The AR Problem Nobody Talks About: Managing Collections Across Dozens of Client Files

Jul 10

Part 3 of the Lumiere AI Stack Series. Start with the pillar post if you haven't read it yet, or catch up on Post 2 (AP automation).

Most AR automation content is written for a company with one set of customers, one aging report, and one collections inbox.

That is not your firm.

An outsourced accounting practice manages AR for clients. You are not collecting money that your firm is owed. You are building and managing collections workflows on behalf of 10, 20, or 50 different businesses, each with their own customer base, invoice format, payment terms, and tolerance for follow-up. You are doing this across multiple staff members, often in QBO or Xero environments that were not built with multi-client operations in mind.

The AR automation tools built for enterprise finance teams solve a different problem. The tools worth your attention in 2026 are the ones designed around the multi-client delivery model, or at least flexible enough to run inside one.

This post covers what those tools are, how the workflow actually runs, and what a realistic rollout looks like for a CAS practice that wants to offer AR automation as a service rather than just pass the problem back to the client.

Why AR Is Harder Than AP for Outsourced Firms

AP automation is largely an inbound problem: invoices come in, get coded and approved, and get paid. The firm controls most of the workflow. AR is an outbound relationship problem: invoices go out, and then you wait, follow up, chase, negotiate, and reconcile what eventually comes back in. The other party (the client's customer) is outside your control entirely.

That asymmetry creates three problems that are specific to outsourced firms:

Client authority issues. Before you can send a collections email on a client's behalf, you need to know their voice, their payment terms, their customer relationships, and whether certain customers get special treatment. This context lives in the client's head, not your system. Every new client engagement requires a structured onboarding conversation before you can automate anything.
Aging across files. When you are managing AR for multiple clients, you need visibility across all of them simultaneously. Which clients have significant overdue balances? Which have customers who are 90 days out? You cannot answer those questions by logging into five separate QBO accounts one at a time. You need a consolidated view that does not exist natively in the accounting software.
Cash application at scale. When payments come in across dozens of client files, matching them to open invoices without errors is a manual, time-consuming task. A partial payment on a multi-line invoice. A wire with no remittance. A customer who pays three invoices in one check with a handwritten note. Each of these requires judgment. At low volume, you handle it. At scale, it becomes a bottleneck.

The right AR automation stack addresses all three. Most tools in the market address one.

The Tools Worth Knowing in 2026

The AR automation category has consolidated significantly over the past two years. Here is where the relevant options sit for a CAS practice.

BILL (for collections management)

Most outsourced firms already have BILL in their stack for AP. What fewer firms use is BILL's AR module, which handles invoice delivery, payment reminders, and online payment collection. The advantage is the integration: if you are already using BILL for a client's AP, adding their AR creates a single platform for the full payables and receivables workflow.

The limitation is depth. BILL's AR functionality is designed for simplicity, which works well for clients with straightforward invoice-to-cash cycles and limited collections complexity. For clients with high invoice volume, layered payment terms, or a customer base that requires active collections management, you will outgrow it.

Best fit: Smaller clients with clean AR, low invoice volume, and a preference for a single-vendor stack.

Lockstep (now part of Sage)

Lockstep is the most CAS-native AR tool in the market. It was built around the concept of connected accounting, meaning both sides of a transaction (the AR team and the AP team) working in a shared environment. For outsourced firms, the key differentiator is multi-tenant architecture: you can manage AR across multiple client files from a single dashboard, segment by client, and run separate collections workflows for each without switching logins. Sage acquired Lockstep in February 2024, which means the roadmap is now tied to Sage's broader product strategy. For firms already in the Sage ecosystem, that is a natural fit. For QBO- and Xero-native practices, it is worth evaluating whether the integration depth still meets your needs before committing.

Lockstep integrates with QBO, Xero, Sage Intacct, and 40+ other accounting platforms. Collections automation is rules-based but flexible: you configure reminder sequences by customer aging bucket, set escalation triggers, and route exceptions to staff by client. The platform also handles online payment acceptance and cash application, which closes the loop without requiring a separate tool.

Best fit: Firms offering AR management as a defined service, with multiple clients and staff organized around client-facing work, particularly those with Sage Intacct in the client stack.

Gaviti

Gaviti is a collections-focused AR tool with strong workflow automation and a clear UI built for teams doing active outreach. Where Lockstep emphasizes the multi-entity architecture, Gaviti emphasizes the collections process itself: task management for collectors, templated communication sequences, dispute tracking, and performance dashboards by collector and by customer.

For a firm where a staff member is actively managing collections as part of their client responsibilities (rather than just configuring automation and letting it run), Gaviti's workflow layer is more sophisticated than most competitors at the same price point.

Best fit: Firms where AR management involves significant active collections work, not just automated reminders. Gaviti works particularly well when you have a dedicated AR staff member covering multiple clients.

Versapay

Versapay takes a different angle: it focuses on the customer-facing experience, specifically giving the client's customers a portal where they can view invoices, dispute line items, and pay online without a phone call or email exchange. The pitch is that a large portion of late payments are not delinquency, they are friction. Customers who cannot easily see what they owe or pay conveniently will delay.

The platform integrates with major ERPs and has strong cash application automation for remittance matching. The tradeoff is that Versapay is priced and positioned for mid-market clients with meaningful invoice volume. It is less relevant for a client with 20 invoices a month than for one with 500.

Best fit: Clients with higher invoice volume, a B2B customer base that would benefit from self-service payment, and enough transaction volume to justify the platform investment.

QBO/Xero

Both QuickBooks Online and Xero have built-in invoice reminders and payment links that many small clients can run adequately with minimal configuration. For clients with simple AR, low invoice volume, and customers who pay reasonably on time, the native functionality is often enough and the right recommendation is not to add another tool.

The failure mode to watch for: firms that configure native reminders and then treat that as AR management. Native reminders are not collections. They do not track dispute status, do not give you a consolidated aging view across clients, and do not handle cash application. Know what you are getting before you call it done.

The Outsourced AR Workflow, Step by Step

Here is what a well-structured AR automation workflow looks like inside a CAS practice in 2026. This assumes Lockstep as the primary platform, but the sequence applies broadly.

Step 1: Client onboarding and configuration. Before you automate anything, you need the client's input on a few decisions: What are the payment terms by customer? Are there any customers who should never receive automated reminders (key relationships, current disputes, prepay arrangements)? What is the firm's preferred tone for collections communication? Who approves escalation to a formal collections notice? This conversation belongs in your new-client onboarding process, not as a one-off when you start the AR work.
Step 2: Invoice delivery. Invoices generated in QBO or Xero sync to the AR platform and get delivered to customers via the platform's delivery system, which creates a trackable record of when the invoice was sent, opened, and viewed. This is a meaningful improvement over emailing PDFs from a staff inbox: you can see which invoices have been opened and which have not, and build your follow-up sequence accordingly.
Step 3: Automated reminder sequence. Configure a reminder cadence by aging bucket. A common structure: a friendly reminder at 7 days past due, a firmer follow-up at 21 days, and a staff-triggered escalation at 45 days. The automated messages run without manual intervention. Staff attention is reserved for exceptions and escalations.
Step 4: Exception handling. Customers who dispute an invoice, request a credit, or indicate they need a payment plan get flagged out of the automated sequence and routed to a staff member. The platform tracks dispute status and ensures nothing falls through the cracks while the exception is being resolved.
Step 5: Payment and cash application. When a payment comes in, the platform matches it to the open invoice. For straightforward payments, this is touchless. For partial payments, split remittances, or wires without remittance detail, the platform surfaces the exception with suggested matches for a staff member to confirm. The goal is not 100% touchless matching; it is reducing the manual matching workload by 70 to 80 percent, which is realistic with current tooling.
Step 6: Consolidated reporting. At any point, the firm principal or the assigned staff member can pull a cross-client aging summary: total AR outstanding by client, overdue by bucket, exceptions in queue, and payment trends. This is the view that does not exist natively in QBO or Xero and is the primary reason a multi-client AR platform pays for itself.

What AI Is Actually Doing Here

It is worth being direct about where AI adds genuine value in AR versus where it is marketing language.

Cash application matching is the clearest AI use case. Machine learning models trained on historical payment patterns can match incoming payments to open invoices with meaningful accuracy, even when remittance data is incomplete or nonstandard. This is real, measurable, and the primary reason cash application automation has improved so substantially in the past two years.
Predictive collections prioritization is emerging but still early for small-business AR. Enterprise tools from HighRadius and similar platforms can score customers by payment likelihood and surface the accounts most likely to go delinquent before they do. For the client base most CAS firms serve, the predictive layer is not yet well-calibrated, and a simple aging report with consistent follow-up often outperforms a probabilistic model built on limited transaction history.
Reminder drafting is the area where AI is least impressive, despite frequent vendor claims. The value in AR communications is not the text of the reminder. It is the timing, the sequencing, and knowing when to take a human off the automation and into a direct conversation. AI-drafted reminder copy is fine. AI-managed collections sequences are not meaningfully different from well-configured rule-based sequences for most CAS clients.

Take a position internally on what you are selling. If you are offering AI-powered AR management, be clear about which part of the workflow the AI is actually running and what the firm is still doing manually. Clients and prospects will ask.

Building AR as a Service: The Pricing and Scoping Question

AR management is a natural value-added service for outsourced firms, but it is priced and scoped differently than close support or bookkeeping. A few things to get right before you offer it:

Scope it by outcome, not by task. Clients do not want to buy "AR software setup and weekly aging review." They want to reduce their DSO and stop chasing customers themselves. Scope the service around a target outcome (for example: all invoices delivered within 24 hours of creation, first reminder sent within 7 days of due date, monthly aging review delivered with commentary) and price against the value of that outcome.
Set expectations on what the firm controls. AR automation reduces friction and increases follow-up consistency. It does not make customers pay faster if they are slow payers by choice. Make sure clients understand that the firm manages the process, not the outcome, and that some customers will still require direct intervention regardless of what the automation does.
Platform cost passes through or bundles. Lockstep, Gaviti, and Versapay each carry their own subscription costs. Decide early whether you will pass these through to clients (cleaner for accounting purposes, sometimes harder to sell) or bundle them into a service fee (simpler for clients, requires careful margin math on your side).

The Rollout Sequence

If you are adding AR automation to your service offering, here is a realistic sequencing for a four to six week rollout with a pilot client.

Weeks 1 and 2: Select a pilot client with moderate AR complexity (not your most complicated, not your simplest). Complete the onboarding conversation. Configure the platform, connect it to their accounting file, and set up the customer list, payment terms, and reminder sequences. Deliver invoices through the platform for the first time.
Weeks 3 and 4: Run the first full collections cycle. Monitor the exception queue daily. Document every edge case that requires manual intervention. These become your playbook for future clients.
Weeks 5 and 6: Review the aging summary with the client. Measure: What percentage of invoices were paid without manual follow-up? What was the average days-to-payment compared to the prior period? What exceptions came up and how were they handled? Use these numbers to refine the service scope and build the case for rolling out to additional clients.

The most common failure mode in AR automation rollouts is skipping the onboarding conversation and configuring the tool based on assumptions about the client's customers and preferences. This produces reminder sequences that do not match the client's voice, customers who get automated messages that conflict with verbal agreements the client already made, and a first impression of the service that requires damage control. The configuration step is not optional.

The Owner Takeaway

AR management is one of the highest-value services an outsourced firm can offer because it directly affects the client's cash position in a way they feel immediately. It is also one of the most underpriced services in the market because most firms either do it inconsistently or treat it as a bookkeeping task rather than a managed service. A firm that can demonstrate a reduction in DSO or a measurable improvement in collections consistency has a concrete, quantified service outcome to sell, which is a different conversation than "we manage your books."

The platform investment is modest. The margin at the right service price is strong. The barrier is the operational design: you need to build the onboarding process, the exception-handling workflow, and the reporting cadence before you can scale it across clients.

The Operator Takeaway

The highest-leverage thing you can do is build a standard AR onboarding checklist that captures the inputs you need before touching any client's AR file: payment terms by customer, exception customers, communication preferences, escalation triggers, and the name of the person at the client who has authority to approve write-offs or payment plans. This document becomes the configuration layer for every AR automation setup you do. Without it, you are making assumptions that will eventually produce an awkward conversation with a client about why their biggest customer got an aggressive collections email.

The second-highest-leverage thing is to get the consolidated aging view running across all AR clients as soon as possible. Once you can see all AR exceptions in one place without logging into individual files, your response time improves and nothing goes unnoticed at month-end.

What Comes Next in the Series

Posts 2 and 3 have covered the client transaction layer: AP and AR. Next in the series, we move to Layer 02, the per-client close and review workflow, where the tools are Keeper, Numeric, and the close management stack that lets a firm run a consistent, documented close across every client file without rebuilding the process from scratch each month.

Back to Post 1: The AI-Enabled Close | Post 2: AP on Autopilot

Resources

The Lumiere AI Stack Map for Outsourced Accounting Firms (PDF download) -- the four-layer visual referenced throughout this series
Two-week time study template -- track where your firm's hours are actually going before you automate anything