Slashing Inaccuracies in A/P Cash Forecasting using AI
An introduction to how AI can help treasurers supersede challenges surrounding accounts payable unpredictability
Factors on which Corporate Treasury’s Success is measured on
A treasurer’s job earlier used to be narrowly defined and focused on managing daily cash positions and liquidity. Today’s treasurer, however, is more involved in devising business strategies and supporting an organization’s operational goals and growth.
In fact, today CEOs are also giving treasurers a seat at the table for strategy meetings and brainstorming sessions. Rapidly evolving technology, new accounting standards, and globalization are just a few of the reasons why the treasurer’s role is expanding to more intensive management of business risks through treasury operations.
While there has been an expansion of the treasurer’s responsibilities, key success factors of a treasurer have also evolved. Nowadays treasurers are not just gauged upon their managerial skills, but also on strategic & forward-looking activities such as:
- Risks Managed/Mitigated
- Accuracy of Cash Forecasts
- Cash Optimization
- Reducing borrowing costs
And among these activities, Accurately forecasting cash is (and always has been) accorded the highest importance since it plays a central role in finance management, balance sheet optimization, protecting investments, improving cash management, and all other activities that steer clear an organization of cash depravity.
However, Cash Forecasting (both direct & indirect) is dogged by challenges that are caused not because of the treasurer’s lack of financial literacy or acumen, but rather because of the lack of visibility into data, process inefficiencies, and most importantly the right technology.
The Main Factors that Render a Cash Forecast Inaccurate
The first step to designing a cash forecast is gathering data; and this data is related to payables, receivables, sales, and return on investments. And considering the scale at which organizations conduct their business, Treasurers are usually inundated with multiple spreadsheets containing this data, which is either extracted from CRMs, ERPs, Billing Management systems, or is received from the finance analysts.
The treasurer then analyses the received data and then employs a spreadsheet-based forecast model that can factor in most of the data and extrapolate a cash position at a future point in time.
However, this method of extrapolation future cash positions has its inherent challenges:
1. Possibilities of Negative Variance: Caused primarily due to the lack of granular visibility into inflows and outflows while using excel as a tool. As remediation, treasurers set up higher cash buffers (either by leveraging short-term debts or utilizing cash pools) to reduce the impact of the variance.
2. Inability to Factor in All the Data: A Spreadsheet is the most used tool to design a cash forecast. However, this tool comes with limitations such as the inability to factor in all the necessary forecast data, since spreadsheets are not designed to do so, which in turn reduces the accuracy of a Cash Forecast
3. The unpredictability of Cash Flow Categories: Along with the limitations created by spreadsheets, the unpredictability of cash flow categories such as accounts payable also adds to the inaccuracy of cash forecasts, since spreadsheet-based models seldom allow deep visibility into the payables data that can be leveraged to build an accurate cash forecast
Of the aforementioned factors that contribute to forecast inaccuracy, the unpredictability of the cash flow categories such as A/P contributes significantly. And this has been cited by 100+ treasurers from fortune 1000s in the 2019 Highradius Cash Forecasting Survey.
It is also important to consider that, the unpredictability of this cash flow category also negatively impacts other factors such as working capital management which is crucial during times of crisis (Covid-19) and long term liquidity – which is an important element in the upkeep of an organization.
The Reasons Behind the Unpredictability of Accounts Payables
Along with the being limited through spreadsheets, Treasurers also have to deal with the unpredictability of accounts payables, while designing a cash forecast. And this unpredictability is owed to the following factors:
Sudden Expenses
During an accounting period, sudden expenses such as breakdowns, an increase in inventory, sudden payments may arise, which contribute significantly to variance. As remediation, treasurers are forced to set up extra cash buffers with an aim to absorb the impact of sudden expenses.
Lost Open Invoice Data
In most of the enterprise’s data associated with payables (invoice data) is stored in systems such as ERPs, CRMs, and Billing management systems. While building a forecast few of these invoices miss getting considered, which results in increased variances, costs of payments, and impact to an organization’s creditworthiness.
Increase in the COGS
Due to fluctuations in market conditions certain expenses such as labor costs, manufacturing costs, and cost of raw materials tend to increase. Which in turn affects the accuracy of the cash forecast models and increases the variance.
Errors caused at record entries
The process of raising a purchase order to issue a vendor invoice is often a manual process and creates a room for errors such as duplication of records and capturing of the wrong invoice amount. And while building a forecast, this incorrect information also gets considered; which results in an increase in the variance and cash buffers.
Handling the A/P Unpredictability and the Way Forward
The Current Remediation in handling the A/P Unpredictability
In order to absorb the unpredictability of accounts payables treasurers set aside a certain amount of cash (cash buffers) by either using cash reserves or by borrowing from financial institutions.
This method of remediation further results in an increase in stagnant idle cash that cannot be used to generate any returns, and also increases costs, since borrowing (if opted as an option) comes with certain additional expenses tied to it.
And if this method of remediation is considered, organizations may have to delay investments that may be crucial for an organization’s progress, since most of the cash is diverted towards making payments that are crucial for an organization’s sustainability. With the current scale of how aggressive businesses are towards expansion, this method of remediation may prove detrimental rather than beneficial.
AI-based Cash Forecasting Applications – The ‘Arsenal’ to Accurate Cash Forecasting
Considering the challenges and their expensive remedies it is imperative that the treasury function adopts a solution that can help treasurers in their initiatives related to cash forecasting. And from the many, one such holistic solution is Artificial Intelligence (AI).
To the uninitiated, AI is a widely popular branch of computer science that builds solutions to intelligently automate key tasks that are heavily manual, error-prone, and taxes. In this context, for the treasurer, AI can automate the process of cash forecasting and help in improving cash & liquidity management for that forecast period.
To build an accurate cash forecast AI-based cash forecasting applications initially have to be integrated (which is done by the vendor offering the solution) with Cash Management systems, Billing Management Systems, ERPs, Treasury Management systems, and any other systems that house data related to cash inflows and outflows.
After integrating with these systems through an API, the AI based cash forecasting application fetches data automatically into its database. It is also important to note that at this phase the data is mostly unstructured and requires a further level of cleansing to make it structured, which is performed by the application, resulting in creating a structured database that has relationships mapped between most of the entities within data.
Upon structuring data, the AI-based application would then automatically choose the best-fit algorithm that can consider all the entities within the data and design a forecast that has the least variance and is extremely accurate.
For instance, considering a use-case of building an Accounts Payable cash forecast, the AI-based cash forecasting system would fetch all the data related to payables into its database. And after fetching the data the AI-based cash forecasting application would then structure the data and run an algorithm that could accurately predict the payments that may arise in the forecasting period.
A simple under the hood representation of how Highradius’s ai-based cash forecasting tool predicts accounts payable
Deep Dive: How AI helps the Treasurer Predict the Payables of Future and Slash Forecast Inaccuracies.
After structuring the data, the AI-based cash forecasting application does a simple historical analysis that allows it to retrospectively identify trends and payment patterns. These trends and patterns would include data points related to:
- Historical instances of which forecasting period had sudden expenses and high variance, along with the identification of what the reasons were
- What were the categories of the payments that suddenly cropped up and the reasons behind it
- Identify instances of late payments, and the associated penalties paid
- Identification of payments made on time and the goods/services purchased on the purchase order
- Taxes paid during an accounting period
- Historical invoices booked during a certain period
- Investments made during the accounting period
- Repayments related to debt and identification of payment patterns
Post the historical analysis the AI-based cash forecasting application would correlate historical data with the most recent data and perform a scenario analysis using multiple AI algorithms. The algorithm that presents the most favorable and realistic cash forecast is then chosen automatically to deliver an accurate A/P Forecast.
This automated method of cash forecasting performed by the AI-based application helps the treasurer predict the expenses that may arise in the particular forecast period, foresee any particular fluctuations in costs and predict invoices that can be booked in a certain forecasting period.
The treasurer is also at a benefit leveraging this information since he will be able to reduce the cash buffer, optimize cash and related assets, and improve the creditworthiness of the organization.
Conclusion: How The Modern Treasurer stands to benefit from AI-based A/P cash forecasting
After structuring the data, the AI-based cash forecasting application does a simple historical analysis that allows it to retrospectively identify trends and payment patterns. These trends and patterns would include data points related to:
It goes without saying that Cash forecasting is indeed crucial to an organization and its importance is only going to increase with time since businesses are always striving to progress, expand, and innovate. And at the center of these activities lies ‘cash’, so the more accurate a cash forecast is, the better is cash optimization and progress, which is only possible through new-age applications such as AI.
Along with the benefit of cash forecasting accuracy AI-based cash forecasting applications augment a treasurer’s workday in the following ways
- Improves financial management
- Enables deep visibility into inflows and outflows
- Reduces the burden of maintaining excessive cash buffers
- Helps the treasurer provide strategic support to the CFO’s office
About HighRadius
HighRadius is a Fintech enterprise Software-as-a-Service (SaaS) company which leverages Artificial Intelligence-based Autonomous Systems to help companies automate Accounts Receivable and Treasury processes. The HighRadius® Integrated Receivables platform reduces cycle times in your order-to-cash process through automation of receivables and payments processes across credit, electronic billing and payment processing, cash application, deductions, and collections. HighRadius® Treasury Management Applications help teams achieve touchless cash management, accurate cash forecasting and seamless bank reconciliation.
Powered by the Rivana™ Artificial Intelligence Engine and Freeda™ Digital Assistant for order-to-cash teams, HighRadius enables teams to leverage machine learning to predict future outcomes and automate routine labor-intensive tasks. The radiusOne™ B2B payment network allows suppliers to digitally connect with buyers, closing the loop from supplier receivable processes to buyer payable processes.
About Treasury Solutions
The HighRadiusTM Treasury Management Applications are a suite of world’s first AI-powered solutions designed to support treasury teams across all industries by automating and enhancing their cash forecasting, cash management and bank reconciliation processes.
The HighRadiusTM Treasury Management Applications are unique in the approach that they are powered by an Artificial Intelligence technology created exclusively to redefine the forecasting, bank reconciliation and cash management processes, so that treasurers spend lesser manual effort but extract better outcomes such as making more accurate cash forecasts across all cash flow categories, increased forecasting frequency and variance reporting, gaining instant visibility into real time cash positions across bank accounts at any level and achieving 99%+ straight through reconciliation of bank statements without human intervention.
The solutions are delivered via cloud which enable them to be seamlessly integrated with multiple systems including your ERP, TMS, accounting systems and banks instantly and simultaneously.
