As AI adoption accelerates across industries, many organizations are rushing to implement these technologies without fully considering the foundational elements required for success. In my 25+ years in information governance, information management, and project management, I’ve observed organizations become enamored with new technologies without first addressing the fundamental elements needed to support them.
Before implementing AI, organizations must assess their readiness across multiple dimensions. In 2024, AIIM released an AI readiness assessment tool. The tool provides an excellent framework for evaluating and helping organizations identify what they need to have in place before adopting new technology.
In my experience, a successful AI implementation still requires the fundamentals that have always been essential in technology adoption:
These questions are foundational and are critical determinants of a successful AI implementation. The foundational questions form a “quadrant” that you can consider when assessing and planning your AI initiative.
Organizations must evaluate AI adoption in the context of different risk categories. A few examples are:
Therefore, performing a complete and thorough AI readiness assessment is critical. The assessment is necessary to understand which risks are the most challenging for your organization. The assessment can provide guidance and help you focus on key risks during your AI implementation efforts.
In my experience, I've consistently seen organizations become captivated by new technologies, i.e., “shiny new object,” without fully thinking thorough implementation requirements. This pattern recurs with each new generation of technology:
"Twenty-five years back, people would say, 'I need to have a policy on information lifecycle management because we have a new IT system. Have the policy ready by Monday.' You can't prepare something like that by Monday just because you happened to read about it in an in-flight magazine or in a blog."
The tendency of C-suite executives to be seduced quickly by what technology can do — without considering all the elements needed for successful implementation — continues with AI-related technologies as the latest “shiny new object.” Organizations often fail to:
For AI specifically, data quality is perhaps the most critical foundation. Bad data can cause bias because your AI models are not being trained on representative information, thereby leading to poor decisions. AI is amplifying the risk of poor decisions.
Data quality has always been an important domain in information management, but with AI, the stakes are higher. Understanding the quality of your data — its completeness, accuracy, and reliability — is essential for AI success.
As I advise clients, if you don't trust your data because of quality issues, then can you trust the outputs from your AI models? If not, then can your outputs lead to incorrect decisions or even legal issues due to bias in the data?
Before rushing into your AI implementation, take the time to assess your organizational readiness, evaluate your data quality, and understand your risk tolerance in the context of the four “quadrants” mentioned above — people, processes, policies, and technology. Only by building on a solid foundation can your organization truly harness the transformative potential of AI.
This blog post is based on an original AIIM OnAir podcast. When recording podcasts, AIIM uses AI-enabled transcription in Zoom. We then use that transcription as part of a prompt with Claude Pro, Anthropic’s AI assistant. AIIM staff (aka humans) then edit the output from Claude for accuracy, completeness, and tone. In this way, we use AI to increase the accessibility of our podcast and extend the value of great content.