USING AI FOR BUDGETING

AI is extremely promising technology, but can we use it to make the budgeting process more efficient – and more effective?

The short answer is YES and leading companies like Microsoft already are.

How to get started with AI for budgeting:

• Use AI to budget for “lesser” expenses such as office supplies, subscriptions, or even travel & entertainment.
• Leverage the time saved to develop thoughtful and intentional budgets for key areas like Contractor Spending or Consulting, for example.
• The budget process leaves a lot of money on the table by assuming nobody leaves the organization. So use AI to predict Employee Attrition rates. Apply those rates to assess how much the overall budgeted Salaries & Benefits should be adjusted downward.
• With that money, identity what value-added spending (such as trade R&D, strategic initiatives, marketing campaigns, etc.) can be funded.
• Use AI to develop objective, data driven Sales Projections at whatever level of detail is most appropriate for your business.
• Compare the AI revenue projection with what your Sales or Marketing department delivers. Wherever there is a sizeable gap, have a focused conversation and determine what the forecast should be.
• Continue to track the AI projections and Sales/Marketing projections to see the comparative levels of accuracy over time.
• For Raw Materials budgeting, create a simple grid of Dollar Value on one axis and Price Volatility on the other. For all-but the “High Dollar Value, High Volatility” use AI to project future pricing.
• Use the time saved to develop market knowledge, expertise and negotiating skills for the High Dollar Value, High Volatility materials.

There are plenty of other examples, but hopefully this provide some inspiration.

Great. But how do we actually “do” AI?

The answer depends on if your organization has Data Scientists on staff and are available to work with you. If the answer is “Yes” then work with them to frame up your AI requests and have them deliver it (preferably linking their AI tool of choice to your budgeting/planning application). If the answer is “No” then consider how much basic AI knowledge you may have within the Finance department.

For Finance teams that have people on staff that know popular tools like Azure AutoML for example, you’re in luck. Support their efforts to create forecasting algorithms. In the best of all worlds your budgeting/planning solution you use for the rest of the budget process can integrate with Azure AutoML for example, and have the capability to write algorithms or do regression analysis.

If the answer is “we don’t have in-house expertise in Finance or with Data Scientists” then look for a solution that doesn’t require that.

In other words, something that can be maintained by the business (and not rely on IT).

There are planning systems with built-in AI. For example, I know of at least one that comes complete with more than 50 sophisticated algorithms in the ARIMA family . To forecast Sales for example, you plug in 24 months of historical data. It uses the first 18 months as a “training set”. Then it runs through each of those 50+ algorithms, one by one, to see what each produces as a 6 month forecast. Now since it’s been loaded with 24 months of history and knows what the actual results were, it can then compare what each of those algorithms produced and compare it with the actual results. The one that did the best job (produced the most accurate forecast) can then be used – in a click – to project the future.

I’ve been around long enough to recall a time when large consulting firms would literally bill clients well over a million dollars for a team of consultants to do the same thing. It's somewhat stunning that the same capability is now available to everyone (even without a degree in data science).

For more on this topic, you might want to take a look at KPMG's whitepaper: AI For Finance: Lead adoption for powerful outcomes