Strategy: Improving AI-Readiness through process changes

Challenge

A large DC based Multinational Finance Institution was implementing an enterprise-wide Agentic AI solution. Their current process was to hand-feed vetted information to the model. This produced stelar outcomes, but a very small-scale footprint as the time that it took to vet the data throttled the project speed. There was a desire to use enterprise intranet content for the project, as this was a primary source of data that supported the effort. There was, however, no way to pull in this much unstructured data in a manner that it would be of benefit to the AI implementation.

Approach

I met with five different verticals across the organization, analyzed their intranet sites, many of which went back nearly a decade. The developers of the enterprise intranet were interviewed to get their perspective on the problem and that data was combined with the perspectives of the employees on the ground who maintained the data. I identified that two key challenges, metadata and filing folder structure, were preventing their data from being AI-Ready.

This required change in three areas of scope:

• Planning remediation to make legacy content AI-Ready
• Planning to update their operating-models to ensure new content was AI-Ready
• Re-defining the scope of the employees to account for AI-Readiness in their role.

The solution desired by the client was to deliver a workshop to those employees. A three-day session workshop was created and delivered to reskill employees and have them develop remedial and forward-thinking plans, while also providing them to understand the shift in the scope of their role.

Outcome

Approximately 70-80 users in that role participated and walked away with an understanding of their new role, and plans on how to make it effective.

The stakeholder interview process also gave me the insight to propose a solution more in line with my transformation model. I presented an automation approach to complete the metadata creation process, eliminating thousands of hours spent manually addressing each file on the intranet individually. This approach would address the pain points vocalized by the learners, while simultaneously saving the organization thousands of hours of work, many months of delay in getting the content AI-Ready, and would save hundreds of thousands of dollars (conservatively) to the business.