AI Tool Adoption Checklist

This checklist is designed to be a general guide to follow when assessing AI tools.

• Does the task or process that you are looking to augment with AI take enough time to accomplish that it is worth implementing a new tool?

• Do the staff who will be impacted and/or involved in the tasks or processes associated with the AI tool have a basic understanding of what AI is and how its adoption will further the organization’s objectives?

• Are there staff who are willing and able to own the implementation and help colleagues get up to speed with the new AI tool?

• Is there a budget in place to pay for a trial, and then full adoption, of the tool?

• Is there the ability to test or sandbox the tool before launching it throughout the organization with a timed review after an appropriate amount of time?

• From a data privacy perspective, how sensitive or high risk is the organizational or customer data needed to complete the task or process? Does the tool have sufficient data privacy controls in place?

• If the task or process involves an opportunity for biases (gender, race, age, etc.) to influence outcomes such as in customer service and HR, has the vendor demonstrated that they have sufficiently addressed issues of bias?

• What is the level of accuracy outputs is required in the task or process? Is it important for the AI tool’s output to be right all the time?

• Is it easy to check the outputs of the AI tool for accuracy?

• Is the AI tool scalable enough to meet the needs of the organization? Does the vendor have the capacity and structure to provide customer support?

• Does the AI tool integrate with, or replace, existing systems and workflows? How disruptive could the implementation of the tool be for the organization?

• If we do not implement an AI tool into the task or process in question how incentivised are staff to use unsanctioned AI tools in the task or process?