A large financial institution which had acquired many smaller institutions and struggled with
multitude of conflicting policies across its many businesses and functions
including Finance and Risk.
Each policy was over 100 pages long and there were over 1000 policies across 40 countries.
In country regulators were not happy with conflicting policies on controlling and
reporting in the Finance function
What I did?
Held Daily Stand-ups/scrum calls to resolve the problem by:
understanding existing business process
identifying weaknesses and recommending solutions
working with business and IT stakeholders to deliver solutions
identifying and escalating risks alongside mitigations
implemented quality control loop as part of the handover at end of the project
Solution
Spoke to the in country teams to develop and publish an inventory of the Finance/accounting policies across 40 countries on SharePoint
Manual review of the policy was not possible due to the huge volume involved, it required use of an emerging technology like Artificial Intelligence.
Worked with the partners from Google and Microsoft to deliver private on-premises versions of Large Language Models (LLM) Gemini and ChatGPT. This removed the problem of data privacy and confidentially as the data was not sent over the internet.
Review by the AI revealed over 3k inconsistencies which were then prioritised according to
materiality and impact
Agreed with the client as to which items to focus upon and helped project manage them right from planning and resourcing to delivery of removing over 500 material inconsistencies