Better Accounting Means Better Business

What are the biggest barriers for brokerages when it comes to revenue accounting?

Between recording deposits, gathering commission statements, reconciling millions of transactions against disagreeing AMS policy data, it’s a labor intensive, manual process done with legacy tools, exception spreadsheets, and PDFs. Acquisitive agencies encounter additional challenges before, during, and after migrations, as they often transact across multiple agency management systems.

This process becomes even more challenging at scale as the ratio of finance and accounting personnel increases. When you’re talking about dozens or hundreds of folks recording hundreds of millions or billions in revenue with the necessary controls at that level, a further resource-intensive coordination layer is required.

What downstream impacts of this manual process have you seen outside the accounting process?

The fractious accounting workflow causes downstream operational challenges, beyond agencies’ and producers’ not earning everything they’re owed. One example is the loss of compensation detail from carriers, particularly in the employee benefits arena. There’s per-employee-per-month in addition to percent-of-premium, among other structures. Many accounting teams are so underwater they don’t record this detail at all, or there are so many hands in the workflows the data isn’t consistent or trusted. Meanwhile, carrier relations teams spend tremendous effort negotiating with their preferred carriers but can’t actually validate that the agreements are being followed.

Another example are bonus and override programs, which for many drive 10-15% of their benefits revenue. Yet practice leaders often say they’re “flying blind” when understanding what bonuses they’re receiving, what they’re for, and where they’re likely to land on persistency and other metrics. This revenue is often non-producer commissionable, so optimizing is uniquely valuable: every dollar is driving up to 15x of enterprise value.

What have been some obstacles for brokers in undergoing this revenue transformation?

It takes a few weeks for some people to shrug off old habits from decades-old instincts they’ve developed to do so much manually. Interestingly, a top-30 brokerage CFO recently told me they were stunned that, within a month of signing, a majority of their revenue reconciliation was already being automated via Comulate. They said it was the first time the decision and demo process was more work than the implementation.

To me, that speaks to a broader reality in our industry, in which the back office hasn’t been the beneficiary of high-quality software that actually delivers—no less on their most complex workflows. Part of our role becomes rebuilding their trust in technology partners, especially when folks have previously tried RPA [robotic process automation] that failed to deliver value.

Given the complexity, what does Comulate’s solution for this problem look like?

Comulate’s foundation is a purpose-built engine that digitizes messy financial documents like direct bill statements and checks, capturing transactional data. Given its focus, Comulate already understands 99% of a broker’s statements at onboarding. A proprietary reconciliation layer then automatically matches 90%+ of transactions or payments to their integrated AMS, with dedicated exception workflows. It’s fulfilling to hear when some accounting teams call it “magic” after they post reconciliations from Comulate to their AMS’s ledger, considering the hundreds of little complexities and nuances the team has solved along the way.

The revenue intelligence is powered by models that leverage high-quality, consistently captured, transactional data. Specialized visualization enables instant drill-downs on large volumes of data to answer questions about variances, compensation metrics, and account growth.

How does the underlying automation work? Is it similar to a large language model (LLM)?

Our secret sauce is the models we’ve built internally, which I’d call “light AI” or “controlled AI,” which are overridden by robust guards to ensure revenue integrity.

LLMs shouldn’t be used for most financial processes. The system must use the data provided and produce consistent output given the same input, without hallucinations or generating new content. I think most controllers, chief administrative officers, and chief financial officers understand LLMs cannot be used for financial processes but are exploring them for areas of the business that don’t require 100% accuracy.

Are there any cybersecurity concerns brokerages should be aware of when using this kind of software?

Interestingly, some broker leaders say a reason they approach us is because it reduces their risk. If offshoring, they have dozens of non-employees on other continents that have broad access to their corporate systems. Many large firms prioritize audited, controlled, enterprise-grade software. The resiliency, predictability, and accuracy of solutions that can do almost all the work without being affected by someone’s vacation or retirement are also important to chief information officers.