here can automate a large share of remittance matching, and the right choice depends on your ERP, your volumes, your payment m,ix and how much of the wider AR process you want to change at the same time.
AI-Native AR Platforms
Monk. New York-based Monk is an AI-native accounts receivable platform covering collections, cash application and forecasting. In July 2026 it launched Cash Application 2.0, built on three matching tiers: customer-specific rules, exact matching, and a background agent that runs daily over transaction and invoice data to handle the cases rules cannot encode. The agent learns from every manual match a person makes and stores that context for each customer, and new lockbox and cheque upload recovers remittance detail that bank feeds often drop. Monk reports an 80% automatic match rate, rising to around 95% once suggested rules are switched on. It raised a $25 million Series A in April 2026 and suits B2B finance teams that want remittance matching run as part of a wider AI-run AR function. Read more in our Monk vendor profile.
Cash Application Specialists
Transformance. Transformance offers ClearMatch, a cash application module built on vision language models rather than OCR and fixed templates. It integrates with SAP, Oracle, NetSuite and Microsoft Dynamics, is typically deployed in four to eight weeks, and is often the first module customers adopt before adding deductions, collections or forecasting. It is a natural fit for enterprise finance teams that want a focused matching engine on top of an established ERP.
Emagia. Emagia provides AI-powered cash application within a wider autonomous finance suite, with its AI assistant, Gia, supporting the cash application team. It connects to more than 170 global and local banks, as well as lockboxes, ERPs, email inboxes and customer AP portals, and captures data from cheque images, lockbox files, PDFs and bank statements in multiple languages and currencies. It is well suited to large, multinational businesses.
Order-to-Cash Suites
HighRadius. HighRadius provides AI-powered order-to-cash automation for large enterprises, with cash application alongside credit, collections, deductions and e-invoicing. It says more than 1,300 companies use its platform and that its cash application delivers more than 85% process automation through AI-based data capture. Its outcome-based pricing model, launched in February 2026, is a notable option for buyers who want fees tied to results.
Billtrust. Billtrust covers invoicing, payments, cash application, collections and credit in one platform, with particular strength in distribution, wholesale and manufacturing. Its cash application uses confidence-based machine learning matching, posting payments automatically and passing only low-confidence matches to a person, and it handles ACH, wire, card, cheque and AP portal payments. Its MCP server, launched in July 2026, makes its cash application data available to AI assistants.
Esker. Lyon-based Esker offers cash application as part of its AR suite, powered by Synergy AI. It triages the remittance mailbox, extracts data from remittance advices and payment files, collects remittances from customer portals using RPA, and learns from user corrections over time. Esker was recognised in The Hackett Group's Digital World Class Matrix for cash application software in September 2025, and its e-invoicing heritage is useful for businesses preparing for mandates across Europe.
Close Platforms and ERP-Native Options
BlackLine. Best known for financial close and reconciliation, BlackLine gained its cash application capability when it acquired specialist Rimilia in 2020. Its Verity Remit agent automates the capture of remittance data, and cash application sits alongside its collections and AR intelligence tools. It is a strong option for organisations that already run their close on BlackLine and want cash application on the same platform.
SAP Cash Application. For businesses on SAP S/4HANA, SAP's own Cash Application service uses machine learning trained on previously cleared items to propose matches between bank statement items and open receivables, clearing them automatically above a confidence threshold the business sets. SAP partners such as Serrala also offer embedded matching inside the SAP interface. These options suit organisations that want to keep matching within the ERP rather than add a separate platform.
For full profiles of the specialist vendors, our Cash Application Software Options 2026 report covers the market in more depth, and our AI Accounts Receivable Software Options 2026 and Accounts Receivable Software Options 2026 reports cover the wider AR platforms.
 |
Build a longlist of remittance matching and cash application vendors that fit your specific systems and volumes with our Longlist Builder powered by HUEY our AI Agent. He's trained on all of our reports, vendor profiles, and lots more!  |
How We Would Approach a Remittance Matching Project
If we were sitting in the buyer's chair, this is how we would run it.
1. Measure where you are. Record unapplied cash at month end, hours spent matching, the share of payments matched automatically today, and how long exceptions take to clear. Note what works well too, so it is not lost in the change.
2. Collect your hardest payments. Build a sample of real payments from the last three months that took the most effort: consolidated payments, short payments, foreign currency receipts, payments with no remittance and payments from unfamiliar payers.
3. Decide how much you want to change. Are you adding matching to an existing AR platform, replacing a cash application tool, or rethinking the whole of AR? The answer changes the shortlist. If you already have a platform, our Stick or Switch Application Review compares improving what you have with moving to something new.
4. Shortlist tightly and test on your data. Share your problem statement through our Technology Matchmaker Service and invite a small number of vendors to run your sample through their platform. A Rapid RFP keeps the process to weeks rather than months.
5. Agree the rules of trust. Set the confidence level above which cash is applied without review, who approves the rest, and what the audit trail must show. Ask your auditors early.
6. Phase the rollout and run in parallel. Start with one entity, one bank or one customer segment, compare results with your current process, and widen the scope as match rates and confidence grow.
If you would like support running the process, our 30-Day Technology Selection takes you from problem statement to a confident decision in a month, and the Enterprise Software Selection Playbook 2026 sets out the same method for teams who prefer to run it themselves.
Remittance Matching Terminology
A few terms that come up in almost every vendor conversation.
Remittance advice: the information a customer sends to explain which invoices a payment covers.
Cash application: the end-to-end process of recording customer payments against open receivables, of which remittance matching is one step.
Unapplied cash: money received but not yet matched to invoices, often held in a suspense account.
Auto-match rate: the share of payments matched and applied without a person, also called the touchless or straight-through rate. Always check how it is defined.
Confidence threshold: the level of certainty above which the software applies a match automatically, with anything below routed for review.
Lockbox: a bank service that receives customer payments, often cheques, and sends payment and remittance data to you as a file.
Short payment and deduction: an amount a customer holds back from what was invoiced, for reasons such as pricing differences, promotions or damaged goods.
ISO 20022: the global messaging standard for payments, which allows structured remittance information to travel with the payment.
camt.053: the ISO 20022 end-of-day bank statement format, which can carry detailed remittance data to your matching platform.
AP portal: a customer's own supplier portal, such as Coupa or Ariba, where invoices are submitted and remittance details are often published.
Common Remittance Matching Mistakes
The first mistake is judging platforms on demo data. Every vendor's matching looks impressive on clean examples. The differences only show up on your own messy payments, which is why we always recommend testing a sample of your hardest cases before a decision is made, and treating any match rate that has not been measured on your data as a starting point for discussion rather than a promise.
The second is automating the matching but not the inputs. Teams invest in a strong engine but leave remittances arriving in five different inboxes, AP portals checked by hand once a week and customer master data full of unlinked payer names. The software then spends its time waiting for data, and the match rate stalls well below what was expected.
The third is setting the confidence threshold once and never revisiting it. Set it too high and people review matches the software would have got right, so little time is saved. Set it too low and errors reach the ledger and the customer. The right level changes as the platform learns, and it deserves a monthly review in the first year.
The fourth is losing the knowledge of experienced matchers. The people who know which customer always pays three weeks late and short by the freight charge are exactly the people who should be training the new platform and reviewing its early decisions. Bringing them into the project from the start protects both the results and the team.
Summary
Remittance matching has been one of the most stubborn manual jobs in finance because the payment and its explanation so rarely arrive together. In 2027, that is changing on several fronts at once. AI can read any remittance, agents can fetch the data that is missing, models can match payments that arrive with nothing attached, and payment standards and e-invoicing mandates are steadily improving the data itself.
The vendors leading the way take different routes, from AI-native platforms such as Monk and Stuut, to cash application specialists, broad order-to-cash suites, close platforms and the ERP itself. The right one for you depends on your systems, your volumes, your payment mix and how much of AR you want to change. Whichever direction you take, start with an honest baseline, test on your own hardest payments, and put the inputs in order alongside the technology.
Talk to Viewpoint Analysis
The quickest place to start is HUEY, the Viewpoint Analysis AI assistant. HUEY is trained on all of our content, including our AR and cash application reports and vendor profiles, so he can explain how the platforms above approach remittance matching, build a shortlist that fits your ERP and volumes, and even introduce you to a vendor's team or arrange a demo.
Compare remittance matching platforms, build a shortlist and arrange demos with the Viewpoint Analysis AI assistant.  |
If you would rather talk it through with a person, we are happy to help, whether you are scoping a first project or reviewing a platform you already have. Just request a call and we'll arrange a time to speak. You can also explore our wider Finance and ERP Technology coverage.
