In my last post here, I outlined the steps I take when managing an international legal project. This follow-up question was both popular and insightful, so I'll attempt to answer. First, it's worth noting that there are many types of international legal projects from expansion projects (e.g. branches or subsidiaries in a new jurisdiction), acquisitions and joint ventures, global project launches and global compliance projects, to name a few. The factual matrix in each is different but, in all cases, the facts come from the business.

Since I've done a couple of posts on AI governance here and here, I'll use this global compliance project as an example. These types of projects usually have a copious amount of information to assess. As a wise mentor told me 27 years ago, preparation is the key to success. For an international project to be successful, it is important that the business organizes and prepares the information that constitutes the "facts".

In the case of AI compliance, I'd want to see:

1. An inventory of all AI in the organization, categorized by type (i.e. traditional or generative), organized by jurisdiction, line of business/functional area, purpose, who uses it and how. 
2. An inventory of all legal agreements in the business - client-facing, supplier-facing, employee-facing and any business-to-business arrangements.
3. An inventory of all relevant disclosures, warranties, pamphlets, brochures, instructions and the like given to customers or shown on the organization's website.
4. An inventory of all relevant policies and procedures, whether they apply to operations, procurement, marketing, privacy, data protection, sales, employees, etc.
5. An organizational chart identifying the individuals responsible for AI at the enterprise level and within each line of business/functional area, and finally
6. A direct line to the data scientists would be helpful.

The organization's guiding principles for the use of AI are also important. If a business uses AI to create shareholder value, and enhance the customer and employee experience, that will likely impact my analysis. If a business indicates that they use AI to create efficiences and replace jobs that require low intellectual capital in the hopes of retraining and moving staff to positions requiring more intellectual capital, that raises different issues that need to be analyzed.

In total, all of this information would constitute the "facts" that I would assess at the beginning of the project. This stage is not as labour-intense as it sounds. I would not read every document. Instead, I would look for representative documents. For example, most businesses use a model client agreement and vary that for different circumstances. I would read the model agreement and ask for a list of the categories of variations (e.g. price, term, termination, etc.). My end goal would be to eliminate duplicates, cull out irrelevant documents and cherry-pick representative examples. I would carry those examples forward through my analysis.