AI ROI Calculator

Enter the values below to calculate the Return on Investment (ROI) for your AI solution.

Annual dollar amount of the benefits ($):

Annual dollar amount of uncertainty (optional, $):

Yearly total amount of resources (units):

Average yearly cost per unit ($):

Return per year: $0.00

Total annual investments: $0.00

Total annual ROI: 0.00%

Step 1: Define your goals and how to measure results.

What objectives should this AI project support? For example, reducing operational costs, improving the citizen experience, and reducing response rate, etc.
Pick KPIs to measure that.

Step 2: Establish your baseline.

Step 3: Estimate your benefits and gains in revenue based on the solution.

Examples of cost savings can come from:

Production line optimization
Quality control automation
SLA integration for model design and fit
Gains in workforce productivity (through automation, increased efficiency, etc.)
Increase in revenue from product or service
Time saving costs

Be sure to also factor in:

New production capacities
Additional products developed for citizens or used internally
Number of internal and external government customers and citizens served
The impact of false positives and negatives in classification or recommendation AI solutions
For AI models, account for the cost of errors and impact of errors and number of predictions and value per prediction

Step 4: Identify and estimate the investments and costs of resources.

Be sure to consider:

Task and resources, compute, and storage (including technology needs and tech stack)
SME resources and labor resources
Training and data science/ AI training resources
Data investments

Step 5: Determine what the intangible or soft benefits are:

Reduced burnout and mental health struggles for workforce
Increased positive agency brand for citizens
Increased employee morale
These will not be included in the calculator, but should be included as part of your ROI analysis and decision-making process.

Step 6: Set a realistic timeframe to evaluate the ROI and continuously reevaluate ROI.

Step 7: Develop a status quo analysis.

Step 8: Do the calculation (see calculator below).

Using the calculator below, calculate the difference between your AI solution and the status quo.
Find reputable references and justify references and calculations (a few are included below).

Step 9: Consider ongoing challenges.

Data quality will likely be a problem.
Factor in risks (i.e. reputational risk if there is a strong bias or data breech).
These will not be included in the calculator, but should be included as part of your ROI analysis and decision-making process.

Step 10: Review and adjust continuously.

Reassess your ROI calculation as the project progresses and communicate with stakeholders.
Account for an AI learning curve. For example, models can learn and improve with data and finetuning.
Remember that calculating AI ROI is an ongoing process and it will change over time. For example, AI can keep learning from your data and be prepared to adapt your approach and your AI solution as the project evolves.

Resources that you can use for benefits and costs/investment estimates:

Salaries for various roles: https://www.bls.gov/data/