The Innovation and Digital Transformation areas of large companies face the challenge of justifying the projects in which they invest. In the case of Artificial Intelligence or Advanced Analytics projects, the challenge is even greater due to the large number of use cases for which considerable operational savings or higher revenues can be achieved.
To prioritise those projects where the business impact is most significant, the following should be taken into account:
- ROI: Return on Investment.
- TTM: Time to Market.
- Maturity of the Technology.
In the case of AI projects, it makes more sense to focus on ROI if you want to move quickly because companies are looking for process improvements over how they are doing it now. These process improvements must produce measurable financial benefits.
On the other hand, the prioritisation of projects by TTM has a strategic component that goes beyond purely quantitative calculations.
As for the maturity of the technology, just say that there are so many use cases where sufficiently mature AI techniques already exist that there is no need to venture into more R&D projects, except for a few for experimentation or strategic reasons.
For the above reasons, we will focus on ROI. Each situation and company is unique, so the method for estimating ROI is also very different. However, in general, these could be the steps to do this estimation, with adaptations for each case:
1- Estimation of Expenditures or Revenues in an Ideal Scenario
One way to do this estimation would be to choose several past situations and calculate how much the company would have spent or earned if the decisions or processes had been optimal.
Based on these calculations, the expenses or revenues for a particular period are estimated. To make this assessment, objective criteria are used to relate the sample of data calculated to the estimated value for the whole period. There are cases where costs or revenues can be calculated without estimating, but this is not very common.
2- Analysis of Historical Data on the Company’s Actual Costs or Revenues
The actual expenditure or revenue earned in the same period used in the estimation in step 1 is summed.
3- Calculation of Expected Profits
Using the same period, the profits for the company are calculated as follows:
- Savings: Profit = Result step 2 – Result step 1
- Increased Revenues: Profit = Result step 1 – Result step 2
4- Calculation of ROI (and payback)
Generally, the development of an AI application is an investment, but the savings or revenue increases are undefined if the business environment does not change. That is why ROI is calculated over a period of time, e.g. for one year. The equation is:
ROI: ((Profit – Investment) / Investment) x 100
Where “profit” is the savings or increase in revenue in the chosen period calculated in step 3, and investment is the total cost of developing the application.
A very effective way to express ROI is by calculating payback. This is the period in which the application “pays for itself” with the profits made.
This value is usually the most appropriate indicator to express the return on investment for Artificial Intelligence and Advanced Analytics applications, where it can be expected to be “financed” in a few months, weeks, or even days of use.
The payback calculation is very simple: for example, if we want to calculate how many weeks it takes to “pay for” the application:
Payback = Investment / Average profit in a week
Any investment in Artificial Intelligence or Advanced Analytics projects should be prioritised by estimating the ROI or Payback.
There are use cases for which it is not worth starting a project because of the low return, but there are many others in which money is lost or money is not earned if Artificial Intelligence solutions are not developed to serve the business.
Want to know more about ROI?
