Author: Richard Oberdieck, PhD
We are seeing a surge in the adoption of AI technologies around the globe among companies from practically every industry. According to a recent survey by Accenture, around two-thirds of businesses are planning investments in AI technologies, and these applications are expected to generate a 30% increase in revenue over the course of the next four years.
Indeed, the implementation of AI technologies – such as robotic process automation, machine learning, and mathematical optimization – has the potential to deliver tremendous return on investment (ROI). However, one of the chief challenges that the “champions” of these AI implementation projects in every organization face is accurately calculating and effectively communicating the business value of their AI applications to stakeholders.
Tackling this challenge is absolutely critical: If you fail to demonstrate the business value of your application and get buy-in from your colleagues, then the AI technology may never truly take hold in our organization.
If you are one of the champions or users of a mathematical optimization application in your company, you have probably encountered this issue and are well aware of the importance of proving the ROI of this technology to stakeholders across functions and levels in your organization.
But how can you ensure that you are able to track and showcase the business value of your mathematical optimization application?
In this blog, I will highlight some of the key steps that you can take to calculate and communicate the business value of your mathematical optimization solution.
Money is – of course – the language of the business world. If the champions and users of mathematical optimization applications (who are often “technical” people such as operations research specialists, computer scientists, and data scientists) want to communicate the value of these applications to the “business” people and other colleagues across their organizations, then quantifying the benefits delivered by these applications in monetary terms (e.g. the revenue growth or cost savings) is important as this will immediately resonate with the “business” people.
Sometimes, this is straightforward. If your application maximizes risk-adjusted investment returns, you clearly can pinpoint the amount of money made. But what about when you are minimizing the amount of employee overtime? Or risk? Or machine downtime? Even in these cases, I strongly urge you to go the extra mile to define and quantify your efficiency gains in monetary terms.
For example, imagine you want to minimize the risk of a machine breaking down in your factory by scheduling the visits by the technician using sensor data. Let’s say that in 2019 each of the 24 breakdowns resulted on average in a four-hour delay of your production, costing USD180,000 per breakdown. If your mathematical optimization solution reduces the breakdown probability from 10% to 1%, then you can expect to only have three breakdowns in 2020, and therefore you will have avoided 21 breakdowns, thereby saving your company USD3.7 million (21 x USD180,000).
As this example shows, sometimes you may need to really to think about the implications of your application from a business perspective in order to come up with a monetary metric that captures the impact of that application on your company’s overall profitability.
Note though that you should always strive to be fair and thorough in your calculations, and – when in doubt – to estimate conservatively. In the example above, we achieved a 10x reduction in machine breakdowns, which translates to 2.4 breakdowns a year. As it is technically not possible to have 2.4 breakdowns (either there is a breakdown or there isn’t), you may be tempted to claim only two breakdowns per year. However, if you are too bullish, you may lose credibility and your project may lose traction due to doubts that you are trying to oversell your application.
In sum, if you want stakeholders across your organization – from the boardroom to the back office and throughout your operational network – to understand the business value of your mathematical optimization application, you need to speak to them in a language that resonates with them (ideally the language of money). This way, you will be able to show them exactly how much the application is contributing to your company’s bottom-line performance.
Calculating and communicating the cost savings or revenue growth delivered by your mathematical optimization application is only one part of the story. To tell the full story of the success of your application, you need to compare the results of that new solution to those of the legacy system.
Prior to the implementation of your mathematical optimization application, your company may have been using manual tools or techniques or a heuristic-based system to address the same business problem. To calculate the true value generated by your mathematical optimization application, you need to determine the difference between what you would have achieved with the legacy system and what you are able to achieve with mathematical optimization.
Let’s say, for example, your company was able to boost annual sales by USD500 million with your mathematical optimization application, but – under the same business conditions – you would have been able to increase annual sales by USD100 million using your legacy system. The delta between the two approaches – USD400 million – represents the true business value of the mathematical optimization application, and this is the figure that you want to report to stakeholders throughout your organization.
The third and perhaps most important step that you need to take is to involve colleagues beyond the technical team in your mathematical optimization project. This is especially important when you are creating the monetary metric to assess value generation, calculating the business value of your application, and communicating the results and ROI of your application to various stakeholders.
Involving your colleagues in this process is important on two levels:
By collaborating with your colleagues to determine the business value of your mathematical optimization application and communicate that value to others, you will be able to build a more compelling business case for expanding the use of mathematical optimization in your organization and get buy-in for that wider roll-out.
Choose the evaluation license that fits you best, and start working with our Expert Team for technical guidance and support.
Request free trial hours, so you can see how quickly and easily a model can be solved on the cloud.