Prioritizing Artificial Intelligence use cases - Calculating Return on Investment (ROI) - Part 2

This post is continuation of my previous post on Identifying and Prioritizing AI use cases. In previous post we emphasized need for identifying use cases that can impact organization strategic goals and deliver significant business value. Link to previous post below

One challenge with AI projects compared to typical software projects is uncertainty on the outcome, possible longer time to experiment than actually accounted, complexity in integration with business process, training deployment skew among others. This makes it difficult to size an AI project either for budgeting or to measure the return on investment

In my previous post, taking uncertainty of AI projects into consideration we introduced below formula for calculating ROI (Refer previous post above for details on individual formula components)

Finally we indicated in Part 1, next series is all about calculating ROI in excel. So here we are

I am not going to walk through each and every sheet in Excel but just going to provide intuition on various parameters used to calculate ROI. The excel template is available in GitHub repository below for anyone and everyone to take, modify as one wishes and use it

Note: The Excel template is not an policy or standard or procedures but an approach taking into factor uncertainty and complexity of AI projects. Anyone can use the template to their need and convenience. Feel free to add/delete parameters, change weightages, modify to your enterprise cost structure among others. In case if you already have template for measuring ROI of AI projects you can skip the remaining section

This has been one of my toughest article to frame and today I am glad to bring this "AI Value Realization Framework" to life and hope it will accelerate and provide intuition into calculating ROI for budgeting and to further prioritize AI use cases based on business value

Who is this "AI Value Realization Framework" for?

• CxO or Head of AI or AI Champion within organization who want to size the opportunity of AI use case for budgeting and understand expected business benefit over time
• AI consulting and service organization that are in business of accelerating AI journey for their customer and looking at way to come up with high level pricing to size AI opportunity
• AI product startup or boutique AI firms using this to calculate ROI over time horizon to position their product or services

Framework in GitHub repository contains detailed notes on each and every tab to walk you through the calculation process. I will summarize some of the core work sheets below

Return on Investment worksheet contains summarized information from different tabs in Excel taking into consideration Total Project Cost as well as Expected Business Benefit

ROI is calculated over 3 years horizon as first year is typically revenue negative due to high cost of engineering AI use cases. ROI duration below can be extended or reduced based on your organization AI strategy

Let us go into couple of worksheets that feed value into ROI calculation. Let us start with Implementation cost sheet which captures one of the major expense in AI projects

AI projects are typically very expensive to implement and picking a problem that generates business value is key

In Implementation cost sheet, first thing is to categorize AI projects based on their expected complexity. It is important for enterprise to have high level view on what resources are required and when these resources have to be brought into the project. Say Data and ML Engineer might be required from the start of projects where as BI Engineer might be required half way into the project. Allocated percentage is used to highlight their role duration in the AI project

Note: Change Blended cost to the internal cost structure of your enterprise. Above is just for illustrative purpose

Next we will look at Complexity Factor sheet

In this sheet we highlight various challenges implementing AI projects and subsequently weight them relevant to one that matters to the enterprise. Based on AI project under implementation we assign relevance value to calculate overall complexity factor.

Complexity Factor is in turn used to calculate implementation cost by baking in possible cost on account of project complexity

Total Implementation Cost = (Implementation Cost*Complexity Factor) + 
                            Infrastructure Cost + Operations Cost + 
                            Product purchase cost

You can find most of the other calculation in the framework template available in GitHub repository.

In case if you have any queries, suggestion or issue please use the comment section below or raise a Issue in the associated GitHub repository

Once you have the ROI for individual use cases under consideration, you can prioritize based on either ROI alone or a combination of different parameters. Have listed few of the parameters below

I hope this template is helpful to get started in your AI budgeting and business benefit calculation journey