Artificial Intelligence (AI) is the buzzword in business these days…
Will AI replace human accountants?
The short answer – no.
Why?
Because AI cannot flag errors, recode, interpret data correctly all the time, nor communicate with emotional understanding.
By the way, did you know that there are three different types of AI?
They are:
- Machine learning – this is where software recognizes patterns in data.
- Predictive AI – this is where the software uses those learned patterns to suggest actions to users.
- Generative AI – where software uses learned patterns to create new content, i.e. writing text and creating images.
AI in accounting is largely of the Predictive type. It is used to extract data from source documents, and to suggest actions (in coding and matching transactions).
Let us explore how AI is used in the suite of software we use for our clients at ControllershipPLUS.
Source Document Extraction
Here in our front-end software called HubDoc AI goes to work in extracting the core details from supplier bills, in addition to other source documents.
Within seconds of a supplier bill being uploaded or emailed to HubDoc the key details are extracted automatically:
- Vender name.
- Date due.
- Total amount.
- Taxes, including GST, PST, and HST.
HubDoc, based on predictive AI will suggest the account code to code the expense to. It will suggest this based on the past coding of that supplier.
For instance, Telus, a large telecom provider in Canada will usually be coded to “telephone” expense, if that is where you coded it in the past.
However, if the Telus bill is not for a telephone expense, rather it is for cell phones purchased, it may need to be coded somewhere else, like a capital account, “Telephone equipment.”
Human intervention is needed at this point because although predictive AI is good at extraction and making suggestions, it is not smart enough to know what to do with a new situation!
A highly intelligent accounting technician is needed to review and watch for exceptions and recode the suggested coding by AI.
In accounting, our human technicians are definitely much smarter than the AI.
Where predictive AI shines here is in the time saving of mind-numbing data entry tasks.
Okay, let us head on over to Xero next, where these extracted documents have been sent to within these now coded transactions….
Xero – Here AI Makes Suggestions
Inside Xero, the predictive AI matches transactions fed into it directly from the bank feeds.
It suggests which transactions likely match what went through your bank accounts.
The risk here in relying on AI is that the AI is not that smart. It only can suggest a match. If there are multiple transactions with the same dollar amount if you blindly accept the suggestion from AI, you may be wrong.
Here again, human intervention is required to manage and review, and not unconsciously clicking the “OKAY” button.
Xero and Accounts Receivable
Inside the bank feeds, Xero’s AI will suggest matches of deposits. This saves time in looking up the correct transactions to reconcile.
Unpaid deposits remain in aged accounts receivable.
Xero sends reminders of outstanding invoices to your customers and clients. I am not sure how this can rightly be called AI as it is just a routine, fixed scheduling task. Nevertheless, it does save time for the accounting technician.
Bills approved for payment now are sent over to Plooto…
AI and Plooto
The key features of AI in Plooto are:
- Two-way synchronization. Bills are automatically synced to Plooto from Xero, along with all attached source documents. Once they are paid in Plooto, they are synced back to Xero and recorded as a payment against the bill just paid.
- Plooto allows for customizable approval processes. These save time by reducing chasing people physically in different locations to sign cheques.
- Plooto leverages AI to encrypt all date keeping everything tight and secure.
The predictive AI here is not really thinking for you in a way that you might imagine AI to be working. It does, though, save accounting technicians from doing boring, repetitive tasks.
Fathom and Smart Prediction
Our high-end reporting software, Fathom, picks up all the month-end date from Xero on a regular, automated basis.
Inside of the forecasting module of Fathom it uses what it calls Smart Prediction to predict future revenue and expenses when doing a forecast.
Again, highly intelligent accounting technicians are required to intervene and not assume that the software’s predictions are written on stone tablets!
In Summary
Predictive AI really has stripped many mundane, repetitive tasks from accounting technicians. It has freed them up to add true intelligence to the mix. To analyse and override and to ensure that transactions are coded correctly, and that cash-flow projections make sense based on our knowledge of our client’s businesses.
Thank you for reading…