Part of your job is being automated. And why that’s a good thing.
Some learnings about the online accounting software ecosystem.
It’s been one year since my team and I began our journey into the online accounting ecosystem. I’ve been reflecting on our key observations and learnings since we started building our company. I’d be doing our bookkeeper and accountant (B/A’s) partners a disservice if I didn’t share some thoughts on technology advancements.
On a day-to-day basis technology continues to disrupt both personal and professional lives. Many professions are forced to reimagine their business models or eventually become obsolete. Unfortunate for the millions of professionals in the accounting and finance industries, these changes are coming, and faster than you’d think.
Things are changing, and fast
What we know is that, although there is some hype and a lot of “AI garbage” out there, many advancements are further along than imagined. Technology is absolutely automating certain aspects of the job, but it is also going to help B/A’s evolve and break new ground — providing much higher client value than they already are, in less time, from anywhere. Here’s what’s coming.
Tax Filing Automation
Tax return automation is something we will see widespread in the very short term future. When I refer to tax returns, I mean “simple” tax returns which are the majority. Digital systems that aid in the filing of tax returns have been around for quite some time, which means there is a lot of historical data available to train a model using machine learning techniques. If you need a simple explanation of what AI is, please read (AI — An Explanation From a Parenting Perspective).
Assuming the data scientists working with this data are top notch, these systems should provide quality automated results when connected to the key financial systems holding the data for the current tax year.
Complex tax returns whether involving complex corporate structuring or foreign assets will still require the planning and expertise of real professionals. So, if someone is pitching you on a system that automates this complexity, be very skeptical in the foreseeable future. There is a pile of hype out there, so business owners should not be too quick to forgo real professional help.
Transaction Classification
This area might be the biggest disruption to the online B/A ecosystem for those who are simply applying an account name to a transaction that through digital means has nicely flowed into the accounting software.
The reason I say transaction classification is coming fast is because it is not a difficult machine learning problem to solve. The vast amounts of transactional data that Intuit, Xero, Sage, etc have for all of the users on the system allows them to train models that can classify a transaction into an account specific to the user’s chart of accounts, all automatically.
A simple example is an expense transaction for a flight. A model would be able to see a parameter like a flight supplier such as Air Canada or WestJet and know this would be somewhere in the domain of a travel type account expense. Note, there are parameters other than a supplier name that might come into play for classification such as locations, bank transaction id’s, amounts, dates, etc. Just know that there is a lot of data in the background that can help the accuracy of the trained model.
Now of course not every business uses the exact same account names or codes a certain transaction the same way. One business might code a flight to general travel for all flights and another might code a flight to general travel or conferences depending on the purpose of the flight expense. These might seem like complex scenarios, but they are simple in the world of machine learning.
You see, a model would classify the transaction based on previous coding work done by the B/A for that client. So, in essence, the system is being taught with each new transaction that is coded and updates the probability distributions for future classifications. What simplifies the process is that many SMB’s have a large portion of the expenses that are somewhat repeated over time, so the accuracy of the system should be a lot higher as the complexity is a lot lower.
There will be rare or new transactions that come into the system that do need account classification. Even though the client’s model may not have experienced these transactions before, it is probable that at least some client using the same online accounting software supplier has a transaction with parameters that closely match the new one. What you will see in the short term is a “suggested” account classification that will have to be accepted by the B/A or SMB as it would be dangerous to automatically classify a transaction that the system does not have a high enough confidence on. With that being said, it is only a matter of time until the trained models get really good at this.
SMBs do need help, even if they are not willing to admit it
If you are a B/A and the previous few paragraphs have you concerned with being automated out of a job, don’t worry too much as the latest trend I am about to discuss will provide one of many opportunities to reimagine your business and provide an enhanced level of value to your clients.
It is a well-documented fact that hiring a dedicated CFO when your revenues are under a certain threshold makes no sense. What we have noticed is the B/A ecosystem has really taken to this fact and in many cases are trying to reimagine their business model. They are not only just coding transactions or taking shoeboxes of receipts any more, but in a way acting as a part time CFO to the SMB.
This is an incredible development for both the B/A and SMB communities as it allows the B/A to provide value add services to the SMB, and at the same time the SMB gets financial expertise to help them stay competitive with the larger corporations that might be invading their turf.
Win-win for everyone! Sounds great right? Not so fast.
The last company I was a co-founder of produced advanced cardiac image processing software. We hit an early growth curve and had passed a revenue threshold where it made sense for me to take over the CFO role in addition to other duties. In essence, I started out this phase of the company as a part-time CFO.
In the few years I was in this role I can say without a doubt that a CFO who does not understand at least some of the operations of the business is really not that valuable. The CFO’s job is to understand how the finances and operations work together to maximize growth and efficiencies of the business. Just knowing how to code transactions and file tax returns is not what a CFO does. I highly caution anyone in the B/A community to not market yourself as a CFO if you do not plan to understand at least a bit of the operations of your client or else this “reimagining” of your business is really not taking place.
A good CFO provides strong financial guidance, but also empowers the owners and managers of the SMBs to run the business in a much more informed and efficient way. The CFO is the connection where the financial and operational data meet and needs to be the one providing educated insights and recommendations. This might seem quite obvious, but in all of our interactions we have seen a disconnect between the perception of what the client wants vs. what they actually want.
We performed user experience (UX) research early in our company’s history. Out of the 40 SMB’s we talked to, all of them found little to no value in the standard financial statements they would receive beyond just knowing the financials were reconciled and some general revenue and expense information. This might be alarming for some who pride themselves on the statements as a value add deliverable to the client, but at the same time one should realize the opportunity to provide value beyond just the statements.
Easier said than done
The B/As we have been able to interact with this year are truly amazing people and really care about their clients. However, they still have a business to run and cannot spend a bunch of extra hours exporting data, putting it into a spreadsheet, doing complex calculations and trying to make some type of operational report that the user will find value in.
It is not economic to do so even if the B/A really understands the operations of their clients to begin with. So, we run into another disconnect on how to provide operational level value while still being able to run a profitable business.
Technology that will help you evolve (And, why we wake up in the morning)
At chata.ai, we consistently discuss how we can be a catalyst to enable this business model transformation. There are two main areas that need to be tackled by innovation:
1) Information retrieval and reporting needs to be as quick and intuitive as possible.
SOLVED: chata.ai lets you ask a question and receive an answer, no more exports, spreadsheets, or incomplete dashboard-based software.
2) A better understanding of what B/A’s clients are interested in on a day-to-day basis.
SOLVING:
One of the critical components for us to achieve these business model transformations with the B/A community is to get chata.ai in the hands of the SMBs that they work with.
chata.ai is most compelling in the fact that it is dead simple to use. The SMB user requires little to no training with the system other than knowing how to write a text message (which pretty much everyone knows how to do). SMB users are living their businesses on a day-to-day basis. The large majority of SMB’s have questions about their business but often they cannot get answers as the data is trapped and inaccessible in some system they do not know how to use.
If the B/A users had access to see the natural language questions their SMB clients were asking, it would provide incredible insight and allow the B/A to provide a whole new level of value and expertise to the SMB. SPOILER ALERT — a B/A interface for this might be in the works J.
At chata.ai we question ourselves daily on things like “How can we help the B/A community add more value to their clients?”, “How do we solve the day to day problems of SMBs by enabling the usage of their data assets?”, “How do we optimize the time/value curve so that all parties come out ahead?”.
Our ultimate goal is to help our partners be successful. These are complex questions, and like any good innovator should do, we continue to listen to people in this incredible ecosystem, adapt where needed and always, always push our own boundaries.
